An Ethnoarchaeological Study of Iron-Smelting Practices among the Pangwa and Fipa in Tanzania 9781841716572, 9781407327334

This work is a comparative study of iron smelting practices among the Pangwa and Fipa peoples of Tanzania. The author di

233 102 20MB

English Pages [188] Year 2004

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

An Ethnoarchaeological Study of Iron-Smelting Practices among the Pangwa and Fipa in Tanzania
 9781841716572, 9781407327334

Table of contents :
Front Cover
Title Page
Copyright
Table of Contents
List of Figures
List of Plates
Preface
1. Introduction
2. Multiple worlds and multiple realities
3. A discussion of technology, the concepts and its contents
4. Re-enactments of smelting, how to describe and transcribe?
5. The Pangwa practice of iron smelting
6. The Fipa practice of iron smelting
7. Iron technology and symbolic reservoirs
8. From ethnoarchaeology to archaeology, from recent history to prehistory
9. Masters of metallurgy, masters of metaphors
10. Towards conclusions
Appendices
Bibliography
Monograph Series

Citation preview

BAR S1308 2004 BARNDON AN ETHNOARCHAEOLOGICAL STUDY OF IRON-SMELTING PRACTICES

B A R

Cambridge Monographs in African Archaeology 61 Series Editors: John Alexander and Lawrence Smith

An Ethnoarchaeological Study of Iron-Smelting Practices among the Pangwa and Fipa in Tanzania Randi Barndon

BAR International Series 1308 2004

Cambridge Monographs in African Archaeology 61 Series Editors: John Alexander and Lawrence Smith

An Ethnoarchaeological Study of Iron-Smelting Practices among the Pangwa and Fipa in Tanzania Randi Barndon

BAR International Series 1308 2004

ISBN 9781841716572 paperback ISBN 9781407327334 e-format DOI https://doi.org/10.30861/9781841716572 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

Contents

List of Figures...................................................................................................................................................................iii List of Plates .....................................................................................................................................................................iii Preface...............................................................................................................................................................................iv 1. Introduction...................................................................................................................................................................1 1.1 The study .................................................................................................................................................................1 1.2 Ethnoarchaeology and the quest for the traditional .................................................................................................4 1.3 The history of the region .........................................................................................................................................8 1.4 The people in the region ........................................................................................................................................10 1.5 Summary ...............................................................................................................................................................13 2. Multiple worlds and multiple realities.......................................................................................................................15 2.1 Multiple worlds and realities .................................................................................................................................15 2.2 Embodiment, between phenomenology and structuralism? ..................................................................................16 2.3 Metaphors..............................................................................................................................................................19 2.4 Comparisons ..........................................................................................................................................................19 2.5 Summary ...............................................................................................................................................................22 3. A discussion of technology, the concepts and its contents........................................................................................25 3.1 What is technology? ..............................................................................................................................................25 3.2 From tekhne to technology ....................................................................................................................................26 3.3 Technology and culture, social representations and styles ....................................................................................27 3.4 Habitus in technology............................................................................................................................................30 3.5 Science, magic and technology .............................................................................................................................31 3.6 Masters and magicians ..........................................................................................................................................33 3.7 Ritual performances in technology........................................................................................................................35 3.8 Towards a framework for technologies .................................................................................................................36 3.9 Summary ...............................................................................................................................................................37 4. Re-enactments of smelting, how to describe and transcribe?..................................................................................39 4.1 African iron working research ...............................................................................................................................39 4.2 What is iron working? ...........................................................................................................................................40 4.3 Description of iron smelting technologies .............................................................................................................42 4.4 Transcription of iron smelting technologies ..........................................................................................................45 4.5 First meetings and the ‘setting’ for research..........................................................................................................47 4.6 Summary ...............................................................................................................................................................53 5. The Pangwa practice of iron smelting .......................................................................................................................55 5.1 Introduction ...........................................................................................................................................................55 5.2 Mobilising a labour force ......................................................................................................................................56 5.3 Making charcoal, ulweko.......................................................................................................................................57 5.4 Smelting in secrecy, guarded by the umboocha ....................................................................................................59 5.5 Building a smelting furnace at Ing'angitoli............................................................................................................60 5.6 Building the furnace at Kuking’ande.....................................................................................................................63 5.7 Making tuyeres, bellows and beans .......................................................................................................................65 5.8 Collecting and preparing the ores ..........................................................................................................................67 5.9 Smelting begins by magical preparations and sacrifices .......................................................................................69 5.10 The chaînes opératoires in the smelting process.................................................................................................75

i

6. The Fipa practice of iron smelting.............................................................................................................................77 6.1 Introduction ...........................................................................................................................................................77 6.2 Mobilising a labour force ......................................................................................................................................77 6.3 Smelting in secrecy, guarded by the umpakasi......................................................................................................78 6.4 Performing magical skills......................................................................................................................................79 6.5 Building the smelting furnaces in Katumba ..........................................................................................................82 6.6 Re-using the furnace in Katumba ..........................................................................................................................86 6.7 Re-using and building furnaces at Tupa ................................................................................................................86 6.8 Making tuyeres and bellows ..................................................................................................................................87 6.9 Collecting and preparing ores................................................................................................................................87 6.10 Making charcoal and collecting firewood ...........................................................................................................88 6.11 Smelting begins ...................................................................................................................................................88 6.12 The chaînes opératoires in the smelting process.................................................................................................91 7. Iron technology and symbolic reservoirs ..................................................................................................................93 7.1 Smelting symbolism ..............................................................................................................................................93 7.2 Gender ...................................................................................................................................................................94 7.3 Iron smelting and embodiment ..............................................................................................................................97 7.4 The body as a container .........................................................................................................................................98 7.5 The open body .......................................................................................................................................................98 7.6 The naked smelter................................................................................................................................................101 7.7 The baby and the bloom ......................................................................................................................................102 7.8 The body and magic ............................................................................................................................................103 7.9 Chaînes opératoires and symbolism ....................................................................................................................106 7.10 Summary ...........................................................................................................................................................110 8. From ethnoarchaeology to archaeology, from recent history to prehistory ........................................................111 8.1 Operation, decoration and location......................................................................................................................111 8.2 Some survey results .............................................................................................................................................115 8.3 Summary .............................................................................................................................................................121 9. Masters of metallurgy, masters of metaphors ........................................................................................................123 9.1 Unity within diversity..........................................................................................................................................123 9.2 About the interior mosaic, several peoples but one tradition?.............................................................................130 9.3 From symbolic repertoires to symbolic reservoirs...............................................................................................132 9.4 Iron as a thing of value ........................................................................................................................................134 9.5 Summary .............................................................................................................................................................136 10. Towards conclusions ...............................................................................................................................................139 10.1 A second discussion of technology; its individual bodily and collective basis .................................................139 10.2 What did we observe in Ufipa and in Upangwa? ..............................................................................................142 10.3 What messages were conveyed and to whom? ..................................................................................................143 Appendices.....................................................................................................................................................................145 Bibliography ..................................................................................................................................................................163

ii

List of Figures Figure 1.1 Figure 1.2 Figure 1.3 Figure 1.4 Figure 4.1 Figure 5.1 Figure 6.1 Figure 8.1 Figure 8.2 Figure 8.3 Figure 8.4

The areas of research ...................................................................................................................................2 Sketch of the Fipa natural draft furnace, iluungu and the secondary refining furnace, ichinteengwe.........3 Sketch of the Pangwa forced draft furnace, liteende ....................................................................................4 Groups of people referred to in the text......................................................................................................11 The problem of scientific distancing (after Hastrup 1992).........................................................................45 Chaînes opératoires in Pangwa iron smelting............................................................................................76 Chaînes opératoires in Fipa iron smelting .................................................................................................92 Sketch plan of Fipa workspaces at the smelting site Tupa .......................................................................117 Sketch plan of Fipa workspace at Tupa imagined as an ‘archaeological’ site ........................................117 Sketch plan of a Pangwa smelting site as observed in a re-enactment at Ing’angitoli.............................119 Sketch plan of the Ing’angitoli smelting site imagined as an ‘archaeological’ site .................................120

List of Plates Plate 4.1 Plate 4.2 Plate 5.1 Plate 5.2 Plate 5.3 Plate 5.4 Plate 5.5 Plate 5.6 Plate 5.7 Plate 5.8 Plate 5.9 Plate 5.10 Plate 5.11 Plate 5.12 Plate 5.13 Plate 5.14 Plate 5.15 Plate 5.16 Plate 5.17 Plate 5.18 Plate 6.1 Plate 6.2 Plate 6.3 Plate 6.4 Plate 6.5 Plate 6.6 Plate 6.7 Plate 6.8 Plate 7.1 Plate 8.1 Plate 8.2

A Fipa blacksmith at work in the village forge, Katumba 1990 .................................................................51 A Pangwa blacksmith at work in his bicycle-wheel driven forge, Mkiu 1995 ............................................52 The furnace guardian, umboocha, Ing’angitoli 1993. ................................................................................59 Making the furnace foundation trench, luutu, and pit, umlindi, at Ing’angitoli 1993 ................................61 Building a furnace at Ing’angitoli in 1993 .................................................................................................62 The furnace, liteende, when ready built, Ing’angitoli 1993........................................................................62 Making the furnace pit and trench for the second re-enactment, Kuking’ande 1995.................................63 Making the furnace shaft, Kuking’ande 1995.............................................................................................64 Washing the furnace, liteende, in ritual water, kiwalasha, Kuking’ande 1995 ..........................................64 Making tuyeres, Ing’angitoli 1993..............................................................................................................65 Drum-bellows, n’goma, Ing’angitoli 1991 .................................................................................................66 Tuyeres, tubes and bellow arrangements, Kuking’ande 1995 ....................................................................66 Preparing imono at Ing’angitoli 1993........................................................................................................67 Collecting ore at Manga.............................................................................................................................68 Ores prepared at the smelting site, Ing’angitoli 1993 ................................................................................68 A goat sacrifice, Kuking’ande 1995 ..........................................................................................................70 The medicine pot, ing’anjo, in front of the furnace, Kuking’ande 1945 .....................................................70 Smelting in liteende at Kuking’ande 1995, note slices of kiwalahsa plant on top of ore...........................71 Smelting in liteende, Kuking’ande 1995, note the ‘windmaker-hole’, kwibepo..........................................71 The front tuyere port, umlomo, with its tuyere inserted..............................................................................73 The furnace caretaker, umpakasi, hanging over the main tuyere port on the high shaft furnace, Katumba 1991. ...........................................................................................................................................80 The furnace caretaker, umpakasi, inside the furnace chamber, Katumba 1990. Note the bundle of medical barks, ingailo ..................................................................................................80 Sorting magical substances, ifingila, in front of the furnace, Tupa 1991 ...................................................81 Yakasike iluungu, a sacrifice for the furnace, Katumba 1990 ....................................................................83 The secondary furnace, ichinteengwe, when ready built, Tupa 1991 .........................................................84 Celebrating the high shaft furnace as a bride, kanawiinga, Katumba 1990, note scaffoldings and flowers .....................................................................................................................85 The furnace in our re-enactment at Tupa 1991, note permanent scaffoldings ...........................................85 A secondary refining of the iron in the low shaft furnace, ichinteengwe, Tupa 1991.................................89 A smelter wearing isuuli, Katumba 1990..................................................................................................100 An old Fipa smelting furnace ...................................................................................................................116 An old Pangwa furnace, liteende..............................................................................................................118

iii

Preface Iron as a material is omnipresent in technologies, economies and politics, while for most of us in our daily lives it is a matter of less significance or value. When we think about iron we imagine rusty old stuff, car- or shipwrecks as huge piles of scrap metal. In Eastern Africa iron was once a material of value and those who won iron out of ores and forged tools out of blooms were highly admired. Most of them were not ‘professional specialists’ but part-time specialist or craftsmen who produced their own food and were farmers as the rest of their populations. In my opinion iron blooms and hoes made in pre-colonial Eastern Africa had an interconnected value and social significance. Members of iron producing communities have had to find ways to understand their technology and to be able to adjust it (both technological development and knowledge of the technology in itself) to their political, economic, social and cosmological views and values. The simple fact that iron metal is produced by people and not by nature implies that one has to know what is an ore in order to successfully extract iron from ores (Kriger 1999:30). One needs to know what kinds of hardwood are necessary as fuel and especially one must know how to keep the temperatures controlled within the smelting furnace. And, one has to know what rituals to conduct and what medicines and magical substances to use in order to transform ores to iron assisted by the supernatural, ancestors and spirits. In the past those who produced the iron, the master smelters, were aware of their importance because when the blooms were forged, the iron they had produced by skills and ‘know-how’ were transformed into tools, weapons and prestige objects (ibid.). Often were those who possessed this valuable knowledge of transformation of ore into iron given a high social status in real life, in myths and in legends (i.e. Eliade 1958; Herbert 1993; Kriger 1999; Maret 1983; 1985; Schmidt 1997). My work is based on and has developed from field studies of iron smelting practices among the Pangwa and Fipa peoples in southwestern Tanzania.1 My encounters have been with master smelters, smelters and blacksmiths in Upangwa and Ufipa who through our re-enactment of smelting practices tried to recreate a technological activity once part of their social life. Some of the aged smelters I met have experienced colonialism, two world wars, attacks from Arabs and much internal turmoil. Many of their sisters and bothers were taken as slaves by the Arabs or forced to work on sisal-, coffee- or teaplantations. They were forced to abandon old ways of living, marrying and working and had to move with their families into new villages. Some of them experienced all this while growing up and while they were boys participating in iron making by blowing the bellows of smelting furnaces or assisting their fathers in the forge. Once they were young men they became apprentices in an iron working team. Finally, some of them became master smelters. In their lifetime they have experienced the decline of value attached to their profession and the decline in value of their locally made iron objects. Today the Pangwa and Fipa live in small villages of mostly square brick built houses. Their villages are commonly located close to the main roads between Njombe and Ludewa (in Upangwa) and Namanyere and Sumbawanga (in Ufipa).2 The villages still cluster around old mission centres. The Fipa and Pangwa are modern Tanzanians, living in rural areas struggling to survive. They are farmers that care for their fields and gardens in order to produce enough maize or groundnuts. Working as blacksmiths in their village forges, using scrap metal that is reforged into tools, doorlocks and other repairs, they are struggling for a surplus that can help them raise their children and enjoy life. What they more than anything want is to provide their children with means to marry and establish their own households. One of the goals with my project is to embark on a discussion of and comparison of magic and metaphors expressed in technology among two different groups of ironworkers. My aim is ‘to produce more raw material’ in Brown’s words (1995) and towards my conclusions I will comment upon how my observations may provide a conceptual framework for ethnoarchaeological studies of iron technology and for archaeological interpretations of iron working sites. Which aspects of technological cum symbolic connotations are relevant to look for in other empirical cases in Africa and which aspects might be of relevance to our interpretations of the past, in Africa or beyond? Some of the ethic problems this task is faced with and must try to overcome will be addressed below. It is anticipated that my approach will illustrate the value of including variation into more uniform models that consider cross-cultural inferences of symbolic systems and meaning of morality to be based on the same use of metaphors and symbolic repertoires. This study is about local perceptions of iron smelting but I will address both aspects of unity and diversity. I have focused on how the Pangwa and Fipa themselves experience, explain and conceptualise the technological transformation that occurs when iron ores are reduced to bloomery iron.

1

In this work I follow conventional ethnography by including lexicostatic or indigenous terms of the smelting process and related objects. I strongly believe in the information being stored in the lexical terms used in the smelting process. Hence, throughout this work I will refer to local Pangwa and Fipa terms (see also Appendices A and B). 2 I have chosen to call the regions where I have conducted fieldwork according to how the Pangwa and Fipa themselves talk about their regions. Thus, Upangwa refers to Ludewa region while Ufipa refers to Rukwa region.

iv

Many people in many places have been engaged in, supervised and supported me in completing this project. First and foremost I acknowledge the openness and friendliness I was met with in Tanzania, at Archaeology Unit, University of Dar es Salaam, especially Dr. Seth Nyagava and Professor Felix Chami. Among the many kind and helpful people I met in the field I am in particular deeply indebted to the former master smelters and blacksmiths who without hesitation left their daily tasks in order to embark on this study of iron working. In Ufipa Mzee Stefano Malimbo, Mzee Andrea Mwanisawa, Mzee Nicolao J. Kanyama, and especially Paskary Mwanisawa, and in Upangwa Mzee Longines Mtweve, Mzee Andreas Mtweve, Mzee Benedicto P. Damakira, Sylvester Mgimba and the late Amos Mgimba. Although we did communicate language barriers were many and grasping their way of conceptualising the smelting process would not have been possible without the assistance from Bilham Kimati, George Gardias, Emanuel T. Kessy and Amin Msuya. In Rukwa, Regional Cultural Officers Mr. S. N. Mwiyane and Mrs. V. J. Kapufi and Sumbawanga Town Director Mr. J. Magogo made my stay in Ufipa possible similar to District Commissioners in Njombe and Ludewa, Mr. O. M. Kapinga and Mr. A.R. Achireka who followed up and assisted me in line with research permits granted from Tanzania Commission for Science and Technology. Professor Randi Håland, University of Bergen, was my supervisor for my Cand. philol. thesis on Fipa iron working and I am also grateful for her involvement during initial stages of this research. Professor Gro Mandt, University of Bergen, and Professor Bjørnar Olsen, University of Tromsø, have supervised me, commented on drafts and guided me through the writing of this dissertation. Bjørnar Olsen’s generosity in sharing ideas, views and sources of inspiration I greatly acknowledge. Gro Mandt has encouraged, inspired and supported me throughout this project, I am deeply grateful for her comments and critiques. Professor Nicholas David, University of Calgary, has provided useful comments on several chapters and I have certainly benefited from his insight in terms of African iron working. Dr. John Sutton, University of Oxford, kindly handed over to me fieldnotes from his surveys of iron working in the Southern Highlands Region as well as an undocumented manuscript on Pangwa iron smelting written by Hans. Strinimann. Among colleagues in Bergen I will especially mention Knut A. Bergsvik, Øystein Geber, Kari K. Kristoffersen and Søren Diinhof, Prof. Nils Gilje and Kathryn Bousman. Ellinor M. Hoff has redrawn my field sketches and maps. Leeta R. Thomson and Greg Rowse have proof-read parts of the manuscript and Brian A. Farrelly has proof-read the final version. Bergen, March 2001

v

vi

Introduction

1. Introduction Are we as Englishmen, after all that has been done, to allow these lands and this interesting people to pass beyond our influence and sympathy (Kerr-Cross 1890:295)1 The third stage was the forging, ukusula, of iron, ulolo, into objects. The forging was conducted over an open fire or a small three-walled stone forge. The Fipa forge, impeembe, was a grass-thatched hut located within the village. Two of the same bellows, umwuuwa, that were used in the secondary refining furnace were also used during forging. Among the most common Fipa produced iron objects were the male hoe, ise, used when making compost mounds, a female hoe, impalanga, a knife, inkuumbwa, razors, pinchers, nails, hammer, axes, sickles, drills, needles, spears, arrows, bracelets, earrings, and finger rings. A detailed and idealised description of Fipa iron working has been presented in Traditional Iron working among the Fipa (Barndon 1992).

1.1 The study. This work is a comparative study of iron smelting practices among the Pangwa and Fipa in Tanzania (figure 1.1). I will discuss local concepts of metallurgy through a step by step observation of the two processes. I have three interconnected agendas. I will embark on a description of technological agency with a partial implementation of a chaîne opératoire approach (Lemonnier 1992) as observed in several re-enactments of smelting practices. Secondly, I attempt to discuss within a comparative frame the multiple and shared levels of meaning and experience in symbolic repertoires (Bekaert 1998) and symbolic reservoirs (Sterner 1992) held by technicians during these observed technological processes. I will demonstrate with a focus on magic and metaphors the link between perceptions of the body and concepts of the technology. Finally, it is anticipated that a shift in our concepts of technology can be of use also for studies of iron technology where there are no living informants to answer questions about the use or meaning of specific objects or actions that were part of reaching technological goals. For most researchers concerned with African metallurgy, the Fipa’s tall natural draft smelting furnace is familiar.

The Pangwa iron working process is less familiar to Africanist archaeologist simply because their smelting practice has not been observed before my field work in the Lugarawa area in Ludewa in 1993 and 1995 (cf. Barndon 1993, 1995, 1996b, 1999a).3 The furnaces of this region varied in height but were of Kense (1983) and Childs’ (1992) Type B; a low forced draft furnace (cf. Appendix D, figures 1 and 2). Based on testimonies of former smelters Stirnimann provides a detailed description of Pangwa iron making practices of Manga and Mlangali (Strinimann 1969, 1976). In 1968 Sutton conducted a survey of the northeastern Southern Highland region (Sutton 1971b, 1985). This survey of Bena, Kinga and Matengo furnaces has confirmed that iron working was common among the highland peoples until the beginning of this century.

The Fipa method of producing iron has been documented by travellers and missionaries and been observed and studied by ethnographers and archaeologists (Barndon 1992; Greig 1937; Lechaptois 1913; Mapunda 1995; Robert 1949; Sutton 1971b, 1985; Wembah-Rashid 1969, 1973; Willis 1966, 1981; Wise 1958a, 1958b; Wyckaert 1914 and see Appendix D, table 1). Fipa iron working, including the forging, was a three-staged process. The first stage was conducted in an impressive 2 to 3 metres tall furnace. These furnaces are generally classified as type C furnaces (Childs 1992; Kense 1983, cf. Barndon 1992 and see Appendix D, figures 1 and 2). ‘Brown ores’ of limonite were reduced in this high shaft furnace, iluungu, fuelled by charcoal and green wood. A secondary refining was carried out in a smaller forced air draft furnace, ichinteengwe, varying in height between 40 to 50 centimetres and of the B type (cf. Appendix D, figure1). This furnace was used in order to consolidate the bloom and was operated by three bag bellows of goatskins connected to bamboo sticks and tuyeres made of clay.2

Pangwa iron making including the forging of the blooms was a two-stage process and the total ore reduction took place within one furnace called liteende. The Pangwa smelting furnace was an approximately 1 metre high shaft furnace operated by three pairs of drum bellows, ing’oma, connected to bamboo sticks, and three tuyeres, ing’elu or ing’elwa, that were inserted into the furnace chamber. When bloomery iron, itaale, (uwutali) had been produced it was taken back to the village and the forge, ulufuumbu. The forge was an open structure that operated with flat wooden bellows. In the forge the most common objects made were the hoe, likimilu, a knife, nyengu, bells worn around the feet during smelting and dancing ceremonies, mangala, axes, maimbachu, and spears for hunting, mamukoha.

1 These words stem from Kerr-Cross and were his concluding remarks after a geographical survey of the country between the Lakes Nyasa, Rukwa and Tanganyika. 2 In the final days of active smelting and in our re-enactment of smelts the Fipa smelters used iron tubes instead of bamboo sticks.

3

Tanzania Commission for Science and Technology (UTAFITI) or Costec, authorises Research Permissions and is located in Dar es Salaam. Costec has reports on file from each field season carried out by the author.

1

Randi Barndon

KENYA RWANDA LAKE VICTORIA BURUNDI

TANZANIA LAKE TANGANYIKA

*

Dar es Salaam ZAIRE

LAKE RUKWA

INDIAN OCEAN

UFIPA UPANGWA LAKE MALAWI

ZAMBIA

MOZAMBIQUE

Figure 1.1 The areas of research

were metaphorically linked because they all transform substances into an irreversible state (Collett 1993, cf. also Barndon 1992,1996b; Childs 1991b). Herbert has discussed how ‘gender’ and ‘age’ categories were used in guiding the smelting process and the smelters. Hebert’s gendered model or ‘procreative paradigm’ refers to the smelters’ notions of the female life cycle as it was expressed in the smelting rituals (Herbert 1993). Schmidt argues that not only human sexual intercourse is metaphorically linked to the smelting of iron but that the complete female cycle is the structuring principle (Schmidt 1996:100). Rowlands and Warnier suggest iron technology to be linked to notions of fertility and violence. They have focused on metaphorical links between women and furnaces in terms of magic and medicines and acknowledge the use of oppositions and binary codes as a structuring principle during smelting (Rowlands and Warnier 1993). Maret has discussed the close association between leadership and metalworking, the ambiguous role of the blacksmith who is ‘playing with fire’ as linked to myths of leadership in terms of the RoiForgeron, widespread in Central Africa (Maret 1982, 1985)

African iron making is a huge research field and Lemonnier has forcefully illustrated that describing technological actions is a full-time field occupation (Lemonnier 1992:28). I will attempt to look into what Lemonnier calls specific knowledge of know-how, conscious and unconscious skills and experiences (ibid.: 6), because this knowledge is connected to the meanings the practitioners of a technology give their actions and processes. Studies of African iron working have been intimately linked to the study of social, ritual and mythical aspects. Cline’s Mining and Metallurgy in Negro Africa (1937), Kense’s Traditional African Iron Working (1983) and Herbert’s Iron, Gender and Power (1993) all address African metallurgy within a cultural framework. From these comprehensive studies it is apparent that African iron smelting was metaphorically linked to fertility, procreation and childbirth. The smelters were the ‘husbands’ of their ‘furnace wife’ who transformed ores to iron and ‘gave birth’ to the iron blooms. The furnace ‘husband’ was the master smelter and taboos against sexual intercourse (seen as adulterous behaviour) were mandated for the participants in a smelt (Brelsford 1949:28). Women (in menstruation) were not allowed entry at the smelting sites because they could disturb the transfomative process.

Cline suggested the Pangwa (Pangwe) iron smelting practice to be analogous to the Fipa smelting rituals and pointed at the importance of sexual taboos and associations with the ancestral spirits (Cline 1937:123). He highlighted four points that made Pangwa iron smelting more ‘than a technical affair’ which it is equally

Collett explains iron smelting symbolism in terms of a semantic principle in which women, pots and furnaces 2

Introduction

NTAANDA PEEP HOLE

PALINYINA

PALISI TUYERES

.

.

UMPAKASI FLOOR

NATURAL SOIL

ULUFULA FOUNDATION DITCH

LEVEL AT FURNACE BASE ON OLD PRODUCTION SITES

0

IRON CLAY TUBE PLASTER TUYERE

200cm

BELLOW

STONE AND GRAVEL PLATFORM

N PIT FOR MAGICAL SUBSTANCES

FOUNDATION DITCH

0

50cm

Figure 1.2 Sketch of the Fipa natural draft furnace, iluungu and the secondary refining furnace, ichinteengwe

essential to focus on and include in studies of meaning and metaphors in re-enactments of iron smelting today as it was then when iron smelting was practised. These were ‘the fire cult, the ancestral cult, a belief that the nature of plants can exert a sympathetic magical effect on the products of industry, and still a more general conviction, ‘widespread throughout Negro Africa; that sexual activity of any sort may cause the work to go wrong’ (ibid.: 125). Although Cline was correct in his observations of Pangwa iron smelting as composed of these ‘non-technical elements’ he based his description of the Pangwa on Tessmann’s work conducted among the Pangwe in Western Africa (Tessmann 1913).4 The important point here is that Cline included multiple levels such as a procreative paradigm, sexual taboos, magic and the

importance of assistance from ancestral spirits. In this study of Pangwa and Fipa iron smelting I will question the (false) categorisations made between humans and non-humans in studies of technology and in particular in archaeology. I will question the common made opposition between technology and magic, technicians and magicians, people and furnaces and gender oppositions through an inquiry into local concepts of the smelting process. In addition to my re-enactments of iron smelting I base my analysis on Willis’ ethnographic work among the Fipa and Stirnimann’s studies of the Pangwa (cf. bibliography). Willis and Stirnimann’s detailed ethnographic research, concerned with social relations, worldview, structures in politics and cosmologies, relate to the cultural contents of and symbolic repertoires that the smelters used in their iron making.

4 Cline has no reference to the Pangwa of Tanzania. Tessmann (1913) studied the Pangwe in Western Africa and Cline describes this peoplegroup in another chapter in his book (Cline 1937:38-39).

3

Randi Barndon

FURNACE, LITEENDE AT ING´ANGITOLI UPANGWA

1m

MAVELE

KILULEPU

0 UMLOMO

ING´ANJO

UMLINDI

1m

FURNACE seen from back with the” wind-making hole” KWIBEPO

KILULEPU

0

UMLINDI

Tegn.E.Hoff

Figure 1.3 Sketch of the Pangwa forced draft furnace, liteende

The motive for many studies of African indigenous iron working has been, as Killick argues, to record some of the processes while it was still possible (Killick 1990:3 and cf. also Barndon 1992:29; Schmidt 1997:ix). Killick lists a number of additional reasons to continue to study African iron working, all of them relevant. He believes it provides us with answers to current issues within anthropology, archaeology, the history of technology and African economic history. There are still an enormous number of questions both in terms of metallurgical, social and ideological issues that are relevant to inquire more information about. Particular topics of interest are aspects of African systems of thought, technological change in pre-colonial Africa, the social context of prehistoric metallurgy and interpretation of metallurgical remains (Killick 1990:3).

1.2 Ethnoarchaeology and the quest for the traditional It is commonly stated that in our search for a better understanding of prehistoric material culture, we must conduct traditional ethnography. The argument is that ethnography is what gives us the raw material for analogies that we can apply in our attempts to understand the past (Brown 1995:ix, cf. also Gould 1978, 1990; Gould and Watson 1982; Haaland 1988). Brown correctly notes that for decades this ethnographic raw material has been absent from anthropological studies (but see Appadurai 1986; Arnoldi et al. 1996; Miller 1987; Moore 1986). Hence, ethnoarchaeology has developed as a subdiscipline. Ethnoarchaeology might be framed as: ‘.. research that involves investigations of the material culture, behaviour and beliefs of present-day populations, with the object of that research being the generation of propositions about the cultural contexts of residues 4

Introduction recovered from archaeological occurrences’ (MacEachern 1996:247).

working practices is however related to archaeology because another object of this study is to describe my observations of smelting activities in such a way that they can make available inferences about technological agency in the past and about societies in change. It is anticipated that parts of this study among these others and their recent and present past can produce paths to or guidelines about where we should gather information valuable to the study of technologies.

Lane’s detailed observations of what is involved in the practice of ethnoarchaeology reads as follows: ‘Ethnoarchaeology is partly a methodological procedure whereby archaeologists use standard techniques of archaeological recording in conjunction with anthropological methods of participant observation, informal and formal interviews, and thus the questionnaires whilst living and working with a contemporary community. Ethnoarchaeological fieldwork also involves observing the inter-relationships between different cultural, physical and social consequences. It is also in regard to this approach essential that the members of the local communities are interviewed about the causes, reasons and meanings of their actions. Some command of the local language, or languages, therefore is also helpful’ (Lane 1998b:178).

How then can I, an outsider, a foreigner and an other, postulate or make any hypothesis and about their modes of thoughts? This question relates to the anthropological debate and philosophical considerations of meaning and otherness (Augé 1998). Bernstein and Geertz were early in their discussions of the possibility of anthropological knowledge of ‘the way natives think’ (Bernstein 1983: 94; Geertz 1976). Latour has however, made us explicitly aware of problems within the modern and post-modern era, which he himself rejects. Latour suggests it is time to recognise and practice that there is no great divide between us and the other (Latour 1996) or actors and artefacts (Latour 1999, Olsen 2000).

Ethnoarchaeological fieldwork contains experiences beyond language, beyond words and definitions. Most of these experiences are very similar to those a field anthropologist is faced with. Therefore, in terms of field work and transcription of data the boundaries between the disciplines ethnoarchaeology, ethnography and social anthropology are not easy to maintain (David 1992; MacEachern 1996). Ethnoarchaeological practice is a multi-levelled theoretical and multi-methodological approach towards the past (David 1992). It is also, or rather has become, theoretical about the relations between material objects and meaning.

But the question of comparison is about us and them and how to comprehend or translate modes of thought and actions of other people. We have become more and more aware that we undertake this translation from a Western baseline and that what we do is a mapping of the ideas and practices of other peoples onto Western categories of understanding, in order to develop a language that can be used in comparisons (Tambiah 1993). Providing relevant data on material culture that can be used in comparisons is the basis in all contemporary ethnoarchaeological work. But as Tambiah has noted ‘the translation of ‘cultures’ and their comparative study raises not only questions about ‘us’ and ‘them’, it ultimately also raises the issues of morality and of rationality within Western scientism as a paradigm’ (ibid.: 3). Being aware of the dominance of the Western mode of thought and the scientific language that has been applied in order to describe and interpret cultural categories allows us to also recognise its preeminence in studies of technology, and in particular iron metallurgy in non-western societies. Trying to describe the Tanzanian Pangwa and Fipa peoples’ iron smelting practices is not a task that can be performed without reflections on how our Western mode of thought, our ‘scientific’ comparative baseline has been and is used for this translation (Appadurai 1988; Tambiah 1993).

The major object of conducting ethnoarchaeological research is that it must contribute to our understanding of prehistory. But ethnoarchaeology is first of all the study of material culture in contemporary, often small minoritygroup societies. Ethnoarchaeologists attempt to show how the results from their studies can make sense of and make inferences about a set of material culture in the archaeological record. If ethnoarchaeology is to contribute to our understanding of prehistory, what matters then is to somehow generalise from the present to the past; to move from individual experiences to a more mutual worldview regardless of time and without loosing the particulars. Is this possible without the loss of high ethical and moral standards in our research and results? My work, although concerned with the past, is not about the prehistoric past as we normally conceive it. Rather, it is about information gathered through reconstruction or re-enactments of the past in the present and encounters with others.5 My comparison of Pangwa and Fipa iron

More than two decades ago Barth advocated a critical attitude to all aspects of one’s own and others’ work because of a failure to transform anthropology from a rich man’s hobby to a concerned human discipline (Barth 1974:99).6 In a critical review of a colleague’s monograph Barth pointed at the crises of anthropology in relation to aspects of responsibility and humanity. He discussed the moral responsibility we have when

5 The term is used here ‘... in all senses of ‘the other’ (Augé 1998:xiiixvi) and I am fully aware that analysis of ‘the other’ are including if not mirroring ourselves. Because ‘anthropologists not only represent their experiences of an ‘other’ culture in the text they also constitute and produce their experiences and themselves (Moore 1995:117, original emphasis).

6

5

Italics by the author.

Randi Barndon inherent danger of exoticising, sexualising and primitivising those peoples who do not seem to understand the real meaning of their actions’ (Bekaert 1998:16, but see also Appadurai 1988). If these constructed notions of traditional include an approach that communicates what is traditional, as something underdeveloped, primitive and native how can ethnoarchaeology, like anthropology, be a concerned human discipline? How can ethnoarchaeologists’ search for the ‘traditional’, the ‘authentic’ or the studies of ‘living traditions’ be compatible with a participation in the anthropological project to better comprehend the world? Within anthropology ‘traditional societies’ have been and are compared with the ‘Western’ or the ‘modern’ world, in line with the idea that this is a part of the anthropologists scientific tasks; the ultimate project of comparison (Eriksen 1995). Critical anthropology is a field with various objectives (e.g. Appadurai 1986, 1988; Clifford 1986; Clifford and Marcus 1986; Comaroff and Comaroff 1990; Crapanzano 1986; Hastrup and Hervik 1994; Hastrup 1995; Kempny and Burzda 1994; Latour 1996). From the early critics followed a substantial amount of self-reflexive monographs that have made anthropologists aware of the actual manner in which their translations into a Western mode of thought are carried out (Tambiah 1993:3).

embarking on ethnographic research and the ethnoarchaeologist is embarking on exactly such research. Both the monograph by his colleague and the review were written in the early 1970’s but the comments Barth made are as relevant today as they were then. Barth also discussed how we convey our messages of indigenous people within the research communities; another issue that seriously needs to be discussed within contemporary ethnoarchaeology. In short, Barth pointed out that the pursuit of research in social and cultural anthropology entails circumstances of fieldwork and analysis which are special and which therefore require special ethics and competence both of a professional and of a personal nature (ibid.: 100). First on Barth’s list of circumstances in fieldwork is the fact that we impose ourselves on other people. We come unasked, uninvited and often incompletely perceived to other people in other countries and in other societies (Barth 1974:100). As ethnoarchaeologists we may tell our hosts that we will study iron working. But, do we also tell them once we are at home at our university desks, when our research among our hosts has been completed, when we have left the village but left behind blankets, metal casseroles and plastic containers, that we will compare their traditions with now extinct and prehistoric ways of smelting iron? This is in itself a huge problem that we cannot merely escape from but must consider as equally important to the actual archaeological applicability of our research (Barndon 1999). Not only should we consider how we impose ourselves on other people but also how we behave when we are there and how we part (Barth 1972:100).

Ethnographic and anthropological fieldwork with its descriptions of research methods has a history of discussing these issues that goes back to Malinowski (1935). Contemporary ethnoarchaeology, on the other hand, is surprisingly immature because there are few recently discussed directions on how to actually conduct ethnoarchaeological fieldwork. The form of participantobservation, the degree of participation as opposed to degree of observation, and specifically how to do all this with the eye of an archaeologist is rarely discussed (but cf. Gould 1978; Stanislawsky 1973; Stark 1993; Stiles 1977). Contemporary African ethnoarchaeology is gradually taking into account these, mostly well justified criticisms within anthropology, and has allowed similar arguments to shed light on the ethics of doing ethnoarchaeology (e.g. David 1992, 2001; Lane 1995, 1996, 1998, 1999a; MacEachern 1996). This is part of an overall essential philosophical shift that considers ethics and moral issues important within our research practices.

Barth’s second point is that we legitimise our work because we call it research; a kind of research that often has only a vague promise of applied usefulness (cf. David 1992, 2001). Barth notes that because of this we bring objectivity into our standards, because we want our research, our results, to appear scientific (Barth 1974:100). Finally, Barth points to the typicality of social or cultural anthropology, and I will add ethnoarchaeology, in that we use ourselves as a research tool in participant observation, including ‘our intuition, our charm, our emotion and our abilities’ (ibid.: 100). Barth discusses how we apply and assimilate our observation within the academic world. For an ethnoarchaeologist it might seem legitimate to look for any kind of materialised traditionality because the raison d’être for ethnoarchaeology lies in the observations and documentation of pre-industrial, pre-mechanical, premotorised ways of producing and using material culture. But are our moral concerns left out or neglected in order to produce more convincing arguments in order to present a more authentic traditionality? One can correctly claim that ethnoarchaeological practice privileges the observer over the observed (Burg and Elison 1992 in Lane 1996b). ‘Since the focus is mainly on other peoples there is an

What I observed and will describe in the following pages is the result of what Schmidt has called an experimental approach (Schmidt 1997:11). An experimental approach is a demonstration and a reconstruction of iron working. But the smelters who participated were also learning from the process because they were no longer themselves practising traditional iron working. Ani thoroughly criticises Ethnoarchaeological fieldwork among contemporary people must take into account the ethical discussion in social and cultural anthropology. It is high time our practices, our way of conducting ethnoarchaeological fieldwork, is seriously considered. How

6

Introduction that might when used in other non-archaeological contexts be dangerous and used against indigenous peoples (MacEachern 1996).

we transform our experiences and observations into our ‘scientific data’ should no longer merely be swept under the carpet but considered an important part of our creativity with the data (cf. Hastrup 1992). Accepting this, we must never forget that what we do, the writing of some of technological, social and symbolic aspects of the lives of marginal peoples or minority groups, might also have an impact (even a dangerous one) that is never fully foreseeable (Comaroff and Comaroff 1992:12). Therefore, we must avoid forms of inferences that ridicule or otherwise discriminate and we must consider this especially in studies that are concerned with people.

The agenda of the post-structuralists was and is a selfcritical view on the practice of doing archaeology and writing prehistory (Perceul 1991a, 1991b; Perceul and Hodder 1996; Potter 1991). This agenda was similar to what was the major concern of critical anthropology (Clifford 1986; Marcus and Fischer 1986). Olsen clarifies many of the problems related to post-processual and specifically post-structuralistic archaeology. It is, as he points out, important to gain a better understanding of all the practices that we as archaeologists are occupied with through the production of archaeological knowledge (Olsen 1997:282).

There is, along with more specified methodologies, a need for a deeper awareness of the use and abuse of contemporary people (Lane 1996), because these people might have or once have had their own histories and archaeologies irrelevant of Western modes of thought and Western archaeology (Lane 1995, 1996). Very little African based ethnoarchaeology has actually been applied within African archaeology while we have a long history of using this ethnography on the rest of the world (Lane 1996). It is important to be aware of the discrimination that lies within the Western philosophical tradition and its efforts at imposing meaning on ‘others’, peoples in other non-western cultures.

Much of the early enthusiastic post-structuralism has now been rejected by many archaeologists because of a disbelief in any pragmatic theoretical applicability of its baseline (Olsen 1997:280). This rejection has been based on ‘a misunderstanding of the post-structural project’ (ibid.). The critics claim that the emphasis on textuality is an impossible concept within archaeology and will even lead us into ‘a hyper relativism’ (Trigger 1989:406). Olsen has demonstrated that this was not the object of the post-structuralists, but that it was, and still is, an important starting point for discussions of the archaeological practice (Olsen 1990, 1997). That is, archaeology in practice, in terms of the activities in the field, the transcription of these experiences into writing, indicates how important it will be in the future that we develop better methods of making explicit how we acquire and transcribe our data. The point is to make visible the processes of translations.7

Numerous societies on the African continent have been a source of analogies for interpretations of archaeological remains, not only from the African archaeological record but other past societies in other parts of the world (Lane 1996b:727). But this is as pointed out by Barth masked and given a ‘scientific label’. It is high time ethnoarchaeologists try to take this debate seriously. My concern is with how we as ethnoarchaeologists use contemporary people in our research and consequently how archaeologists use these same people as ‘empirical examples’, as ‘evidence’ and as ‘parallels’ or ‘analogies’ to prehistoric people. What notions of the prehistoric or recent past are we forcing upon the people in contemporary societies through our material culture studies? What notions of the present and a supposed ‘traditionalisation’ of contemporary people, often minority groups, are we imposing on the past when we use their traditions (technologies, customs and rituals) as references for analogical and experimental thinking about the past.

Post-structuralism has taught us that the past is continually recreated in accordance with our contemporary image of the world and ourselves as research tools. The only ‘thing’ that remains in a static, unchanged position and remains unquestioned within ethnoarchaeology as a research practice, in spite of our new awareness of subjectivity, is the search for, and use of indigenous people as ‘traditional’. The present is neglected in most ethnoarchaeological studies. Empirical data from ethnoarchaeological surveys have been applied in reconstructions of the past. But we have a moral responsibility and it should not be legitimate to transform the actual observed events into appearing as traditional as possible for the benefit of other human beings. Within social anthropology this is criticised (Appadurai 1988; Barth 1972; Bekaert 1998; Vansina 1990) because it contradicts the anthropological project, while within ethnoarchaeological studies this same search for and representations of the other as traditional is believed

A cautionary tale about how Africans are represented as traditional comes from South Africa where the peoples traditions were used as a discursive weapon that denied Africans full participation in the country’s political life (Spiegel 1994). There is a common ethnoarchaeological framing of material cultural studies in stills. They, the people with traditional ways of life, still believe in evil spirits, still practice initiation rituals, still smelt iron or still use stone tools. Spiegel has illustrated the dangerous political rhetoric in this tradition (Spiegel 1994). Ethnoarchaeologists are in a constant quest for these stills

7

Hodder (1999) has, like to Olsen (Olsen 1987, 1990, 1997; Olsen et al. 1993), demonstrated a post-processual approach towards archaeology as a field practice. This had led to an implementation of a more reflexive approach towards data collection.

7

Randi Barndon necessary and hardly ever questioned. Contrary to the realities that face anthropologists, ethnologists and their quests for the present and real world, the ethnoarchaeologist is trying to contribute to an understanding of the past and wants to contribute to the writing of prehistory. Clearly, the intentions within the two disciplines are different. However, this alone is not enough to legitimise the ethnoarchaeological quest for the traditional. Following Barth’s early considerations ‘it (ethnoarchaeology) must be human in recognising the social and cultural construction of reality while yet seeking inter cultural translatability and universality. It must be concerned in its striving to transcend complacent tolerance and value-freedom to create deeper understanding of the human condition’ (Barth 1974:99). Thus, another crucial question raised through Barth’s (ibid.) efforts at making social anthropology a concerned human science, is whether studies of material culture or specifically ethnoarchaeology have any direct practical consequence for ‘science’ and our understanding of our world or not.

Africa as underdeveloped and especially as nondeveloping and a continent without history before the first arrival of the Europeans is a European made myth. This myth arose because of prejudices, lack of knowledge and economic or other political interests. The interior of the Dark Continent suffered from this and even after the ‘opening up’ of the interior this region was under the tyranny of the coast (Koponen 1988:43, Roberts 1968:v).9 Tambila has criticised the tendency, more prevalent in the past of dividing Africa into modern and traditional regions, and Ufipa has mostly been labelled ‘traditional and remote’ (Tambila 1981:7, cf. also Clark 1974:7), a notion from which Upangwa has also suffered (e.g. Kerr-Cross 1890:295). Most of Africa south of Sahara was scarcely known to Europeans before the late 14th century’s trade expeditions. In particular the interior of East Africa was unknown because the European expeditions that reached Africa at first kept their activities safely within the trade centres along the African coasts. But in spite of the fact that the early European tradesmen seldom left the coast they continuously made an impact on the interior and this affected the local trade systems, which goods were traded and how goods were valued.

1.3 The history of the region Vansina, among others, has discussed the concept traditional and the problems of homogeneity on connotations of the others as traditional in ethnographic accounts.8 He has specifically illuminated some of the problems in both writing and in the photography of ethnography that are equally relevant and problematic within ethnoarchaeology. According to Vansina ‘by implication traditional refers to an epoch without change, covering pre-colonial times far into dim antiquity’ (Vansina 1990:20). Observers often leave out of their accounts anything that refers to colonial practice or the modern and ‘this applies even to photographs, no bicycles, no kerosene lamps, no office buildings etc. And naturally ‘traditional’ clothing and housing was a must’ (ibid.: 21). Apart from more recent comments (i.e. Agrosah 1990; Andah 1995; Atherton 1983; Eggert 1977, Lane 1995, 1996; MacEachern 1996), ethics in ethnoarchaeology are rarely questioned. Few of the indigenous voices behind the description of the ‘traditional’ of indigenous people, such as their basketry, pottery or iron working that ethnoarchaeologists use are ever heard. Archaeologists working in Africa, as well as other poor, third world countries have, although not intentionally, conveyed that everything before kerosene lamps and plastic containers is traditional. We have tended to think that everything before what is generally believed to be due to colonial influence is traditional. Therefore pre-colonial is often mistakenly for and uncritically believed to be the same as authentic and traditional.

Before the European and other non-African (Arab) penetration of sub-Saharan Africa, several events disrupted and changed Africa. Our knowledge of these events has challenged the idea that Africa constituted a continent of peaceful farmers, herders and hunters. Some of these events were fierce and warlike disruptions such as the establishment of new kingdoms and royal courts. Nevertheless, there were other more peaceful migrations most likely connected to the development of iron producing communities. People moved and were able to settle in new areas, often after a period of labour on land and fields belonging to the settled (host) population. Equally important as the European intruders were local and long distance interior trade systems, rights to land and movement of specialists guided by local demands. Everywhere there was local exchange, often arranged in markets (Simensen 1997:208). Gray suggests the existence of trade between the Indian Ocean littoral and the lakes long before the colonial impact (Gray 1957). Koponen is more modest, claiming that trade was ‘nothing new among inland Africans themselves, only that they had a different form from the modern market concept’ (Koponen 1988:102). Despite the possibility that markets and the market as an institution were absent over most of Tanzania goods were traded and exchanged. Pre-colonial trading systems in the interior of East Africa have been divided into two categories, local trade and long-distance trade or international trade. The first, local

8 Although cultural/social anthropologist probably will not include Vansina among the well-known within the school of Critical Anthropology, he is known for his historical and linguistic research on Bantu-speaking peoples and it is therefore especially interesting to note his points on these matters in Paths in the rainforest (Vansina 1990).

9

In Tanzania before 1900 Roberts notes that the coast has been given much more attention than it deserves in studies of long distance trade (Roberts 1968).

8

Introduction trade was basically subsistence oriented trade and prestige trade while the other systems were non-subsistence or long-distance trade oriented. However, the two systems were intertwined in such a way that they cannot be discussed in isolation from one another (Mutoro 1998:186-187). Both the local and long-distance trade centred on transactions by barter in agriculture and domestic or wild animal products. Mutoro points to the fact that trade is among one of the universal human practices, therefore, it is incorrect and produces a false picture of the interior if pre-colonial trade systems and trade routes are underestimated or even excluded from the histories of the lakes region (Mutoro 1998:187). If iron working is considered, this commodity constituted an important part of trade together with cloth, pots and salt.10

territories into Tanganyika (Tanzania) and Northern Rhodesia (Zambia) or, as they were then called German East Africa and British Central Africa. The frontier between the two new territories cut across several of the tribal peoples in the area such as the Mambwe, the Lungu and the Nyamwanga (Willis 1966:xiv). When these new states had been established they controlled the interior while tribal wars and raiding continued. The Bemba of Zambia raided until 1895, as did the ruga ruga throughout large areas of East Africa.13 The late pre-colonial and early colonial history of contact between the interior and the coast marks the beginning of some of the most significant cultural and economic changes. These changes occurred among the numerous groups of peoples in the interior. Even the most remote areas were affected by internal turmoil, to some extent initiated by colonisation. The Ngoni invasion from southern Africa into Tanzania made an impact on both Fipa and Pangwa societies. The Fipa had to build palisades around their villages as protection against the forceful Ngoni intruders. Many peoples in the region did not leave their villages but kept their cattle and grain well behind palisades (Koponen 1988; Willis 1981). The Ngoni forced the Nyika into Ufipa (Poppelwell 1937:100), and Pangwa fled their villages and lived in caves in the Livingstone mountains during the Ngoni raids in their country (Iliffe 1979; Koponen 1988; Stirnimann 1976). Some Fipa men (and the Sukuma and Nyamwezi) went to work on the railroads and plantations. Others left because of the Maji-Maji disastrous rebellion. Although few in number, the Fipa and Pangwa were among the plantation – or railroad-workers and some of their villages were almost without men (Koponen 1994:637-640). In recent history the interior trade networks and plantation work along with the establishment of missionary stations has made a particular impact on the Pangwa.14 According to Stirnimann especially the 1930’s brought changes to the Pangwa who then experienced severe changes because of impact of economic and other cultural and religious impulses due to the establishment of missionary stations all over the region (Stirnimann 1976:263).

New social structures occurred because of migrations, local trade and exchange and because of long distance trade such as for instance the famous Nyamwezi trade (Iliffe 1979).11 For the Rukwa region Mapunda provides archaeological evidence for early trade networks confirming the early existence of a long distance trade (Mapunda 1995:66). In general this trade included goods such as ivory, iron and salt, and luxury objects of gold and copper (Iliffe 1979; Kjekshus 1976; Koponen 1988; Sutton 1990; Willis 1981; Wright 1980, 1982).12 The precolonial trading systems influenced pre-colonial and early colonial iron working traditions and the contacts that were founded between peoples of the Corridor region before colonial administration was established. It seems likely that the people of the Corridor area, of which Upangwa and Ufipa today form parts, contributed to networks of barter trade that criss-crossed the plateau and the surrounding lowlands. Relics of these trading systems have been maintained up until the present. Unfortunately, the existence of an internal and pre-colonial trade network in the interior of Africa before the European and Arab caravan routes were established is still not well documented. During the colonial 1800’s Europeans penetrated the African continent and established themselves as colonial administrators and missionaries in the Rukwa (Ufipa) and Ludewa (Upangwa) regions. Ufipa and Upangwa are located in the natural corridor linking eastern and central Africa, the interior and the coast, the area that later accommodated one of the major transcontinental trade routes of the late pre-colonial era (Willis 1981:196). Two major European states, Britain and Germany had influenced Eastern Africa for several years when in July 1890 they signed an agreement that demarcated the

In 1916, two years after the outbreak of the war between Britain and Germany, the German territory between Lakes Nyasa and Tanganyika (including Upangwa and Ufipa) came under British control and after 1919 the peoples were governed under ‘indirect rule’. In 1961 Tanganyika became a sovereign state. In 1962 two major administrative changes were initiated in Tanzania. Political officers were established at regional levels to oversee the working of national policy. And secondly, officers were appointed by local governmental authorities to take over the administrative functions formerly exercised by traditional chiefs and councils of chiefs. This

10 Fagan and Yellen have dated the occupation and production at the salt pans at Ivuna southeast of Ufipa from the early thirteenth century AD (Fagan and Yellen 1968). 11 Changes have not only occurred within this area but of course in Tanzania as a whole before, during and after colonisation (Iliffe 1979:8; Koponen 1988:390). 12 Koponen suggests that salt, like iron, was a prestigious commodity and commercial salt procured over long distances was accorded a status akin to other trade goods (Koponen 1988:109-110.fn.101).

13

Ruga ruga were young adventurous men, often criminals, who were armed by the local chiefs. 14 Cf. appendix C for historical events in the study area.

9

Randi Barndon system has remained up until the present.

hoes and cash replaced the locally produced ‘wedding hoes’.

Another aspect, similar to western attitudes towards precolonial contacts between the peoples in the interior and the coast is seen in the invisible problem. Many of the interior civilisations and towns, with kingdoms and royal courts, suffered from this problem and still do, although they most certainly existed (Connah 1989:214-215). Within this region of kingdoms and chiefdoms, iron working and more generally craft production was strong and usually connected to intra- and even extra-regional exchange (Wright 1980:2). The kings claimed supernatural and metallurgical knowledge; they were ‘king-smiths’ (Maret 1985). Their kingdoms were founded on their power and wealth as it was expressed not only in their regalia but also in their cattle, their fields, their wives and their claims for knowledge of iron working. In the interior of Africa the king or chief claimed to be a master smelter, or a master blacksmith, a Roi-Forgeron (Maret 1981; 1985). In Bantu- and Swahili-speaking Africa iron was (and still is in its various local forms) called chuma, meaning ‘a thing of value’ (Maret and Nsuka 1977; Phillipson 1988; Wainwrigth 1954).15

Ever since the first maritime contacts between the interior regions of Africa and the Western world were established at the end of the fifteenth century, numerous European observers have been intrigued by African metallurgy (Killick 1990:3). The flow of reports has never stopped although they decreased after independence was established throughout sub-Saharan Africa perhaps mostly due to the increasing availability of imported iron objects and scrap metal and the fact that there were fewer colonial administrators in the country. While this is true, Mapunda notes that in the most remote areas the European influences were minimal and local indigenous iron working continued until the late 1930’s and even 1950’s (Mapunda 1995:96, see also Barndon 1992, 1996b; Grieg 1937; Schmidt 1996b; Wembah-Rashid 1969; Willis 1981; Wright 1982). 1.4 The people in the region Two well known anthropologists have worked within the region. In the 1950’s Monica Wilson, initially with her husband Godfrey Wilson, worked among the Nyakyusa who are close neighbours of the Pangwa (Wilson 1958, 1959, 1977) and Wilson’s study of Nyakyusa ritual life is still important and influential.

Iron was not only a material of prestige for kings or chiefs, men of power; it was a matter of local community importance. It was important in agriculture, in hunting and when establishing local social relations such as marriage bonds. The concept wedding-hoes indicates the value of iron in societies where men had two or more wives, as was the case among many of the peoples in the Corridor (Culwick and Culwick 1935; Stirnimann 1979). The significance of hoes must not be underestimated in the societies of the interior. I believe that accepting iron as ‘a thing of value’, not only economically, but in many integrated social and ideological contexts, is essential for a broader understanding of African metallurgy and in particular iron smelting traditions.

Roy Willis did his first fieldwork among the Fipa in the early 60’s and has continued research based on his knowledge of the Fipa until the present (e.g. Willis 1964, 1966, 1981, 1991 and see bibliography). Willis’ early structural model of Fipa cosmology has proved in many was to be of equal relevance today as it was in the early 70’s. It is relevant in terms of their iron-smelting practice, to which I will return in Chapters 6 and 7 (cf. also Barndon 1992; Childs 1993).

During the colonial period old crafts such as iron smelting gradually diminished. Smelting was eventually completely abandoned while blacksmiths continued their work using scrap metal. Railways effected many spheres of life; they ruined the Zinza iron smelters who had produced iron for hoes which caravans used as currency (Iliffe 1979:137). Lower transport costs allowed European manufacturers to penetrate new areas. Therefore, by the late 1920’s domestic weaving and iron smelting were increasingly rare. Because of this the blacksmiths’ repertoires of tool making diminished to minor repairs of iron tools and the forging of odd wedding hoes. No longer did blacksmiths forge the king’s regalia, the magician’s bracelets or dancing bells for ceremonies or any other objects of prestige. Imported

Hans Stirnimann is the only ethnographer who has ever worked among the Pangwa in the Livingstone Mountains (Stirnimann 1967, 1969, 1976, 1979, 1983).16 Stirnimann, like Willis, conducted his fieldwork in the early 60’s. Wilson, Willis and Stirnimann all made brief but valuable comments on the corridor peoples’ iron working although none of them have observed the entire process (Willis 1966, 1981; Wilson 1958, 1959; Stirnimann 1979). Their comments and descriptions of iron working were familiar to me before I went into the field and will be commented upon and compared with my own observations, description and discussion of Pangwa and Fipa iron smelting. Numerous other researchers have visited Ufipa, both before and after my fieldwork and commented upon Fipa iron working (i.e. Grieg 1937; Mapunda 1995; Sutton 1985; Tambila 1981; Wembah-Rashid 1969; Wise

15

16

The Bantu form –uma for iron was previously postulated to be a core word associated with the early spread of Bantu metallurgists (i.e. Wainwright 1954), while –uma may be a root word entering through later Arabic contacts (Maret and Nsuka 1977).

The Pangwa in the Livingstone mountains must not be mistaken with the Pangwe of West Africa studied by Tessmann in 1913 who are confusingly accounted for in Cline’s publication about African metallurgy (Cline 1937).

10

Introduction RUNDI HA Ujiji

Tabora Zanzibar

Tanganyika

LUBA

Dar es Salam

HE HE NYAKYUSA

IP

Mweru LUNGU

A

Lake

Rukwa

F

TABWA

Lake

Kilwa

KINGA LUNDA Lake

PAN G W

BEMBA PHOKA Bangweulu CHEWA Lake

KISI NG

A

IINDIAN OCEAN

NYAMWEZI Lake

ON

I

Malawi

N

Tegn.E.Hoff

Figure 1.4 Groups of people referred to in the text

1958a, 1958b; Wrigth 1980, 1981, see also Appendix D, table1).

From Wilson’s perspective the Fipa were rather distant to ‘her’ Nyakyusa-Ngonde cluster. Wilson included among the Corridor People the people of Mwika and the Tumbuka-Tonga group, the Tumbuka and Tonga group being represented along the western side of Lake Malawi (Wilson 1958:19, 64). The Mwika were according to Wilson a linguistically and culturally united group that constituted the Fipa, Pimbwe, Rungwa, Mambwe, Iwa, Lungu, Wanda and Namwanga (ibid.). The Mwika group is not mentioned or discussed in any of Willis’ publications and his groupings are slightly different (e.g. 1964). The association he makes of groups of peoples alters according to which aspects are emphasised, such as social organisation (e.g. Willis 1964, 1966, 1981) or cosmology and the use of metaphors (e.g. Willis 1991). In his initial studies Willis associated the Fipa with the Lungu, the Mambwe, the Pimbwe, the Wanda, the Lungwa (Rungwa), the Kuulwe18, the Nyamwanga and the Cile (Willis 1966). Kifipa, being a Bantu language, Willis states to be similar to Mambwe, Pimbwe, Rungwa, and Lungu languages (Willis 1981). In his earliest publications Willis argued for an association between these peoples into a separate group within the region because they all share a common ‘sub-culture’, that of the agnatically organised village community. They speak languages that are largely mutually comprehensible. And finally, as regards descent, they are distinguished from the ‘matrilineal belt’ peoples to the south by possessing systems which were either patrinlineal or modifications of patrilineality incorporating the notion of general alternance (Willis 1966:xi).

The best description of the area I have ever come across stems from Monica Wilson who wrote about the region, and in her words: ‘The mountainous country between the Great Lakes, Tanganyika and Nyasa forms the main corridor between East and Central Africa through which migrants, travelling with stock from East Africa to the grasslands of the south, probably passed. The vast plateau of Central Africa is broken here by four peaks - Rungwe, Mtorwi, Chaluhansi, and Mbeya - which reach over 9000 feet, and by throughs in which the lakes Tanganyika, Nyasa and Rukwa lie. It is the watershed between the Congo, flowing in to the Atlantic, the Zambesi flowing south and the Rufiji east into the Indian Ocean, and smaller streams which flow north into Lake Rukwa. From Rukwa there is no outlet, but less than 300 miles further north there is another watershed from which streams flow into the Nile. This wild belt of country, stretching from high and healthy uplands to the stifling lake shore plains, therefore lies in the very heart of Africa’ (Wilson 1958:1). Wilson’s introduction to the Corridor Region quoted above certainly captures how the area which includes Upangwa and Ufipa has been a geographically important region, a region that over time has accommodated, as it does today, several peoples differing in language and custom. Wilson’s very heart of Africa has later been labelled the interior mosaic referring to the ‘culturalethnic’ and historical complexity of the region (Newman 1995:159). Through time this mosaic of cultural and historical complexity in the interior has changed several times.17 17

Tanzanian societies. Seen in retrospect pre-colonial societies were not stable unchanging structures and we have no clear cut understanding of pre-colonial or colonial ethnicity structures (Koponen 1988). 18 Mkulwe in Iliffe (1979).

Koponen points at a crucial problem concerning pre-colonial

11

Randi Barndon The true origin of the Fipa people is still questioned, as is the origin of the Pangwa and most of the peoples of the interior. Iliffe suggests that considerable influence and exchange has occurred from the western side of the lake and that western Tanzania was strongly influences by the kingdoms in the west (Iliffe 1979:24). Although the chiefs in some legends among the Fipa were believed to have come from the north other traits emphasise a more direct connection with Zambia. Willis suggests Fipa origin to the regions south and southwest of the Ufipa plateau and proposes that the present Fipa are truly connected to their myths that account for how the first Fipa came from the marsh near Mweru Wantipa (Lake of Mud) in northern Zambia (Willis 1968:82-83). This area seems relevant in the light of their iron smelting practices.

Malawian Sukwa, Wandya, Lambya, Tambo, and Phoka and Chewa peoples. She believed in a strong association between the Pangwa, the Kinga and the Matengo, while excluding the immediate neighbours of the Pangwa, the Kisi (Tew 1950:71).20 The Pangwa language is closely related to the languages along the Zambian border (Feldberg 1999; Murdock 1959). Most likely the Pangwa origin is also connected to Zambia and by tradition they are an ‘offshoot’ of the same group as the Kinga in the Livingstone Mountains. From the Kinga group or cluster, the Iwa or Iba and the Nyamwanga/Miamwagna or Niamwanga chiefs were also said to derive. Linguistically the Pangwa have been associated with the Hehe-Bena group and according to Guthrie classified as a G-60 language together with Hehe, Bena, Kinga, Sango, Wanji, Kisi and Matumbi (Guthrie 1967 in Stirnimann 1983).

Many similarities in terms of furnace morphology, lexical terms and ritual practices are shared among the Fipa and peoples towards the west and southwest of the Ufipa plateau (Barndon 1992). The arrival of the first settlers in Ufipa can at the moment only roughly be placed between the wide limits of AD 1500 to AD 1700 (Willis 1981:14). Mapunda’s archaeological research in the Lake Tanganyika region supports the ‘sixteenth century‘ date (Mapunda 1995:56.fn.16). Thus, the route of contact if iron working is in focus may indicate a late transmission of iron working knowledge from peoples along the western and southern shores of Lake Tanganyika and into Ufipa.19 Later immigrants might well have originated from the north and influenced Fipa concepts of iron working, however, the technological practice, as observed in my re-enactments of smelting, does not originate from the north.

Wilson related the Pangwa both to the Bena and the Kinga but she associated them with the smaller groups of people in the region, such as the Safwa, Wanji, Kinga, Kisi and Bungu (Wilson 1958:40-46). The Nyakyusa claim that the Kinga provided them with their chiefs (Wilson 1958:11). The Pangwa are liguistically associated with Tumbuka, Tonga-Sisya, Safwa, Wanji, Kinga, Kisi and Bungu (Wungu) (ibid.: 65).21 From Stirnimann’s point of view the Pangwa are seen as surrounded by and related to the Matengo and Ngoni to the south, the Kinga, Wemba, Bena, Wanji and Nyakyusa to the north and even further north the Sangu and Hehe (Stirnimann 1976:13). The Kinga are connected to the fragmentary memories that Pangwa hold about their origin (ibid.: 16). Pangwa legends tell of how strangers came to Upangwa bringing with them the knowledge of fire making, cooking and iron working. Before that people in Upangwa had eaten their food raw and been cold in the winter (ibid.: 17-19). It is generally believed among the Pangwa that the Kinga were the truly skilful ironworkers of the region. The Kinga’s first famous master smelter was called Mlanga (ibid.: 17). From a view that emphasises iron smelting, the Pangwa are associated with the Kinga who were influential on cosmological and political matters for the Nyakyusa, Pangwa and Bena.

Thomson introduced the Pangwa, as a name given a larger group of Bantu-speaking peoples constituted by refugees from Ubena and Ukinga (Thomson 1881:247251). Describing Upangwa Fülleborn wrote ‘Die Südhälfte des Livingstone-gebirges wird von dem volkreichen Stamm der Wapangwa bewohnt. Die Wapangwa-Ansiedlungen beginnen südlich von KilondoBach und reichen nach der verbleibede allersüdlichste Abschnitt des Livingstone-Gebirges’ (Fülleborn 1906:439), setting their country on the geographical map. Several decades later, Tew notes that on the eastern side of lake Nyasa ‘the Wanji and Poroto north of the Kinga, the Pangwa and Matengo on the south and the smaller Kisi, Sandida and Mpoto appear as names on the maps, but little else is known of them’ (Tew 1950:72). At the time of Tew’s research very little information had been documented about the Pangwa and she took them together with the Ngonde and Nyamwanga people studied by Wilson (Tew 1950:72; Wilson 1958). She suggested that they were most strongly related to the Kinga and Matengo peoples (Tew 1950:71). Tew saw the area as the North Nyasa Region and therefore from another geographical point than Wilson. She did not exclude peoples from other modern nations within the region such as the 19

In my discussion with the Pangwa and Fipa they claimed both diversity from their neighbours and a unity with them. Many of my Pangwa informants claimed they were Bena or as the Bena because they spoke the same language. Many of my Fipa informants claimed they were Lungu or Mambwe for the same reasons. At the same time I was confronted by material objects or traditions, rituals or ceremonies that they claimed were typical Fipa or typical Pangwa, as they claimed for their iron smelting 20

The Kisi did not practice iron smelting but were skilled potters. The Tumbuka, Phoka and Chewa are their neighbours to the west on the other side of Lake Nyasa and are also related to them in terms of iron working. 21

Cf. Barndon 1992; Mapunda 1995; Willis 1981.

12

Introduction ethnographers’, nonetheless contributed to a specific image of the relations between the peoples of the interior. This image was most likely never the one the indigenous peoples themselves held. This has created a false picture (Koponen 1988:43).

practices although they share these traditions and technologies with their neighbours. In the past the Pangwa and Fipa were probably in exchange relations with numerous peoples in the region. Stirnimann reports that the Pangwa obtained their wedding hoes from Manga but salt from the Bena and their pottery from the Kisi (Stirnimann 1979:12), while less is known about Fipa trade relations beyond Ufipa.

My concern is, as I will argue, not primarily an empirical one, it is mainly a conceptual one. It is not merely a question of what empirical research may show to be the case (as perhaps my re-enactments still may demonstrate) but of ‘what analysis reveals about what it make sense to say’.22 Our concepts will provide us with a better comprehension of the objects we study. In this study, I therefore define and discuss my conceptual approach. Through what I find reasonable to say and specifically how I say it, I hope to demonstrate the relevance of a conceptual shift in studies of technology and in particularly recent African iron working. One can only hope that the written text is not discriminating or to be used in any dangerous political rhetoric (Lane 1998b:181; Spiegel 1994).

It is important to observe the focus of the researcher when ethnic issues are questioned. Koponen and more recently Kriger (Kriger 1999:27n.20; Koponen 1988) suggest that we should ‘de-ethnicise’ the mosaic of peoples in the interior of Africa. Kriger suggests that when master blacksmiths are our focus we must acknowledge that they were able to cross ethnic or social boundaries because of their skills as blacksmiths and the prestigious craft they mastered. Koponen, from a different point of view suggests that early European travellers and ethnographers believed that Africans belonged to ‘tribes’ led by ‘chiefs’. This he takes as the mental framework under which they perceived African countries and peoples and took down notes, which they bequeathed to us as our ethnographic and historical sources. As a result almost all our data are couched in ethnic and more specifically tribal terms (Koponen 1988:42). But as he continues, they were in this respect misconceived, basic social units in pre-colonial Tanzania were not based on ethnicity, tribes or ethnic groups (ibid.: 42). Still, there is no general agreement to what these basic units were. Koponen argues that the basic units were determined by political and economic criteria that he calls societies. Within societies social relations were functioning and reciprocal social action took place (ibid.). As Koponen sees it pre-colonial ethnicity (if at all a valid term) was fluid and constantly changing and many of the societies we think of as ethnic groups were probably more categories created by outsiders.

It is anticipated that continuing ethnoarchaeological research with its specific focus on material culture will improve our comprehension of our world, its prehistory (ies), history, recent past and present. Through a constant refining of our concepts this is possible while ‘how we carve up our world will also depend on what is out there independent of us, and equally on the referential scheme we bring to bear, given our purposes, interests and goals’ (Johnson 1987:202). Latour (1999) and Tilley (1998) convincingly demonstrate that material culture is an important aspect of life that science cannot capture and enclose within disciplines or university departments. Studies of the material culture among contemporary people anywhere, as among ourselves, in its many forms, can inform us about the significance of materiality and how we construct our everyday lives and how we cope with the world through this materiality. Simultaneously, I believe that studies of contemporary material culture with an ethnoarchaeological perspective may inform about what it was like living in the past, if the objects or technologies we study were also a part of the past. But more than anything, studies of material culture among living people informs us about how they organise and structure their lives, what they are concerned with and which values they hold. It must be emphasised that conveying Pangwa and Fipa concepts of iron technology and writing about parts of their lives must not uncritically be projected upon any or anyone’s past. I am particularly concerned with respect and admiration for those people who generously demonstrated their iron smelting skills to me in the mid 1990’s and included me in their way of comprehending technology and how to practice it.

1.5 Summary In writing the ethnographic history of the region it is important to be aware of how those who first met the people living there and who were also the first who commented upon them, reported ‘home’ and described their hosts’ ways of living. Their focuses, points of departure and settings have been the main factors in the making of ethnic diversity of the peoples of the interior. This has created and maintained some peoples as more powerful, famous or central than others. Hence, ethnographers have contributed to the establishment of a focus and a relation between the various peoples because of their objects of research and their own personal point of view; their way of relating all other peoples as their friendly neighbours or enemies. For instance Audrey Richard’s Bemba (1958), Victor Turner’s Ndembu (1967) and Monica Wilson’s Nyakyusa (1958) are such ‘famous’ peoples. My point is that while these studies are excellent, they have, although this was not the intention of the

22

13

The formulation is taken from Winch 1958:71-72.

Randi Barndon

14

Multiple worlds and multiple realities

2. Multiple worlds and multiple realities The past is too complex to be rendered in all its richness in one book, but to refuse a clear choice means the fabric will remain on the loom forever (Vansina 1990:xii). Theories or philosophical programmes are no longer merely available as something we can apply as our ultimate explanation for the various phenomena we want to understand or describe. Some form of acceptance of objectivism has existed in the mainstream Western philosophical world for at least the last three- and a half centuries and has formed ‘the Descartes-Locke-Kant tradition’ (Johnson 1987; Mitcham 1979; Rorty 1979). But Kuhn has made us all aware of how science as a social, cultural and historical practice and knowledge is always situated (Kuhn 1962; after Lakoff and Johnson 1999:89). It is our being situated in relation to things, our embodied experiences, that assures us we are realists argues Johnson (Johnson 1987). However, we need not fear that we have lost our world if we also acknowledge the theory-impregnated nature of science. This does not mean that we can or should try to reject our philosophical tradition and start again from scratch, something that is almost impossible to imagine (Latour 1999).

2.1 Multiple worlds and realities Vansina’s above quoted metaphor points at a problem that certainly has been experienced within the postmodern era of philosophy and which has affected archaeology. While ‘making a clear choice’ in terms of what approach and what theories to apply is not something that will complete the ‘fabric’ at all, I believe our concepts and attitudes towards our data may refine our approaches and hopefully lead us into new directions of research. Schutz has made the important point that the activity of science is a circumscribed activity, undertaken in very specific and restricted circumstances by partial selves of human beings, and that, therefore, this is a special ordering of reality, only one of several others (Schutz 1962; Tambiah 1993[1990]:103). The world as constituted of multiple realities was Schutz’s thesis and ‘the reality of the world of the daily life’ was as the centrepiece. The familiar is in our everyday worlds (cf. Tambiah 1993:102-104). The world of daily life Schutz described as an intersubjective world common to all of us in which we have not a theoretical but an eminently practical interest (ibid.: 102). Schutz treated the other realities as provinces situated around it, which one could step in and out of.

Philosophers within neo-pragmatism such as Bernstein and Rorty, grant the existence of an external world (Bernstein 1998; Johnson 1987; Lakoff and Johnson 1999; and cf. Olsen 1997:17-19). The existence of the external world assures us that there are things in the world that exist independently, there are facts, a certain physicality and uniformity that we can take as our point of departure. Lévi-Strauss exemplified his views on ‘science’ in relation to what he called the ‘world of smells and tastes’ pointing at how we have been accustomed to think that taste was entirely subjective. However, ‘now chemists are able to tell us that each smell or each taste has a certain chemical composition and they can give us the reason why subjectively some smells or tastes feel similar to us and some others widely different’ (Lévi-Strauss 1980:6-7). The chemical compositions of taste is merely one side of the story, what food we like or dislike is another. While as Lakoff and Johnson put it ‘there are stable scientific results; now that we have photographs of the earth from the moon, any lingering doubts that the earth is round has been removed’ (Lakoff and Johnson 1999:98). Similarly ‘we are not likely to discover that there is no such thing as cells or DNA’ (ibid.: 89). Perhaps this is forcing my argument too far since my point is that this view on science does not contradict the value of subjectivity and ‘an embodied thinking’ with its roots back to Merleau-Ponty and Heidegger (Dreyfus 1996; Heidegger 1972; Merleau-Ponty 1989[1962]; see also Tilley 1994:13).

My study is about people and furnaces and about technology and metaphorical imagination. This, however, leads to discussions of the long lasting dualism in science between mind and body, subject and object and humans and non-humans. Latour suggests that we should think of actors and artefacts as actants in order to prevent the dualism between humans and non-humans (objects, technologies) in what he sees as collectives (of subjects and objects) instead of societies as consisting of only humans (Latour 1999:194-196). In line with this view I hold a critical view upon the way technology as a concept has been used in archaeological and anthropological research and how it has been disconnected from more holistic frames of understanding the relation between humans and material culture. Hence, our concepts have ignored individual and communal forms of experience as part and parcel of technological agency and therefore we have missed how ‘making meaning’ was part of making iron. For this reason the following comparative study of Pangwa and Fipa iron working is meant as a contribution to our understanding of symbolic forms (verbal and material metaphors) and its various individual and social levels within a technological practice.

Objects or spaces and landscapes might through an impressive physicality form our cosmologies as do the difference between seasons, day and night, sun and moon, 15

Randi Barndon western Tanzania.

female and male. Hence, one may say that ‘objects’ outside of us talk back to us and proclaim their presence with a very loud voice most of the time (Johnson 1987:204). Because of this we can accept that there is a world independent of our understanding of it that we may even achieve stable knowledge of (Lakoff and Johnson 1999:90). A framework for science, as somewhere between an ‘absolutism’ and ‘relativism’ requires a theory of reference that can explain how language, in particular scientific language, can more accurately map ‘reality’ (Johnson 1987:200 and cf. Johnson 1993; Tambiah 1993).

Empirically observed perceptions of phenomena (i.e. gender or cosmological matters) as binary oppositions in symbolic systems (codic oppositions) have commonly been depicted in anthropological studies of cosmology in Central and East Africa (e.g. Heusch 1956, 1975, 1980; Willis 1967, 1981). Most archaeological studies of African metallurgy concerned with East African cosmology and Bantu modes of thought have concluded that categorisations and oppositions between categories were central (i.e. Barndon 1992; Cline 1937; Collett 1985; Herbert 1993; Kense 1983; Schmidt 1978, 1997). This work will deal with dichotomies and classifications that people make precisely because of a belief in metaphorical thinking as grounded in a personal embodied experience that is projected upon the collective. I will attempt to illustrate how and why metaphors were used within the Pangwa and Fipa iron working practices. The smelters did not only use metaphors; they also used metonyms and multiple combinations of oppositions of conceptual and material pairs and dichotomies.

Most of us now agree that we should reject the false (scientific) dichotomy between the two opposite and incompatible options of either absolutism or relativism (Johnson 1987:196; Trigger 1995:456). Does this imply that it is time for analyses that take into account both sides and which with equal seriousness include a certain uniformity as well as the specific or historical particular? I believe so, and also that it is possible through an embodied philosophy. Without a fear of being labelled objectivists we must acknowledge the progress that has been demonstrated within science while remembering that not everything can be translated into a scientific language. The fear of never arriving at an ultimate truth, a fear that Geertz has made us aware of in his Distinguished Lecture: Anti-Anti Relativism (1984) does not need to be anything but a challenge. We can of course no longer simply apply a classical objectivist view of knowledge, one that assumes that ‘science’ produces successive theories that approach closer and closer to the correct description of reality (Bekaert 1998; Johnson 1987:xiii; Latour 1999). We cannot fully implement relativism either. Not because in its extreme forms it advocates various ‘post-structural anything goes attitudes’ (Trigger 1989:406),23 but because the doctrine of the psychic unity of mankind (or human universals) and the doctrine of diversity of cultures/societies are no longer contradictory dogmas (Tambiah 1993:112).

These dualities do not necessarily and always include a polarity or codic opposition wherein one of the components is valued over the other. As I will discuss they were essential or powerful through their combination, hence more like twins (Ani 1996:77; and cf. my discussion on gender in Chapter 7). Oppositions such as between male and female, day and night or hot and cold were never evoked or taken into use without the social actors who give the categories a meaning (Moore 1995:73). The smelters used a structural system as part of the constituted symbolic repertoire (Bekaert 1998). This symbolic repertoire was active at various levels from individuals to the collective and social sharing of a symbolic reservoir (Sterner 1992). The symbolic reservoir was based on deeply experienced bodily imagination. Taking account of codic oppositions provides a logic and is seen as an integrated part of the local model of the smelting process.

2.2 Embodiment, between phenomenology and structuralism?

According to Tilley the method of Lévi-Strauss as for instance described in Totemism (1963[1962]: 16), can be simplified to the following:

Within African metal production there have been two major areas of research. On the one hand researchers have been occupied with chronology, technology and metallurgy while on the other hand environmental, economic and social and cultural aspects have been the main interest. My focus is rather in between the two domains. I am concerned with the relationship between metallurgy and metaphors in iron making and the unity within diversity of smelting practices primarily between two groups of people, the Pangwa and the Fipa in south-

(1) Define the phenomenon under study as a relation between two or more terms, real or supposed; (2) Construct a table of possible permutations between these terms; (3) Take this table as the general object of analysis, which, at this level only, can yield necessary connections, the empirical phenomenon, considered at the beginning being only one possible combination among others, the complete system of which must be reconstructed beforehand (Tilley 1991:21-22).

23 I have attempted to include post-structural analytical levels in my text when considering the actual field practices and writing of ethnoarchaeology. This post-structural level in my work has no direct relation to ‘the common pragmatic use that archaeologists commonly attach to theory’ (Olsen 1997: 281).

Lévi-Strauss firmly believed that meaning was structured in terms of binary oppositions and explained his thesis by

16

Multiple worlds and multiple realities oversimplification. Today, few, if any, accept the entire Lévi-Straussian structuralistic programme with its search for an objective truth (e.g. Lévi-Strauss 1962). LéviStrauss’ linkage to linguistic theory on the treatment of binary oppositions and parts of his structural principle is still appreciated, often as a foundation for further explanation (cf. Tilley 1990, 1999). Nevertheless, I think that in order to systemise in a creative manner we have to separate and divide, to translate and reduce. Perhaps dichotomies are indeed necessary for the anthropological enterprise or even in all science.26 If so, we should keep two critical points in mind. First, the modes or structures are not identical with the social world but are mere aids in organising fact from the social world. Secondly, dichotomies may be envisaged as scales marked by difference in degree rather than as absolute contrasts (Eriksen 1995:289).

the contrasts people make between cultural categories such as ‘nature’ and ‘culture’, ‘men’ and ‘women’ or ‘raw’ and ‘cooked’. Olsen explains the theory of structuralism clearly and concisely: ‘The focus of structuralism is not the surface of social life as experienced by the members of a society, rather structuralism seeks behind the chaos of rules and customs in order to discover a structure or structural ‘schemata’ that operates independent of time-space dimensions’ (Olsen 1997:197). Olsen defines structuralism to be ‘first of all a theory in which cultural meaning is produced through the differences and contrasts between signs. The most meaningful differences are those that can be classified as binary oppositions and observed cultural phenomena are products of an underlying logic or structure’ (ibid.).24

As early as in the late 60’s Barthes and Derrida published their scepticism towards the structuralism that they initially had followed (cf. Olsen 1990:164-166; Schmidt 1985).27 The most influential implementation of poststructuralism within archaeology has been Moore’s study among the Endo in Kenya (Moore 1986). In a study of Endo organisation of space as related to gender she illustrated how the then new theoretical perspectives in social anthropology could be applied within contemporary material culture studies and be of relevance in archaeology. Moore focused on houses, fireplaces, cooking, treatment of dirt and disposals and relations with space and gender that made Endo processing of their material world into meaningful structures and guidelines in their every day lives. As pointed out by Moore in her use of Bourdieu (1977, 1993) and Ricoeur (1978): ‘as a result a number of writers had observed that the meaning of symbols was given by their operationalisation in different contexts, rather than merely from their position in an abstract system of differences’ (Moore 1986:3).

Saussure introduced the crucial distinction between language and speech; a language which is collectively shared while speech is the individual act in which the codes provided for in the language are utilised to express something (Tilley 1991:17-18). Saussure’s work included the well known division between parole, the utterances of individual speakers and langue, the underlying system of language, making possible a particular speech act (Olsen 1990:183; Tilley 1990:6, 1991:17).25 To understand a language involves not only an understanding of the individual utterances but also of the underlying principle (Olsen 1990:13). As Lévi-Strauss argued, in order to understand cultures, myths or legends one needs again to understand the underlying principle. The structure in language consists of series of units or sign with two sides or faces the signifier (a sound image) and the signified (the concept or object) (cf. Hodder 1989[1986]: 47; Tilley 1990:6). It is the structure of the system which allows signifiers and signified to posses signification or meaning (Tilley 1990:7). Within structuralism a sign obtains its meaning by its position within an abstract and internally structured system of codes of presence and absences, similarities and differences (Preucel and Hodder 1996:299).

We cannot deny the creativity in Lévi-Strauss’s ‘semiology of anthropology’ in his reduction of mythologies, belief-systems and totemic clans (Tilley 1990). As stated by Tilley ‘the abiding significance of the work of Lévi-Strauss for the study of material culture is not the trappings of a structuralistic mode of analysis isolating binary oppositions, but that he is a master of metaphor’ (Tilley 1999:272).

Lévi-Strauss was greatly inspired by Saussure’s semiology, a semiology that included linguistics within many disciplines (see Olsen 1990:172). Lévi-Strauss’ structuralism does not seek to reveal similarities in actual reasoning, but universal underlying principles for thought and symbolisation. As such the fundamental cognitive processes among people regardless of time and space are seen as identical (Eriksen 1995 [1988]:226; Lévi-Strauss 1963:27-28, 1966[1962]: 75).

The distinction made between structuralists such as LéviStrauss and the phenomenology of Merleau-Ponty and his use of Saussure is based on the overall assumption within early structuralism that we can establish a complete objectivism and truth (Schmidt 1985:160). Bourdieu’s

One of the constant dilemmas in trying to describe other mental worlds or cosmologies is the danger of

26

In Latour’s words: ‘At the card table with so many trumps in hand, every scientist becomes a structuralist’ (Latour 1999:38). 27 As noted by Olsen ‘it is common to divide the work of Roland Barthes into an ‘early’ (and structuralistic) phase and a ‘late’ (poststructuralistic) phase (Olsen 1990:164). Barthes therefore rejected his ‘early dream of scientificity’ and became more influenced by Kristeva and Derrida in his post-structuralistic textuality’ (ibid.: 165).

24

Translation by the author. cf. Olsen (1990) and Tilley (1990) for references on Saussure’s semiology. 25

17

Randi Barndon Outline of a theory of Practice (1977) and Giddens’ New rules of Sociological Method (1976) and Central Problems in Social Theory (1979) led to the emergence of anthropologically and sociologically oriented critiques of the objectivism embedded within Lévi-Strauss’ structuralism.28 It was the stress on the material and embodied dimension that separated Bourdieu from LéviStrauss (Bourdieu 1977, 1979:220; 1993; Strathern 1996:26), while at an earlier stage this had separated Lévi-Strauss’ views from Merleau-Ponty’s (Schmidt 1985). Merleau-Ponty argued that philosophy is always a break with objectivism and a return from constructa to lived experience, from the world to our selves (Schmidt 1985:52). I believe that what matters is to be aware of and try to account for the multiplicity of lived experience. Even if we accept that there is not just one communal form of experience, but many emerging from a baseline, we may contribute with new data, knowledge and analytical results. This was Merleau-Ponty’s initial intention when he tried to conceive a relationship between phenomenology and the results of empirical enquiries (ibid.: 56).

For Merleau-Ponty human action at its most basic level is motivated by a need to get a grip on the world (Dreyfus 1996:12.fn.1). Merleau-Ponty argued that we are both inhabiting and embodying our social worlds and that our being-in-the-world is rigorously grounded in bodily experiences (Merleau-Ponty 1989). Thus, for MerleauPonty ‘to be a body is to be tied to a certain world…our body is not primarily in space, it is of it’ (ibid.: 148). Space then is a set of orientations while the anchor is the body by which we inhabit space (Weiss 1996:6). The body is according to Merleau-Ponty ‘a meeting place of past, present and future because it is the carrying forward of the past in the outlining of a future and the living of this bodily momentum as actual person’ (Langer 1989:32). One of Merleau-Ponty’s valuable contributions was to show the influence of spatio-temporal factors on perception, through the concepts of perspective, field and horizon (Merleau-Ponty 1989; Strathern 1996). His aim was to remove the concept of the body as an object and the mechanistic physiology that goes with such a concept and to reintroduce the body as ‘our point of view upon the world’ (Langer 1989:25).

Merleau-Ponty’s use of Saussure was mainly to focus on how subjects polarise a set of constituted signs in ways that enable them to say something new. What MerleauPonty wanted was to show how conventions could produce something new, how an infinite number of signs could be put to an infinite number of uses (Schmidt 1985:162). Bourdieu following Merleau-Ponty has argued that the representation of mental categories is embodied in two ways; by the material constructions people make (e.g. houses, fireplaces or as in my case smelting locations), and by positioning their own human bodies in relations to these constructions (Bourdieu 1973, 1995; Strathern 1996). Accounting for Schmidt’s (1985) reading of Merleau-Ponty’s Perception Bekaert provides a new approach to the study of iron working symbolism in the African context. Bekaert, like Schmidt, sees MerleauPonty’s phenomenology as somewhere between phenomenology and the later developed Lévi-Straussian structuralism (Bekaert 1998; Schmidt 1985:35-39, 5052).

How we interpret cultural expressions or symbolism is a major problematic issue not only within archaeology but perhaps even more so within anthropology (Bekaert 1998:6). Bekaert’s suggested combination of phenomenology and structuralism provides a holistic approach towards how and when symbolism, in its various forms or multiple levels, is at work in the smelting process. Bekaert convincingly argues for an interpretation of the sexual symbolism in Sakata smithing inspired by Lévi-Strauss (1962; 1966), Schutz (1962) Geertz (1976) and Merleau-Ponty (1989). He states that he finds both the phenomenological ‘lived experience’ and structural ‘meaning’ to be concepts that have empirical possibilities (Bekaert 1998:16). Within a phenomenological frame meaning is the result of a process that forms units of experiences, whereas within structuralism meaning is created by the tearing apart of these same units into negatively opposed parts; the binary oppositions. As argued by Bekaert both ways of making meaning and rendering cultural expressions meaningful are relevant (ibid.).

The preface to the Phenomenology of Perceptions opens with the question ‘What is Phenomenology?’ (MerleauPonty 1989:vii). Merleau-Ponty concludes that it is ‘the study of essences’ (ibid.), perhaps not too different from Latour who argues for collectives of humans and nonhumans instead of societies and objects. In Langer’s view: ‘Phenomenology is a transcendental philosophy that brackets the claims springing from the ‘natural attitude’ to the world; yet it recognises the pre-reflective presence of the world and devotes its whole efforts to re-establishing a ‘primitive contact’ with it and directly describing our actual experience’ (Langer 1989:163).

28

In some ways our concepts are based on both a bodily experience and our ‘savage mind’ (Bekaert 1998; LéviStrauss 1994[1966]; Merleau-Ponty 1989). In other words this is a way of thinking that continuously criss-crosses our modern (false) categorisations (Latour 1996; Rio 1998:33). As with metaphors, we continuously expand our repertoire because of a combination of imagination, experiences, discoveries and creativeness, while other metaphors may be more primary and perhaps even universal because they are based on our lived experiences, or our being-in-the-world. Yet, phenomenology in its purest form deals only with the world of a human consciousness that results in a strict divide between objects and subjects that certainly should

Cf. also MacCormack and Strathern 1980.

18

Multiple worlds and multiple realities (Lakoff and Johnson 1999:90; Tilley 1999:34). Discussing metaphors in material culture Tilley concludes that ‘there might be in fact only one invariant cultural rule of any significance; the rule of metaphor’ (Tilley 1999:273). I will return to unity/diversity in metaphorical imagination in my description of Pangwa and Fipa iron working and aspects of their embodied metaphorical thinking in Chapter 7. My discussion of metaphors refers to rule or deep metaphors, metaphorical imagination and moral aspects that I will argue were connected to the technology and the master smelters and the value attached to their objects.

be rejected (Latour 1999:9). Latour argues that phenomenology leaves us with the saddest of all stories; a complete split or divide between the worlds of science and the intentional strategies limited to humans (ibid.). Latour sees objects or artefacts as in a one-to-one relation with humans, thus technologies may, as I will argue, contain information on how bodies are experiences and how technologies then are conceived (cf. Latour 1996, 1999). Merleau-Ponty stated that the body is our general medium for having a world (Merleau-Ponty 1989:146). He acknowledged that phenomenology appears to say everything and because of this he was unable to define its purview (Langer 1989:163). Notwithstanding, the reflections made in Merleau-Ponty’s approach enable us to be aware of our presuppositions and restores a sense of ‘wonder’ vis-à-vis the world. The point of departure one obtains with a phenomenological insight and that the fact that it may ‘be practised and identified as a manner or style of thinking’ (ibid.) is of more importance than trying to define what it involves. This is essential and allows us if we are concerned with technologies, humans and the material world, to discuss and describe technological agency itself from a phenomenological standpoint without taking a fully phenomenological approach. I will try to demonstrate that Merleau-Ponty’s emphasis on the human body as an anchor is a valuable point of departure if focus is on the processing of both iron and meaning. In the following chapters I will describe and discuss the elaborate use of metaphors in iron working among the Pangwa and Fipa within this framework.

Lakoff and Johnson who appreciate Merleau-Ponty’s ‘lived experience’, embodied philosophy and a bodily basis encompassing experience, in philosophy as in our everyday world, have argued for the same attitude with respect to language and specifically our metaphorical constructions (Johnson 1987, 1993; Lakoff and Johnson 1999). Lakoff and Johnson stress the significance of the human body and its physicality as a base for metaphorical imagination. According to Johnson ‘a metaphor is not merely a linguistic expression (a form of word) used for artistic or rhetorical purposes; instead it is a process of human understanding that we can make sense of’ (Johnson 1987:15). A metaphor in this sense is a process by which ‘we understand and structure one domain of experience in terms of another domain of a different kind’ (ibid. and cf. Barndon 1996b). As Johnson, and anthropologists like Fernandez (1980, 1991), Geertz (1973), and Wagner (1986) see it, metaphors or tropes are meaning organised in terms of verbal references. However, as they argue, metaphorical construction it is not only a linguistic phenomenon ‘it is one of the chief cognitive structures by which we are able to have coherent, ordered experiences that we can reason about and make sense of’ (Johnson 1987:xv). Through metaphor we make use of patterns that obtain in our physical experience to organise our more abstract understanding’ (ibid.).29 Johnson and Lakoff, in line with Merleau-Ponty, see the construction of metaphors in connection with our bodily experiences and they give the body pre-eminence in meaning constructions. They suggest that our construction of meaning is constrained by our bodily experiences and that bodily imagination plays a fundamental role in construction of our experiences. Yet these experiences are used in different contexts, moving ‘from one domain of experience to another’ (ibid.). Johnson exemplifies this when he notes that if several objects are placed on top of each other we create the metaphor ‘more is up’. This metaphor argues Johnson illustrates how we think and explain phenomena with a bodily basis. Because this metaphor is again based on another root metaphor ‘the body is a container’. We experience our body as a container and the metaphor ‘the body is a container’ gives a physical base for our abstract understanding of quantity. Johnson’s argument is not

2.3 Metaphors Meaning has to do with the ways in which we function meaningfully in the world and make sense of it via bodily and imaginative structures (Johnson 1987:171, 190; Lakoff and Johnson 1999:78). Lakoff and Johnson have argued that our entire conceptual system is built up of, or structured in accordance with oppositions. At least one kind of experience (our bodily) is structured by moving through metaphors from one domain in life to another (Johnson 1987; Lakoff and Johnson 1980, 1999). Lakoff and Johnson have the Western world and Western metaphor making in mind, but acknowledge that the same sort of metaphorical processes are at work in other cultures. In their 1980 publication they concentrated on the Western world while in the recently published Philosophy in the Flesh (1999) their views are more generalising, suggesting most metaphors to be based on mutually shared bodily imagination and experiences. As stated by Lakoff and Johnson, humans create their lives and actions in accordance with categorisations (Lakoff and Johnson 1999). Following this, metaphors may become systematically linked in the functioning of the human mind (Lakoff and Johnson 1999, cf. also Tilley 1999), and some of the metaphors that we construct are universal, while others are more culturally specific

29

19

Italics by the author.

Randi Barndon individual and a collective level and I will discuss how magic treats both the human body and the anthropomorphic furnace. The use of magic (its symbolism and ritualisation) in the technological process of making iron serves, as Turner described for the Ndembu: ‘to make visible or reveal and bridges between the unknown and the known’ (Turner 1967:48). Hence, magic and medicines being symbolic and homeopathic are based on metaphorical constructions that partake in this process of revealing and making sense of. Later in this work I will provide a discussion of concepts that account for individual levels of experience in the use of metaphors and more shared levels of experiences played out during iron making.

confined to containers. We experience by a vast number of other forms or embodied structures, such as balance, path, scales, cycles, compulsion, blockage, attraction and center-periphery (ibid.: 206). Another important aspect with metaphors is how visual imagination works in the formation of metaphors (Tilley 1999:269). Stressing the importance of material metaphors Tilley describes how different objects, such as pots, furnaces, calabashes or houses have openings and orifices, like the body, and they are containers (ibid.: 268, see also Barndon 1996b). ‘The body is a container’ seems to be an embodied rule metaphor taken into use in thinking about fire or heat transformations such as in iron making (Barndon 1996b).

2.3 Comparisons The significance of a material, an object or process is what makes it a potential metaphor. This also makes it worthwhile to use it within different domains (e.g. Turner 1967). It is used both on individual and social levels of experience and transmission of knowledge. The basis for the metaphors and thus the material metaphors or symbolism can therefore be seen at work on more than an isolated or individual level. If we accept that metaphors are grounded in embodiments (Johnson 1987; Lakoff and Johnson 1999), metaphors are about personal experiences that are played out in public where they achieve social significance, control and power. Hence, they are played out in iron making, pottery making and rainmaking rituals or wedding ceremonies.

Is it still possible to conduct ethnographic comparisons while at the most basic levels maintaining a philosophical foundation beyond to use Bernstein’s words (1993), objectivism and relativism? I think so and this suggests a standpoint that rejects an objective science or as we came to know it within archaeology, processualism or logical positivism (Gellner 1995; Gibbon 1989; Olsen 1987, 1997), while it also rejects relativism on the same premises (Strathern and Lambek 1998:14). We cannot ignore or reject but should suggest an approach that includes a certain degree of physicality and unity. Still, within this unity there is always diversity. As Wylie puts it ‘although there are certainly no such things as wholly neutral factual ‘givens’ it is also not the case that data are entirely plastic, that they are so theory-permeated that facts can be constituted at will in whatever form a contextually appealing theory requires’ (Wylie 1993:25).

Strathern and Lambek have argued that the importance of the body for social and cultural theory has been the realisation that we can understand the nature of such things as selfhood, practice, sociability and religious experience when we bring the body explicitly into the picture (Strathern and Lambek 1998:12). I will argue that by an inclusion of embodiment in our concepts of technology, its practices and symbolism can help us to understand social and cultural practices as being part of pre-industrial iron smelting.

Most ‘things’ or objects are experienced at a personal and embodied level (Langer 1989:40; Merleau-Ponty 1989:67), because we experience our worlds through embodiment (Johnson 1987; Lakoff and Johnson 1980, 1999). It is a better understanding of and description of this personal and collective experience that we need to get a grip on and how it connected to our material world (Olsen 2000). This assures me that we live in a world of multiple realities while some scientific results have made us aware of a certain uniformity across cultures and time barriers (Geertz 1979, 1995; Lakoff and Johnson 1999; Tambiah 1993). When we move about in our (ethno-) archaeological or anthropological descriptions we translate from the unknown to the familiar. This does not mean that the familiar, in this context, our Western based philosophical traditions, our scientific language and concepts (e.g. concerning technology) are the best in the sense of encompassing the ultimate or correct. There are numerous examples of how different and alien this language might be for indigenous peoples (Ani 1986; Kamalu 1990; Wiredu 1996). Therefore, I will argue for an application of phenomenological insight in studies of smelting symbolism which is in line with more recent embodied philosophies and theories that try to take

‘Material metaphors are solid and spatial, rather than spoken, and the process of reading them is more immediate’ (Tilley 1999:264). Johnson announced that he wanted to ‘explore the ways in which the body is in the mind’ (Johnson 1987:xxxviii). This forms a basis for my own starting point. I will explore how embodiment and the body as an anchor is part of technological agency and experience. Below I will describe the use of bodily based symbolism and metaphors in Pangwa and Fipa iron smelting combined with insights from Lévi-Strauss, Merleau-Ponty and Lakoff and Johnson. As my point of departure I will take ‘lived experience’, embodiment and bodily metaphorical imagination. I will focus on ‘magic’ rather than ‘science’ (or metallurgy as a series of chemical reactions). Magic in iron smelting lies within the core of metaphorical imagination and has the body as an anchor. Magic and preliminary rituals operate at both an

20

Multiple worlds and multiple realities Strathern and Lambek argue one may take Douglas’ original hypothesis and use it in an expanded comparative context (ibid. and see also Strathern 1996). Their context is concerned with comparison in terms of agency and history, but this means that the comparative approach moves beyond mere classifications. I will suggest that there is a purpose for comparison that moves beyond classifications, the purpose of dialogue’ and of seeing social and historical changes at work (Strathern and Lambek 1998:71).31 As Strathern and Lambek formulate it ‘ethnographic comparison leads us to continuously enrich and refine our language and reshape it into new contexts of debate’ (Strathern and Lambek 1998:24).

meaning into account (e.g. Johnson 1987; Lakoff and Johnson 1980, 1999; Strathern 1996). This leads on to another problematic question; whether comparison can at all serve as our vehicle when we move in time and space between past and present, between here and there and us and them. When working in the present this is basically a concern with how we compare ‘us’ with ‘the others’ of a different world. Secondly, it is about how we better understand the Western world through a comparison with the rest, with the exotic and unfamiliar (Augé 1998; Eriksen 1995; Fabian 1991; Hastrup 1996; Latour 1996). Within archaeology comparison is between us and ‘the other’ but those we compare ourselves with are others of the past. This is problematic but in order to grasp some of the forces behind the material means and created meanings associated with the symbolic repertoires in African iron working processes, past and present, I believe that one can continue to make comparisons between various smelting practices (e.g. Herbert 1993; Schmidt 1997).

Throughout this work my field experiences and observations of the Pangwa iron working process will be presented while compared with the Fipa. The Fipa constitute the first group of iron workers that I have had a personal encounter with and whose smelting practices I have had the opportunity to observe. Considering my descriptions of Pangwa and Fipa iron working I am of course aware that my translations are partly subjective and my focus is narrow, but there is a common platform on which comparisons can function as a dialogue.

As I see it the ultimate implementation of social anthropology and archaeology is still the comparative method (e.g. Eriksen 1995; Hastrup 1996; Lemonnier 1992; Strathern and Lambek 1998; Tambiah 1993). However, it must be emphasised that I am not through comparison, seeking to provide substantive empirical generalisations in order to turn comparison into classification (Strathern and Lambek 1998:3). Rather, as Strathern and Lambek suggest, by turning it the other way around, from the Pangwa and the Fipa to technology as ‘lived experience’ and embodiment, I will in the following move from technology and embodiment in general to the Pangwa and the Fipa in particular. Because what I first of all seek to get a grip on is how they conceptualised and practised the technology of iron making.

I will argue for an approach that accepts that we have to reduce from the empirical to the analytical in order to give our otherwise chaotic world meaning (Hastrup 1992, 1996 and see figure 4.1). This no longer means that only one paradigm is the correct. Henceforth, I have argued for a somewhere ‘on the middle ground’ (Johnson 1987; Rorty 1989) because all scientific approaches have some level of ‘reality’ within them. According to Bekaert, what we need within the human sciences are analytical distinctions between when one or the other paradigm or approach applies (Bekaert 1998:6). David suggests that an application of a combination of realist and hermeneutic approaches perhaps is the best way (David 1992:333). This means that, depending on the nature of the ‘real things’ and systems under investigation, such as the smelting technology as a system of chemical reactions or symbolism and meaning, different approaches and methodologies may be appropriate and therefore, the ability to predict may be greater or less’ (ibid.). The only natural thing to do then is perhaps to combine approaches and make comparisons (cf. Latour 1996). But what is to be compared with what and what is the frame of reference within which we make our comparisons?

This study might be criticised for its resemblance to Douglas’ classical methodology of explanatory comparisons of two or more ethnographic cases (Douglas 1966,1973). The social aspect of body symbolism was an important part of Douglas’ argument in her thesis on rituals, body symbolism and natural symbols. I will try to demonstrate that iron smelting was an embodied experience and that body symbolism and the body as an anchor was played out in social life as in technology.30 Strathern and Lambek have argued that Douglas’ basic comparative methodology is still useful (Strathern and Lambek 1998:15). The approach is capable of producing statements that are not necessarily of a functionalist cast but certainly belong to the realm of what Strathern and Lambek call ‘comparative social logic’ (ibid.). This view does not imply the more holistic notion of social logic as in Douglas’ analyses; rather it suggests that logic of sociality differ in some systematic ways. Hence, as

In line with Latour (1996, 1999) I will argue that in our time we have had a tendency to separate our cultural phenomena into separate categories, and especially those of distant (exotic?) places and times. We have because of this and the path that Western science took, tried to explain various categories as completely separate domains. In archaeology, Binford’s famous tripartite technomic, sociotechnic and ideotechnic categories

30 This of course was also a theme continuously stressed in Turner’s descriptions of the Ndembu (Tuner 1967, 1969).

31

21

Italics by the author.

Randi Barndon commonly associated with functional models and not a field for semantic studies. As correctly noted by Bekaert much of the theoretical debate and scientific production within archaeology during the last twenty years have been coloured by the great divide, ‘the war’ between these two standpoints (Bekaert 1998:6). Functionalists have questioned why humans acted as they did while the semantic oriented archaeologists have tried to understand what the actions meant. If as Bekaert notes one is occupied with causes for human behaviour, one uses concepts such as ‘interest’, ‘strategy’, ‘intention’, ‘function’ or ‘motif’ while if one is interested in the form of human behaviour one has to look at semantics and meaning (ibid.). Not accepting this great divide, rather placing oneself on the middle ground between them, does however not imply that our explanations or descriptions will provide any absolute ‘truth’. I think that they may provide better scientific results in the form of better descriptions. As such an embodied philosophy tries to ‘get a grip on the world’ and also on ‘the others’. Schutz with his multiple realities certainly describes what Latour later has criticised (Latour 1996). Of major importance is how both views make us aware of the setting in which we conduct our research, our descriptions, transcriptions (Lemmonier 1992) or categorisations of what we have observed.

referring to cultural phenomena as primarily within social or ideological spheres is an outstanding example (Binford 1962). For decades this approach narrowed the focus of archaeology towards a primary occupation only with the lowest steps in ‘Hawkes ladder of inference’ (cf. Trigger 1989), while post-processualists have been too focused on the upper levels in their studies of symbolism. This has resulted in a partial ignorance of the connection between the multiple steps in ladders. This has greatly influenced studies of African iron working as being merely related to metallurgy as understood and conceptualised within Western science. Within this frame technology and hence metallurgy as an object of study has been excluded or classified as peripheral. Iron working has been regarded as a technological phenomenon and not a cultural or social phenomenon, or rather both. Since metallurgy has been understood initially by its base within natural science, other expressions than those available for depiction in chemical formulas or economic models have not been fully explored. Metallurgical approaches miss the human component in metal technologies, the human activity and human resources. In metallurgical approaches the human resources are not seen as essential to directing the technological process used in producing metal out of ores (Kriger 1999:30). Another result of the path Western science has lead on to has been our tendency to separate technology from embodiment (objects or matter from subjects and symbols) and thus technology from experience and transmissions of skills and experiences. Because of this the cultural content or semantic aspects were for along time not issues to explore in studies of technology. I am particularly thinking of processes of the imaginative use of metaphors moving from the self to society and between objects and humans.

The distinct approaches and theoretical levels that I will attempt to implement in my writing all deal in one way or the other, with iron technology and more recent iron producing communities. They also take into consideration the actual practice of doing ethnoarchaeology among ironworkers. In sum this study is about the people that demonstrated iron smelting processes to me. It is about aged men who once were active smelters, blacksmiths and highly regarded members of iron using communities. They had various levels of understanding of the process and knowledge of the technology that they practised. Thus, their level of insight and understanding of the technology, both in our common Western conceptualisation of the phenomenon (Doner and Bugliarello 1979; Gosselain 1992; Ingold 1990; Mitcham 1979), and indigenous African based conceptualisation of the technologies, are issues that I will address.

My approach incorporates insight from phenomenological thinking without taking a fully phenomenological position. Simultaneously I believe that with Lévi-Strauss’ structuralism as a method we can see various levels of meaning at work and how the ironworkers made sense of their technology. Some aspects of meaning in iron smelting activities seem to be structured into oppositional pairs although others were conceptualised as being forceful in their combinations. I am concerned with the conscious use of embodied metaphors and how they were implemented into social control and control of power among the participants in iron working teams. Finally, I will demonstrate how parts of basically grounded metaphorical thinking were materially manifested. This materiality worked back on how the body was understood or conceptualised.

2.4 Summary The Greek philosophers asked how we could know, and Aristotle concluded that we could know ‘because our minds could directly grasp the essence of things in the world, thus a metaphysical realism (Lakoff and Johnson 1999:94). For Aristotle there was no direct opposition between ontology (what there is) and epistemology (what we can know), because the mind was in direct touch with the world (ibid.). A focus on the embodiment of the mind as argued for by both Merleau-Ponty (1989) and Lakoff and Johnson (ibid.) brings us a lot closer to Greek realism than modern and post-modern (analytical) philosophy does. Hence, there are certain aspects about being human

Until recently within archaeology the distinction between the two opponents objectivism versus relativism manifested itself in a distinction between functional models on the one hand and semantic or symbolic models on the other hand. Consequently, technology was most

22

Multiple worlds and multiple realities that are grounded in embodiment and more general metaphorical thinking. We use our body in living in the world and in making sense of the amalgam of activities that constitutes this living. Looking for embodiment and embodied thinking in technology will yield insight into the rich symbolism, combining many levels and categories that once characterised African iron working.

construction and use of metaphors, regard categorisation (and dualism) as central and universal in our experiences with our selves, our bodies and our environments. Perhaps, then, the way we have classified and made categorisation has been wrong, rather than the fact that we all make categorisations based on our own experiences and in encounters with others and objects.

Latour argues against the all-encompassing and constitutional divide between mind and matter that our philosophical foundation has upheld and forced upon all sorts of categories in our world. He criticises phenomenology as being part of the false modern categorisations that we make (Latour 1996:78). The ‘modern’ divide between subject and object, person and society, nature and culture, has according to Latour prevented the development of a more holistic understanding of objects and material culture production, not only within one category but within them all (ibid., Latour 1999:194). Lakoff and Johnson from another standpoint and being more concerned with the particular

The following discussion of meaning in iron making symbolism is as Schutz formulated it ‘a special ordering of reality’ and I will move between two opposed sets of theories or approaches; phenomenology and structuralism.32 At some levels in our analysis concerned with people and their material culture we can still accept and implement a middle ground between absolutism and relativism because there is a mutual platform or spatial foundation of lived experience and embodiment where we can meet and continue to explore. However, I will first explore how we have within a Western scientific baseline approached technology as a phenomenon and the relevance of this for studies of African iron working technologies, past or present.

32

23

Cf. Chapter 7.

Randi Barndon

24

A discussion of technology, the concepts and its contents

3. A discussion of technology, the concepts and its contents Our future, not only its shape, but indeed its very existence - depends critically on our ability to relate our technology to our humanity (Bugliarello and Doner 1979:vii). 3.1 What is technology?

technology.

According to Latour the greatest characteristics of modernism has been its efforts to divide and classify phenomena of all sorts into specific categories that have been regarded as in opposition to each other or nonrelational (Latour 1996). Because of modernism, which Latour claims never really has existed, apart from in science, technology has been treated as an object of study isolated from other realms, such as politics, religion, symbolism or art. Within modernism technology and magic have been seen as distinct categories. Pfaffenberger connects our separation of technology from the cultural and the approach that archaeologists and anthropologists have had towards technology with modernism (Pfaffenberger 1992). Another characteristic of modernism, although a cultural, artistic and literary period, is therefore ‘a period that represents an extreme ambivalence towards technology’ (ibid.: 495).

Plato was the first to deal at length with this notion and in his work, tekhne and episteme, art and systematic or scientific knowledge were closely associated (Mitcham 1979:173). Socrates initiated this line of thought in his distinction between two types of tekhne, one which consisted of mainly physical work but required minimal use of language (such as painting or sculpture) and another which was more intimately bounded up with language, such as arithmetic, logic or astronomy. On the other hand, Socrates did not consider cooking a technology, rather a atechnos and a tribé, a mere knack of routine based on experience (empeiria) (ibid.: 173).33 Later Plato, in Philebus, extended his thoughts on tekhne and moved in the direction of modern notions of technology. Our modern ‘Standard view’ of technology is based on Plato’s classification of knowledge and his division of tekhne into two classes (cf. Bugliarello and Doner 1979; Pfaffenberger 1988, 1992). One class is concerned with education and upbringing and another with making and producing (Mitcham 1979:174). Of the second, technical knowledge, there were again two types, those (such as music, medicine and agriculture) which proceeded by conjecture and intuition based simply on practice and experience and those (such as carpentry) which consciously involved the use of numbering, measuring and weighing. Activities such as carpentry possessed more exactness or precision akribeia (which implied more insight), and it is this which is tekhne in its primary sense (ibid.: 174). Tekhne was therefore clearly distinguished by Plato from all human action and knowledge of a political sort (education and by extension, governing), as well as any kind of logos and pure theory. Because of this it was associated more closely with the activities of making or producing which operate upon the non-human material world (ibid.). Finally Plato’s opinion that ‘of these activities those were most truly tekhne which involve the greatest degree of quantitative precision’ (ibid.), has strongly influences archaeological thinking.

Although perhaps elementary, the ‘womb’ of all Western science lies in the origins of Greek notions of science (Tambiah 1993:8). We have to acknowledge as science the great developments over stretches of time in medicine, metallurgy, geometry and astronomy in the Near East to which Greece was heir and also the achievements of Ancient China that precede Greek science (ibid.). The ancient Greeks did not develop any scientific mentality as such. They had no conception of ‘science’ that can be considered equivalent to our own which developed first in the seventeenth century. The Greeks however had a notion of philosophy in their term philosophia which refers to love of wisdom; episteme which means knowledge; theoria referring to contemplation, and peri physeos historia, inquiry about nature. All these were later to merge into what we call science (ibid.). The word technology as we have used it in our studies of iron metallurgy stems from the Greek technikon which belongs to tekhne, commonly translated as ‘art’, ‘craft’ or ‘skill’ (Ingold 1990:9; Mitcham 1979:172). In The Questions Concerning Technology Heidegger discusses two things in regard to the meaning of the word tekhne. He addresses the distinction between the Western-based Standard view on technology with its origins in Aristotle as opposed to his own view that technology is part of the existential structure of man’s being (Heidegger 1977; Hood 1983:352). For the ancient Greeks, tekhne was a word or name not only for the activities and skills of the craftsmen, but also for the arts of the mind and the fine arts. As such tekhne belonged to bringing-forth, it was something poetic, belonging within poiesis. And hence, there was no distinction to be made between the arts and

In Aristotle’s work we find the first ventures to unite tekhne and logos (Mitcham 1979:183). The first appearance of the concept technology is found in Aristotle’s teknologia, which referred to an awareness of the world. Aristotle continued to include episteme within his notions of technology, hence for Aristotle, as for Plato, tekhne was a special knowledge (ibid:175). 33

25

Cf. also Latour 1999:231.

Randi Barndon 1979).

Aristotle’s approach to technology then, reads as follows: ‘tekhne is a habit or stable disposition to act in a specific manner with a true logos concerned with, or ordered towards, making the human production of material objects. This is in line with the original Greek notion of tekhne as a capacity of action bounded in habits. Tekhne according to Aristotle was also connected with language and therefore it was teachable although not in our own conception of teachability. Like virtue, tekhne is learned primarily through practical imitation and Aristotle postulated that ‘men only become builders by building’ (ibid.: 182-183). In sum, then, teknologia as once perceived is more in line with my own view of technology.

By contrast, in the ancient worldview, for the alchemist, matter was conceived as an aspect of God (Eliade 1958:183; Mitcham 1979:187). In short, there is apparently a void between the early more holistic Greek usage of tekhne until the English term technology is established and its connotations enter into common use. Mitcham believes that technology has its origin in the early Greek use of tekhne through a later Latin source, although the Medieval Latin dictionaries do not go beyond the twelfth century (Mitcham 1979:184). It was not until the last half of the seventeenth century that technology in its modern form was established as a concept (Ingold 1990; Mitcham 1979).34 Ingold defines technique as referring to skills, regarded as the capabilities of particular human subjects, and technology to mean a body of generalised, objective knowledge, insofar as it is capable of practical application (Ingold 1990:7). Furthermore, quoting Mauss and his Techniques du corps (1935) Ingold confirms ‘that it is not necessary to use a tool to implement a technique. For instance dancing is a technique without an instrument’ (Ingold 1990:7).

Mitcham concludes that tekhne was associated with a passion or love for matter and that it would take great concentration to become deeply acquainted with any material object. Abstract knowledge, the ancient Greeks argued, was easy to acquire and identify, while concrete knowledge was a different thing. Because technology is absorbed by means of the sense organs and muscles, concrete knowledge of matters uses quite different channels (Mitcham 1979:182). Thus, in tekhne lies a component that in itself is non-logical. Here, according to Mitcham, lies the difference between the Greek tekhne and the modern technology. Tekhne involved logos, but only in grasping form, not in directing the actual process of production, the activity qua activity: whereas this is exactly what modern technology tries to grasp, a logos of the activity, a rationalisation of the process of production. As discussed by Ingold, tekhne was separated from mekhane, which referred to the manually operated devices that assisted its application. The classical dualism between tekhne/mekhane was eventually transferred to the modern dualism between technology and machine, also resulting in the movement from the personal to the impersonal (Ingold 1990:9). This is the way modern science (archaeology and social anthropology) has treated technology as a phenomenon; a notion of technology as independent of any conception of eidos or form, and which claims to be neutral and dependent in its use on what humans want to do with it or how they experience it.

What relevance have the ancient Greeks concepts of technology for our present understanding of recent African metallurgical practices? Are these ancient concepts of any relevance at all in gaining a better understanding of non-western iron working traditions? An awareness of the history of philosophical dialogues on technology (and matter as material culture) is relevant because it has framed all archaeological concepts of the role of technology. It has framed the role of the technicians and processes of technological change. Discussing the concept itself makes us aware of the background to and content of our ‘Standard view’. It acts as a point of departure and encourages a more holistic approach towards technology that includes matter and material culture in its direct relation to humans. Few archaeologists have ever been concerned with the actual contents and historical background to their use of or their notions of technology as a phenomenon in prehistory. Equally, few have been aware of the fact that in our own worldview, a view that has positioned matter as something inactive and natural, we have tended to separate material culture and physicality (including our bodies) from our minds and our thoughts. Technology has been treated in very much the same way, isolated from how we render our actions meaningful. Our efforts in understanding worldviews and cosmologies have been disconnected from studies of technologies or rather technological behaviour even if technology (i.e. tool

3.2 From tekhne to technology In the later half of the seventeenth century western man’s understanding of the ontology of matter underwent a radical transformation. Under the influence of Galileo (1564-1642), Descartes (1596-1650) and Newton (16421727) and their followers, the material world began commonly to be regarded in much the same way as Aristotle looked upon words. The material world was seen as something separated from cosmos and matter was no longer thought of as anything living. The trend is easily seen in the Cartesian theory in which matter was seen as a pure, lifeless extension in itself ordered towards nothing else. Consequently matter was regarded as something to be done with as one pleased (Mitcham

34

Technology was not in the 1658 edition of Philllip’s The New world of English Words but included in the 1706 edition of the same dictionary (Mitcham 1979:184).

26

A discussion of technology, the concepts and its contents distinction was also taken up by Leach when he distinguished between technical and ritual types of aspects of behaviour (Ingold 1990:10; Leach 1976:9). Leach argued that technical behaviour is defined in purely pragmatic means-end terms and produces observable results in a strictly mechanical way. Ritual behaviour by contrast, is essentially communicative, and serves to convey information in a symbolic code, about group membership or social identity (Leach 1966:403, 1954:12, 1976:9 in Ingold 1990:10).

making) as an aspect of material culture is among the primary foci in archaeology. An awareness of the history of philosophy of technology, tekhne and episteme, arts and knowledge thus clearly confirms that the concept is in itself a product of the Western worldview (Ingold 1990:14). Our modernist frame around technology has been too narrow to account for technologies in non-western, non-industrial and prehistoric contexts (e.g. Spier 1973). This same view has been too narrow for any accounts of the social and cultural significance of our present day technologies.

In spite of a rejection of the Western-based ‘Standard view’ (Bugliarello and Doner 1979; Pfaffenberger 1992:493, and exemplified by Ingold above) we cannot and should not totally ignore it, because at some levels in our analysis the technical actions that serve to alter our world can only be explained in terms of our specific scientific language, our old fashioned ‘Standard view’. Within metallurgy it is simply not possible to produce forgeable iron without the correct chemical formula and correct temperatures, carbon content and combustion. Without these factors, among many, it is not possible to make iron no matter what ‘language’ is used. However, this language of chemical formulas is ours and it does not contain any absolute ‘truth’ either. Within prehistoric and pre-industrial metallurgy there are many unsolved problems, many unanswered questions. How can we deny that these are as are all material production and associated technologies, partly determined by factors in our environmental and substantive world? Within the technology of iron smelting there are matters of the physical world that beyond dispute are facts and factors in the metallurgical process. Technologies, although perhaps not so perceived in local terms, have their particular environmental, physical and chemical constraints.

The modern Western view, the ‘Standard view’, does not capture Heidegger’s ‘arts of the mind’, nor the ancient Greeks poetic attitude towards the technologies of the craftsmen. Our Standard view has not taken seriously into account the cultural or social frames that form technological processes and activities and techniques. Aristotle’s thesis that ‘men only become builders by building’ refers to another notion of technology. This was an embodied view of technology which later influenced Mauss (1935) and consequently more recent notions of technology and techniques within French anthropology. 3.3 Technology and culture, social representations and styles As technology itself has developed we have moved away from the close ties that were once recognised between tekhne and poiesis, technology and culture.35 In our time technology is mostly grasped as something controlling its own path. This has led to a ‘scientific’ logic as the answer to what lies behind all technological traditions. The second point of importance in Heidegger’s discussion of technology is that until Plato, the word tekhne had been linked with the word episteme. Thus both words are names for knowing in the widest sense and Heidegger postulates that technology is a mode of revealing (Heidegger 1977:13). Technology comes to presence in the realm where revealing and unconcealment take place, where the truth happens (ibid.). Heidegger concludes that this definition of technology best applies to the techniques of the craftsman and does not fit machine-powered technology (ibid.), an idea that is elaborated in Mitcham (Mitcham 1979).36

If we are trying to grasp the metallurgical process in iron reduction through chemical analyses and when detailed knowledge of the nuances in this field is needed, our ‘scientific model’ based on the western traditions of knowledge is at the moment the only applicable one. My point here is simply that only in the Western world do we separate technology from other aspects of cultural and social life. We separate technology and science from magic in order to understand certain phenomena of the world, but building on this knowledge is not necessarily wrong in our sense of right and wrong (Rorty 1987).

Ingold has argued against the commonplace anthropological absolute distinction between the domains of technical and social phenomena (Ingold 1990). This

Simultaneously with the development of the critiques of the ‘Standard View’ (Bugliarello and Doner 1979; Mitcham 1979) and the Anglo-American anthropological neglect of cultural aspects of technology, the ‘French School of Technology’ emerged.37 In France, a specific ‘anthropology of technology’ has gradually taken into consideration technology as understood within a cultural

35 Cf. Collett’s early discussion of a cultural frame for the studies of ancient and recent African metallurgy (Collett 1983), and Gell’s discussion of technology and magic in anthropology (Gell 1988). 36 According to Mitcham existentialist philosophers such as Ortega y Gasset, Jaspers and Heidegger along with technologists, such as Dessauer and others associated with Verein Deutche Ingenieure, out of a desire to understand what was being recognized as the dominant influence on our times, gave birth to the philosophy of technology as a distinct discipline (Mitcham 1979:166, and cf. Sigaut 1994).

37

The history of technology and interdisciplinary studies of technology is also often referred to as ‘Science and technology studies’ (STS) (Pfaffenberger 1992:492).

27

Randi Barndon for all the unity of their broad outlines and their extension over vast multi-ethnic territories, is their strongly marked local and individual character (ibid.).

and social frame (Dietler and Herbich 1998:262.fn.3; Hegmon 1992, 1998; Lemonnier 1986; Cresswell 1983 in Lemonnier 1992; Pfaffenberger 1988, 1992; Schlanger 1990:23; Sigaut 1990, 1994). Mauss’ initial discussion of body rhythms and techniques and how all body and muscular habits and behaviours are socially or culturally acquired was of special significance to the French school. As examples Mauss used the way we walk, way we sit down, sleep, swim and run (Mauss 1979 [1935]: 97-105). Mauss argued that techniques are traditional and influenced by their past, they are based on reason and practice, they are like all social phenomena, arbitrary, and particular to the community which engenders them (ibid.: 23). In his discussion of swimming Mauss illustrated the cultural specific and social component in body techniques. The swallowing of water to speed up the swim was when he was a boy believed to be efficient while twenty years later swallowing water during the swim was no longer practised (ibid.). Mauss introduced a preliminary list of body-technological phenomena that he considered worth while to study cross-culturally, such as walking, running, swimming or eating (ibid.: 110-119). In Mauss’ thesis the referent is the physical body itself, as moulded by society and culture (ibid.: 10). Technical behaviour has always a social foundation as Mauss demonstrated in his ethnographic observation of the diversity of ways in which various members of different societies used their bodies (Schlanger 1990:23). Haudricourt, in line with Mauss, criticised the ignorance of ethnographers towards these aspects of technologies. He believed that they could have been observed and collected (through description) in the same manner as artefacts have been ‘brought home to museum collections’ (Haudricourt 1968:57).38 As correctly noted by Lemonnier if Mauss’ thought on body technology had been heard by anthropologists earlier ‘the anthropology of technology’ would already have reached a lot further (Lemonnier 1992:1).

For Leroi-Gourhan operational sequences constitute the building blocks of technology, indeed of culture (White 1993:xviii). As pointed out by Schlanger, to fully grasp the technical act in all its efficiency it is necessary to comprehend it as traditional (Schlanger 1990:23). By traditional Leroi-Gourhan, as Mauss, meant that the actions by the performers in technological processes are culturally or ethnically conditioned and highly structured. But through repetitions and conditioning, at a young age, they become more-or-less subconscious. Studying these structured actions, these chaînes opératoires, is as LeroiGourhan suggested, an accessible entry for archaeologists and anthropologists into social organisation as well as cosmology. Leroi-Gourhan’s gesture connotes ‘material action’ as it refers explicitly to the manual creation of a material culture that is extra corporal. Gesture according to Leroi-Gourhan paralleled speech as a form of expression of mind and language. This notion of gesture was linked to his chaîne opératoire theoretical approach (White 1993:xvii). Leroi-Gourhan made it clear that apart from humans the behaviour of animals is deeply embedded in operational sequences but only among humans do we see how these sequences take material form and become more-or-less permanent manifestations in the environment (ibid.). In his work we see a belief in technological systems, or rather the altering of the material world or matter, to be technological systems with social and cultural representations. This approach is still valuable and relates to a basic philosophical foundation that is argued for in our times, our ‘taking back our bodies’ and our world conceived in line with our ‘lived experiences’. We perceive according to our embodiment, connected to and through our technological behaviour. This encapsulates a philosophical foundation for studying technological behaviour and a conceptualisation of technology that gives room for technology to be understood as something more than something one has to do before one can go on to something else (Hood 1983).

Leroi-Gourhan, inspired by Mauss’ essay on bodytechniques developed a method and a theoretical approach towards the techniques of the body. His famous ‘operational sequences’, chaînes opératoires, are defined as a series of technical operations which are those that bring a raw material from its natural state to its manufactured state (Leroi-Gourhan 1964 and cf. Sigaut 1994). In my particular case, the actions and behaviours by the smelters when they are processing and transforming ore into iron constitutes a multiplicity of operational sequences. Leroi-Gourhan noted that ‘the technical life of the hunter and later the farmer and the artisan, involves a large number of sequences that correspond to the many actions needed for their material survival’ (Leroi-Gourhan 1964:253). These sequences are empirical, borrowed from a collective tradition that one generation passes down to the next. Their principal trait, 38

In line with Mauss’ ideas, technology can be defined as ‘embracing all aspects of the process of action upon matter, whether it is scratching one’s nose, planting sweet potatoes or making aeroplanes (Lemonnier 1989:161, 1992:1). Mauss’ socially founded notion of the actions upon the material world and connotations towards the body and person relates well with our contemporary thinking about the body and technology and our general re-embodiment of conceptions of all phenomena (Johnson 1987; Lakoff and Johnson 1999; Strathern 1996). Mauss and his predecessors have all been more or less occupied with individual embodied technological practices. My own work will attempt to employ parts of this insight in my analyses of the teamwork that constituted iron working. My argument is that the

Translation by the author.

28

A discussion of technology, the concepts and its contents his third category. He also believes that influences can go from general social representations to technological ones or vice versa (ibid.: 81). As I see it, Lemonnier’s ‘projects schema’ is connected to the use of metaphors in technologies. The most interesting question in Lemonnier’s social, rather than symbolic perspective on technology is why there is such variation in solutions to technological systems (Lemonnier 1989). In his words ‘the crucial question in an anthropology of technological systems becomes: given the existence of social representations of technologies, up to what point do they arbitrarily constrain choices made by societies among possible modes of action on matter’ (ibid.: 161). And further ‘these social representations have an obvious influence on style that is to say on those technological traits that carry immediate information and are nonutilitarian in physical action on matter. Thus are there domains of material culture more prone to arbitrary variation than others’ (ibid.). And finally he states that ‘although archaeologists only recover small parts of the social and cultural elements in technological systems doesn’t lessen the fact that they are part of the general theory of material culture and technology’ (ibid.: 156). However, technology, like other social products cannot be reduced to just a single dimension (Lemonnier 1992:118). Several dimensions or aspects of technology, parts that together may be said to form a ‘technology and cosmology’ of iron smelting in East Africa, must be considered. This approach has gradually been adopted by archaeologists such as Gosselain (1993, 1998), van der Leeuv (1993) and Dietler and Herbich (1989:fn.3, 1998).

technological process was a context in which individual socially acquired body techniques and socially embodied techniques were at work. Many actions were unconscious, learned after many years of apprentice, but, as I will discuss below, the technological process was a context in which embodied experiences were played out in order to stage morality and prestige. Lemonnier among others, and some inspired by him, such as Sigaut (1990), Dietler and Herbich (1989) and Gosselain (1992, 1993, 1998, 1999) have taken Mauss and Leroi-Gourhan’s insight as their sources of inspiration for a contemporary reinvestigation of technology.39 Lemonnier’s theoretical approach, as well as his concepts, derive from Leroi-Gourhan (1964) and Cresswell (1990, 1983 in Lemonnier 1992). Lemonnier states that if we are to understand technological systems as social representations we must record them within a chaîne opératoire approach (Lemonnier 1992:6). His approach therefore entails that technological agency can be understood as ‘a technical process composed of a series of operations that result in the production of an object’ (Dietler and Herbich 1998:262,fn.3). Lemonnier believes technology to contain all sorts of aspects of the process of action upon matter, although he is first of all interested in the social representations of technological systems (Lemonnier 1986, 1990, 1992). He is not specifically concerned with technology as understood within a semantic frame (e.g. Collett 1985, 1993). Nor is he concerned with magical practices as part of technological behaviour, although these aspects are included in his approach. According to Lemonnier there are four elements connected to technological activities that are central: the matter to which an action is directed, objects (tools or means of work) including the body, gestures and movements organised in operational sequences. These elements comprise a system and are the basis of the diverse technologies in any given society (Lemonnier 1990:156, 1989:156). These elements are observable in iron smelting contexts, even if these contexts include several actors.

The study of style has a long history within archaeology (cf. also Barndon 1992). This makes the linkage between technology and style relevant (Gosselain 1992:560). Plog (1983), Sackett (1977, 1990), Wiessner (1984), Wobst (1977) and Conkey (1990) have been the major contributors to the ongoing debate. Sackett’s often cited definition of style reads as follows: ‘Style enters the picture when we see that the artisans or any given fraternity (or sorority) are aware of only a few, and often choose but one of the isochrestic options potentially available to them when performing any given task, and that the choices they make are largely dictated by the technological traditions within which they have been enculturated as members of the social group that delineates their ethnicity’ (Sackett 1990:33).

Lemonnier defines ‘social representations’ as sets of ideas shared by members of a given social group. He divides social representations of technologies into three categories; tacit knowledge or Aristotle’s ‘one can only become a builder by building’, specific technological knowledge or know-how and finally ‘projects schema’ which represents other mental representations alongside the operational sequences and other sequences in the process (cf. Lemonnier 1992:79-81). Lemonnier proposes that since technology is also social representation, meanings can be inferred from technological systems. Thus, material culture as symbols and social representations constitutes 39

Sackett concludes that with an isochrestic perspective a specific butchering technique may well convey as much ethnically significant information as the typology of the tools with which it was carried out (ibid.: 35).40

40 The link between style and ethnicity is complicated in view of more recent debates on ethnicity (e.g. Jones 1997; MacEachern 1998).

Haudricourt’s first publication on technology stems from 1936.

29

Randi Barndon Lecthman correctly states that ‘the much more difficult step is to argue from a confident understanding of the style of technological behaviour to more fundamental, deeper cultural patterns which inform that behaviour’ (ibid.: 11). An ‘anthropology of techniques’ as advocated by Lemonnier does not dismiss, for instance, metallurgical understandings of the smelting process as irrelevant. On the contrary, Lecthman and Steinberg suggest that this aspect of a technology must be understood as from the inside (Lecthman and Steinberg 1979). While the concept ‘technological style’ (Lecthman 1977:4) opens for an inclusion of ethnicity within studies of technology, but within this approach, as well as Lemonnier’s, there is a danger in framing nontechnological behaviour and styles as reflections or social representations rather than integrated parts within the technology or chaînes opératoires.

Lecthman introduced the concept technological style as early as 1977 (ibid.).41 Asking what technology can tell us about culture and how we can read it, she suggested technological style as a fruitful concept in our processes of discovering non-metallurgical aspects in metalproduction (ibid.: 3-4). Like her French contemporaries (e.g. Lemonnier and Sigaut), Lecthman revealed by her notions of technology that technology as culture contains much more than our Western-based Standard view has been able to grasp. Echoing Mauss’ formulation that ‘there is not only technology where there is an instrument’ (Mauss 1935 in Ingold 1990:7, and cf. Sakett 1990:33), Lecthman points at how we seem not to have recognised, or at least paid much attention to the style of the activities which produce the artefacts (Lecthman 1977:6). She calls technological style: ‘That behaviour which is characterised by the many elements that make up technological activities - for example, by technical modes of operation, attitudes towards materials, some specific organisation of labour, ritual observances - elements which are unified nonrandomly in a complex of formal relationships. It is the format or package defined by these relationships that is stylistic in nature, and it is the style of such behaviour, not only the rules by which any of its constituent activities is governed that is learned and transmitted through time’ (ibid.: 6).

David, Sterner and Gavua placed the discussion into perspective by their article on Mafa and Bulahay decorated pots and body decoration by suggesting that the pots were persons (David et al. 1988).42 Perhaps, as Gosselain suggests, that technological behaviour is influenced by metaphors, with steps of the chaîne opératoire becoming the locus of a symbolic discourse (Gosselain 1999:205). I will follow this line of thought combined with Johnson and Lakoff’s thesis discussed in Chapter 2 that the construction of metaphors is one of the most fundamental forms of structuring experiences, crossculturally. Consequently, this opens for an entirely new field of inquiry only recently starting to be explored (e.g. Tilley 1999). Gosselain argues for a framework for pottery technology in which technology and culture influence one another (Gosselain 1992:560). Taking account of Lemonnier’s notions of choice and options in chaînes opératoires, he has explored Cameroonian potters and their technological styles.43 He argues that, since external and environmental aspects do not restrict technologies, choice and style merge. The variants that occur are choices of ‘different ways of doing the same thing’ (Gosselain 1992:560 after Lemonnier 1983:17).

Lecthman argues that ‘the style is recognisable by virtue of its repetition that allow the underlying similarities in the formal arrangement of the patterns of event’ (ibid.: 7). Of major importance is how Lecthman and Steinberg have criticised the distinction between technology and culture and technology and society on the basis that technology is culture (Lecthman and Steinberg 1979:138). Lecthman and Steinberg suggested that it would be fruitful for archaeologists to investigate technological style as a phenomenon as well as the manner in which individual styles of technology relate to other aspects of the cultures in which they occur (ibid.: 3). Lecthman, as did Mauss, postulates that technological behaviour is stylistic and that it is learned and transmitted through time (ibid.: 5-6). Lechtman suggests that by asking what is the cultural component in technologies, we may also ask what can technology tell us about culture (ibid.). Furthermore, since technology is like art, music and dance, it is a cultural phenomenon that reflects cultural preoccupations that are expressed in the very technology itself. Thus, technologies are performances that constitute cultural messages and they are communicative systems. The styles are their symbols through which communication occurs (ibid.: 13). But, what are the messages and how can we read them?

3.4 Habitus in technology Mauss defined the body as ‘man’s first and most natural instruments, or man’s first and most natural technical object, and at the same time, man’s technical means, his body’ (Mauss 1979:104 in Strathern 1996:13). Mauss was the first to introduce the concept habitus after his discussion of swimming, marching and running (Schlanger 1998; Strathern 1996:11). Because technological knowledge is very often tacit knowledge that must be experienced (Barndon 1992; Dietler and 42 The same may be said of Latour’s discussion of subjects and objects as seen as human and non-humans (Latour 1999). 43 Gosselain refers to Lecthman (1977) as well as Lecthman and Steinberg (1979) but does not mention her specific and early introduction of the concept ‘technological style’.

41

Lemonnier’s first publications on technology as systems of social representations stems from 1976 in Techniques et Culture, 1:100-151, thus simultaneously with Lecthman’s first publications.

30

A discussion of technology, the concepts and its contents morality. I have chosen not to incorporate the concept habitus as used by Bourdieu because it does not fully account for agency and the multiple possibilities that lie within paths of operational sequences in smelting technologies. Finally, habitus as such does not account for the final judgements of products produced. The general definition of habitus, and its homogeneity is valid. Certainly many technical performances within the smelting context were based on habitus similar to how we go about and cope with most of our technical tasks in our daily lives.45 However, when combined with a chaîne opératoire approach it does not really add new dimensions of active and conscious behaviour in reaching technological goals.

Herbich 1989, 1998). Mauss saw habitus as historically variable rather than universal, in other words as the embodiment of cultural differences encoded in education (Strathern 1996:12). Hence, Mauss insisted that habitus is learned and not natural (ibid.). Mary Douglas, whose emphasis on the body derives from Mauss, argued that bodily control is an expression of social control. Bourdieu following this emphasis on the body (as a source of encoding memory) claims that the body is an index of society (Strathern 1996:27), and that the body is socially informed (Bourdieu 1977:124). Habitus may be defined as ‘a practical logic and sense of order that is learned unconsciously through the enactment and repetition of everyday life’ (Gilchrist 1999:xv). Bourdieu suggests that habitus is ‘the generative principle of regulated improvisations’ (Bourdieu 1977 in Dietler and Herbich 1998:246). Bourdieu argues that space only comes to have meaning through action, something which has also been demonstrated in Moore’s study of the Endo in Marakwet (Moore 1986).

3.5 Science, magic and technology The Greeks were, as noted by Tambiah familiar with the category of magic (magi, magea) referring to magicians and their religion (Tambiah 1993:8). Of interest in technological studies is how the Greeks combined an interest in magic with their other work in peri pyseos historia, the inquiry of nature, and what resembles science. They did also at some level distinguish between medicine (science) and magic, but neither of these two categories were opposed to religion (ibid.: 11). These combinations were later to be regarded as absolutely incompatible because of the ‘mind-body gap’ that developed within Western philosophy.

Bourdieu combines his notions of habitus and practice with two approaches that have proven to be of value also for studies of the past; the phenomenological philosophy of for instance Heidegger, Husserl and Lévi-Strauss’s classical structuralism. This combination resembles Merleau-Ponty’s phenomenology with its focus on the body (Schmidt 1985), and Bekaert’s various repertoires in iron working symbolism that will be discussed in Chapter 7.

If we again return to Mauss, he stated that ‘magic is linked to science in the same way as it is linked to technology’ (Mauss 1972[1950]).46 ‘Magic is not only a practical act it is also a storehouse of ideas but as far as magic is concerned knowledge is power’ (ibid.: 143). Malinowski was the first anthropologist to demonstrate empirically within the confines of one single society or culture, how ‘symbolic’ activities like ritual and magic were linked to and interacted with activities of a practical or pragmatic character (Tambiah 1993:68). The Trobriand gardener said ‘the belly of my garden swells’, and by this he transferred the scheme of pregnancy in the human body over to the scheme of garden fertility, thereby setting up a correspondence that recognises both contexts (Malinowski 1935; Strathern 1996:28). For Mauss, as for the Trobrianders, there was a link between magic and the body. Mauss elaborated this when writing about how bodily movements accompanied rituals (Strathern 1996:27), as they do technologies. That there were or are relations between practical and ritual activities at all, has been unfamiliar to archaeologists who have had a tendency to divide these into two spheres of life. In archaeology we commonly recognise phenomena as practical or functional and state that we have understood

In spite of the approval of Bourdieu’s use of habitus combined with Lemonnier’s chaîne opératoire approach the concept is problematic when questions of social change are in focus.44 Habitus reproduces itself over time unless external factors intervene (Strathern 1996:28). As noted by Strathern ‘Bourdieu’s approach appears to render moot the question of agency and consciousness and questions of morality become hard to handle in Bourdieu’s scheme because how are we to understand the numerous cases when people deviate from or actively resist moral injunctions or when they afterward experience remorse’ (Strathern 1996:28). Iron smelting, as a technology was not only based on habits and routine. The technological agency seen as steps in operational sequences was not unconscious and merely a representation of other realms. Messages about morality and the well-being of persons were merged with perceptions of the technological process and the furnaces. Some metaphors were deep or rule metaphors based on embodiment while others were consciously used. Symbolic repertoires and symbolic reservoirs may be seen as part of the iron workers habitus and when used in the iron smelting context these concepts more specifically touch upon the well-being of the body and folk models of 44

45

This notion of habitus has recently been accounted for (Dobres 2000:137). 46 Mauss first published his theory of magic in 1902, the first English translation appeared in 1950.

I.e. as suggested by Dietler and Herbich 1998:246.

31

Randi Barndon understanding of technology in general.

their significance, something we find more difficult with objects we define not as functional but ritual.

Apart from the work by Mauss (1935, 1950), Malinowski (1935) and later Leroi-Gourhan (1964), few researchers have ever been concerned with technology as a cultural and social phenomenon with its specific link to magic. This technological-cultural-magical void lasted from the late sixties up until the early eighties among most English-speaking archaeologists. They never developed ‘a theory that accounted for the cultural or social aspects in technological practices’ (van der Leeuw 1993:238). Technologies were seen as adaptive responses to other, mostly external, factors. Matter and the material was something that was not in itself active with its ‘own life’, or connected to anything living (active). Thus, changes in technology were seen as determined by ecological or other economic factors. The use of magic and the explanations it provided were not considered.

Malinowski believed that magic served two functions, one psychological and the other sociological, while Tambiah suggests magic to be directed towards the uncontrollable agencies which affect the success of practical activities (Tambiah 1993:72). This is a specific explanation and he continues ‘that one may label as ‘anxiety reduction and compensatory action’ (ibid.). Adding to the context of iron smelting the fact that the master smelters had to demonstrate their pre-eminence as masters and control access to the craftsmanship opens for another social and economic dimension in the use of smelting magic. In African iron smelting many things were at stake and one can easily acknowledge the use of magic. Is it then the case that magic begins where technology ends (Tambiah 1993:72)? Is it only when something is really at stake, as in Trobriand deep-fishing or African iron smelting, that magic comes into play? Malinowski argued that this is not the case. Magic is not primitive science or a confusion of natural and supernatural. Magic works with and is additional to practical knowledge and technique. Malinowski demonstrated this in his study of the Trobrianders and found that magic played a central role in ocean fishing but not in lagoon fishing, it was central in yam cultivation though not in coconut cultivation. These two cultivations are from a general perspective very similar. Is the use of magic, elaborated rituals, sacrifices and use of medicines, then related to the fact that yam cultivation among Trobrianders was bound up with prestige? Was the performance of magic connected to prestige in possessing magical knowledge and performing magic? One is tempted to compare this with African iron smelting magical practices and the fact that knowledge is, as Malinowski argued, an expression of power (ibid.). I believe that in iron working, both smelting and forging, it was essential to announce and establish power and prestige relations. However, if so, this is only one part of the ‘repertoire’ that the smelters played out during the smelting and forging activities (cf. Bekaert 1998). As argued above, more individual repertoires were also played out and based on bodily experiences of deep symbolic reservoirs.

Anthropologists were, of course, occupied with technological change but their concern was directed towards other social and economic mechanisms or, as in more recent anthropology symbolism and symbolism in everyday practices. These and related aspects were studied cross-culturally in terms of ideology or cosmology, while technology and magic were not included in these categories. Although as Mauss suggested, a specific cross-cultural study of body techniques would reveal new aspects of cultures there was no focus on how technologies might be about cosmology or morality. Few, if any, were enchanted by the fact that studies of technological practices may provide answers to aspects of both mundane and mystical character in past societies as in contemporary societies. In more recent debates within both anthropology and archaeology the Western view of technology, our ‘Standard View’ (Bugliarello and Doner 1979), has been questioned and criticised. It has been accused of not giving social and cultural aspects in technology enough attention (Collett 1985, 1993; Ingold 1990; Lecthman 1977; Lecthman and Steinberg 1979; Lemonnier 1986, 1990, 1992; Sigaut 1990, 1994). Malinowski’s accounts of the Trobriand gardeners is famous for its descriptions of the magical aspects and magical language of gardening (Malinowski 1935). It seems strange that his direct discussion of the relationship between magic and technology (cf. also Malinowski 1948) has not been more influential in debates within archaeology in general and in African metallurgy in particular (e.g. Merwe and Avery 1987; Rowlands and Warnier 1993, 1996). Rowlands and Warnier point at how we as ethnoarchaeologists must not make a division between productive techniques and magic as EvansPritchard did (ibid.). They criticise van der Merwe and Avery for the same mistake, although they do not take Malinowski’s Coral gardens and their magic fully into account (Rowlands and Warnier 1993:512).

In the two volumes of Coral gardens and their magic Malinowski illustrated how the Trobrianders themselves made a distinction between the language of magic and the language of ordinary speech (Malinowski 1935 and cf. Tambiah 1993:74). Tambiah suggests that magic for the Trobrianders is, as for us ‘a part of the theory of human life in which the path of ritual action is seen as an indispensable mode for man anywhere and everywhere of relating to and participating in the life of the world’ (Tambiah 1993:83). This is essential and reveals a relation between the two categories that I believe is important to bear in mind in all studies of pre-industrial iron making processes as well as for a broader

32

A discussion of technology, the concepts and its contents believe Tambiah is correct when he states that actors engaged in technological activities, as in other activities, make a shift into and out of different orderings of reality (Tambiah 1993:136). So, what kind of reality? What kind of cultural meaning is embodied and expressed in technological artefacts and in technological activities, and how can we explore into them and describe them?

In Malinowski’s words, we see how in Trobriand life ‘magical or ritual acts and technical and practical acts were interlaced and interdigitated into a larger amalgam which we can label Trobriand yam cultivation or canoe building and so on’ (Tambiah 1993:136). As Winch did 40 years ago we must recall Collingwood’s claim ‘that some accounts of magical practices in primitive societies offered by ‘scientific’ anthropologists have often masked a half-conscious conspiracy to bring into ridicule civilisations different from our own’ (Winch 1990 [1958]: 103). This most certainly was the case in the Enlightenment period but the uncritical and non-reflexive manner of writing has prevailed until the present day. The descriptions of ‘primitive’ technologies coated in magic as opposed to Western scientific chemical formulas to explain what really happens in the process of transforming ore to iron is one example. But what is magic and its significance in technological activities? Is it accurate to state that in pre-industrial societies (lacking ‘science’) there is a natural magic linked with technology and opposed to magic (Tambiah 1993)? Is it correct that in some pre-industrial and pre-technology-concept societies, thus all prehistoric societies, ‘natural magic’ is material production, (thus technologies), and different from other forms of magic such as religious magic or the use of magic in for instance initiation rituals or procreative rituals.

3.6 Masters and magicians The centre of smiths and smelters activities lie in the knowledge and power of the techniques and of the supernatural, of magic and medicines. Both in Ufipa and Upangwa iron smelting was a seasonal activity and the master smelters were farmers as were their co-villagers. They were part-time blacksmiths, and forging was carried out on a more regular basis, related to the demand for new iron objects and repairs. In Upangwa regular magicians and medicine men were specialised into more than ten different categories (Stirnimann 1967:417-18). A Pangwa master smelter was socially highly respected as a magician and medicine man but would not be a village headman. As one of my informants put it ‘the Pangwa did not live in enclosed and well organised villages before 1960!’47 The master smelter, umkololo and umkoyo, was still one of the most respected within his the patrilineal decent group, lutanana.48 He was both an iron master and a medicine man, he was respected for his skills and knowledge in transforming ore into iron and he often became quite wealthy because of this knowledge.

When it comes to debates on iron smelting magic, research in the coming years may illustrate that the division between science and magic with its origins back to the Enlightenment period is indeed irrelevant. During the Enlightenment technology could be connected to science, but not magic (Tambiah 1993:24), while today David suggests an ethnoarchaeology of African iron smelting traditions that inquires into how magic and science impose on each other (David 2001).

A master blacksmith, umponji, was not necessarily identical with the iron master, the umkoyo, although commonly he was a master of both techniques. Hence, once iron was extracted the forge, ulufuumbu, was a family business. Forging was (and is) carried out in the mornings or afternoons in between gardening or other seasonal activities. The most common objects produced by the master, who also was a blacksmith, and often an admired one, were hoes, axes, spears, sickles, bells and bracelets. Iron hoes produced by the technology mastered by the umkoyo were the most important bridewealth objects, imali, in Upangwa (Stirnimann 1967:405), in Ukinga (Koizumi 1995:169), in Ubena and Umatengo (i.e. Culwick and Culwick 1935), and, as commented on

The question is then, how are we to describe and interpret the interlacing of magical and technical acts that forms an amalgam, or a total activity that we may label ‘yam cultivation’, ‘canoe building’ of the Trobrianders or ‘iron working’ as among the Pangwa and Fipa. Tambiah’s question is, as he admits, inspired by Malinowski’s descriptions of ritual and technical sequences as being ‘parallel’ (Tambiah 1993:71). Malinowski wrote that ‘the association between technical pursuit and its magical counterparts is, as we know, very close, and to the natives essential. The sequences of technical stages, on the one hand, and of rites and spells, on the other run parallel’ (Malinowski 1935 quoted in Tambiah 1993). This observation introduced in Malinowski’s study long ago has not had much influence in archaeological studies of technology. Archaeology has not developed approaches or tried to demonstrate how the most available means to both conduct or understand technical behaviour, the practitioners own embodied experiences or simply their own bodies and their own beliefs, cosmology and morality have been at work in technological agency. I

47

This is how he expressed the changes that followed the carrying out of Mwalimu Nyerere’s Ujamaa political program in Upangwa. Before the 1960’s Upangwa settlements had no headman or chief with formal contacts and obligations towards the Ludewa District Commissioner office. They based their social unit and residence more on the patrilineage, uloxolu, then territories (cf. also Koponen 1988:210-211; Stirnimann 1976:263). 48 According to one of my informants the umkoyo would have more wealth than others because people would buy his iron or his hoes. He said a master smelter got one rupee (or rupia) for one hoe, and most people bought them to get married.

33

Randi Barndon above, in pre-colonial Ufipa (Willis 1981:146).49 Because of this the Pangwa master smelter could afford more wives, more arranged workgroups and ‘beer parties’ for himself and his clan. His wealth also made it possible for him to build new furnaces. An old man proudly told me that his father who had been a highly respected umkololo and master smelter, umkoyo, in Lugarawa had built more than three furnaces on his own land.

-The umukanga vaksuhusa, a rainmaker. A Fipa master smelter, who shared his title, umwaami, with the leader of an elephant hunt, was a master who because of his medical skills and magical basket intaangala, was able to build a furnace and smelt iron (Barndon 1992). Early records claimed that the master smelter acted as a priest and that he was often a medicine man (Robert 1949:240, Wise 1958:232). He was not a regular doctor, sing’aanga, who also used an intaangala basket, but was often appointed the village headman and was highly respected and had well-established bonds with the Fipa kings at Lyangilile or Nkansi (Willis 1981). Modern Tanzanians considers the Fipa as the southwestern people most strongly involved in witchcraft and magic (cf. also Willis 1968). A Fipa master smelter, as others associated with the group, was also known to have magical skills beyond the iron making context. He was able to go into a trance and could trace thieves. This has not been paid enough attention in studies of iron working since the focus has been on sexual or fertility symbolism and gender models. Hence, in order to achieve a broader cultural and social understanding of the significance of iron working in pre-colonial and colonial East Africa one has to look at the value of iron as a metal and material and secondly as a source for metaphorical thinking and therefore the role of the smelters and smiths. The master smelter was a man of power because from the moment when the smelting season or each smelt was initiated he was a magician as much as he was a technician. This made him famous and strengthened his social reputation as person of prestige and power. Because of his magical skills he was able transform essences and to communicate with the supernatural, the ancestors and spirits.

My informants confirmed that a Pangwa master smelter was a magician and medicine man, umukanga, and his medical knowledge and magical skills went far beyond what was needed in iron working. Our master smelters stated that they were magicians and medicine men. They were wearing special iron bracelets around their arms as an indicator of their skills and knowledge of magic. One of them claimed to be a rainmaker, umukanga vaksuhusa, while the two others were herbalists and healers, umukanga vaksutuleksa, and specialists in plant medicines. As opposed to a Fipa master smelter who had special knowledge of the magical and magical contents in his basket, intaangala, that he only used in iron working, an umkololo among the Pangwa smelters was often, or would become, the umkoyo of the lutanana. The Pangwa master was therefore a particularly skilled magician. The Pangwa are known throughout Tanzania as medicine men, magicians and poison makers. Stirnimann has observed how they were divided into eight groups of specialists (Stirnimann 1967:417): -The umukanga vaksulondola or the ‘prophet’ that is able to prescribe the correct medicines through a reading of the pattern observed in ashes; -The umukanga vaksutuleksa imikoda, the herbalist or medicine man of plants end herbs and who knows all sorts of poison (two of our aged men claimed to possess and practice this knowledge); -The umukanga vaksabutula, a specialist of body treatments who used special medicines for inhalation; - The umukanga vaksuvung’anya, a magician that was/is able by the use of medicines to settle a dispute between relatives; -The umukanga vaksuteka iliteko, a specialist that by the use of medicines can prevent death and infertility; -The umukanga vaksupinga iliteko, a specialist that can prevent the death maker from succeeding in his use of magic; -The umukanga vamwafi, a specialist that by putting poison in drinks can prevent a magician in succeeding in his doings;

Childs argues that before the building of the smelting furnaces, offerings and sacrifices to gods and spirits inhabiting the locations or otherwise being connected to the place, were generally performed (Childs 1991b: 340). Before the actual smelt begun it was once again common to perform rituals and often a sacrifice was provided. The master smelter would sacrifice for the gods and ancestors, spirits and old master smelters. Food and drink were presented for their spirits (Herbert 1993:92-93). The beer and food provided was like a feast that marked the end of the long period of preparing the smelt and if the furnace used induced draft it also marked the beginning of a time of hard work; the time for blowing the bellows (MacLean 1996). If we can recognise the role of the master and thus his conceptions of the smelt, the many preliminary rites, prescriptions and prohibitions that followed the smelting process will also become clearer. Looking for when rituals were or are performed might inform about which steps along a path in the process, in terms of local conceptions of technology, are the most essential (i.e. Childs’ list of sequences 1991b:340). It might also illustrate how the Western made dichotomy between

49 Further north in Tanzania, in pastoral societies, cattle rather than hoes were used as bridewealth. During colonial and post-colonial times cattle and goats (and currency) have become more common in marriage transactions. Stirnimann notes that wedding hoes were made in Ubena while Koizumi claims that in Kinga rituals lwembe hoes were forged by Bena blacksmiths (Koizumi 1995:139; Stirnimann 1979).

34

A discussion of technology, the concepts and its contents ritual behaviour to be seen as a form of quasi-textual representation. The rite’s function is to communicate shared values within a group and reduce internal conflict (ibid.: 50). The third and historical explanation claims that rites cannot be understood in terms of their internal structure alone. All rituals were invented at one point and over the historical span in which they remain in existence they are susceptible to a change in their meaning (ibid.: 51). Connerton’s distinction of rituals is relevant and combines aspects of ritual activity in the context of iron working, while I will argue against the emphasis on representations or connotations as the primary aspect of rituals in iron making.

magic and technology may possible merge (cf. David 2001; Rowlands and Warnier 1993, 1996; Tambiah 1993). Finally, it may illustrate an activity where one sees individual experience, entrepeneurship and collective memory at work (Connerton 1989). Tambiah validates Malinowski’s discovery of the sociological function of magic (Tambiah 1993:73). Malinowski saw how the ritual expert or specialist by conducting imperative ceremonies or preliminary ritual and rituals mobilised manpower and resources. The rites served as triggering mechanisms for the sustained conduct of practical operations (ibid.). Before the building of a furnace and the smelting started preliminary rituals and magic, sacrifices and offerings were prepared and prohibitions and prescriptions were mandated.

It is true that rituals are by definitions different from everyday communication and signification and they are carried out for specific purposes (Bloch 1992:46; Turner 1969). In almost all societies rites are different and set apart from everyday life such as cosmic festivals, rites de passage, rites of affliction and exorcism ceremonies (Bloch 1992; Tambiah 1993:71). On the other hand it is important that most people conduct ritual cycles or intermittent rites that are interlaced with practical activities, such as agriculture, fishing or crafts (Tambiah 1993:71). Because actions in these settings are both technical and magical/religious they constitute totalities so that simplistic disaggregations into separate technical and expressive aspects is not possible. It is of even greater importance, however, that even if they occupy the same space they are often internally distinguished. The magician being a technician used his experience as a metallurgist when looking into the peephole to control the temperatures but he used aspects of his repertoire as a magician when he spoke to his ancestors and spirits of the dead, while he used his power as the socially defined master smelter when he excluded women or men from his crew.

If we consider the role of the master smelter as a magician and the value attached to the iron bloom and iron objects we may better understand the rituals in the process. This leads on to how and why metaphors were made and materialised. The making of iron was not only a matter of transforming ores into iron using the correct magical and metallurgical ‘formula’; it was a time for social performances. As I will return to in Chapter 9 the masters had to perform rituals, offerings and sacrifices in order to maintain his reputation as a successful master. 3.7 Ritual performances in technology A chaîne opératoire approach in studies of technologies as systems of meaningful behaviour is the best ever suggested (David 2001). It is highly relevant within studies of African iron making (ibid.) and might, when practised, produce new sources for comparisons between technologies in the past and more recent times. A partial implementation of Lemonnier’s approach will be followed when Pangwa and Fipa iron smelting is described in Chapters 5 and 6.

Ritual and magic in iron working, and the material symbols made and used in the process have commonly been seen as external to the smelting process itself and as representations of something other than the process itself. What is not metallurgical is seen as something external, semantic or symbolic. This is a result of our Western scientific baseline and has made it more difficult for us to conceptualise or grasp totalities (cf. Budd and Taylor 1995; Latour 1996; Tambiah 1993). The objects produced are generally classified into agricultural tools, weapons or prestige objects, ignoring the overall ritual and magical context in which they were all produced. This context in itself made the things into valuable objects. In most studies of African iron working the amalgam that constitutes the process is distinguished into separate categories and researchers are specialists in either metallurgy or iron working symbolism. This inevitably makes it difficult to grasp intricate multiple realities or multiple levels within a smelting context. It is still harder to escape the hidden notion we all have of the metallurgical part of the technology to be more ‘true’ or

A ritual is specifically defined by its difference from mundane events because it creates a specific ‘ceremonial space’ for the performers and onlookers (Bell 1992; Bloch 1993; Damm 1998).50 Definitions of rituals lie at the core of most ethnographic and cultural anthropological work while the explanations of them are diverse. Connerton divides the various explanations of ritual action into three separate categories of what he terms psychoanalytical, sociological and historical explanations (Connerton 1989:52). The psychoanalytical position views ritual behaviour as symbolic representation. Rites are the systematically indirect statements, encoded in the symbolism of the rite of conflicts which that rite disguises and to that extent denies (ibid.). The second and sociological explanation believes 50 Damm (1998) situates the discussion of ritual more specifically within studies of material culture and archaeology.

35

Randi Barndon ‘basic’ than ritual or magical components. But as noted by Avery ‘iron ore although certainly necessary, was merely one of the ingredients that went into the furnace’ (Avery et. al. 1988:262).

smelters structuring their mental images of the smelting process? In what way was their mental world part of the smelting process? How and why was magic used in the smelting context?

If we accept the above arguments for a holistic approach towards technology, it is not only necessary to record presmelt and smelting activities but also the preliminary rites and ritual performances which were equally essential parts of the process. The performance of preliminary rites was important among the other preparatory tasks. Sacrifices, prayers to ancestors, taboos, body decorations or the use of various sorts of magic and spells, songs and dancing were part of, and at different steps in the process integrated into, the process. Among the most significant of the preliminary rites was the use of magic, which by definition may be said to include offerings and certain prohibitions such as food taboos and taboos against human sexual intercourse. I will argue that by an incorporation of the use of magic and medicines, a new dimension is added to the step by step transcription of what can be observed in iron smelting as an embodied technological process. Following from this we have to accept that the forged tools contain identity associated with their fabrication.

As discussed at length above, Mauss, Leroi-Gourhan and Lemonnier and Lecthman have all been aware of the weakness in the Standard view. This view has prevailed among most Anglo-American anthropologists and archaeologists. The French school advocated in particular an approach that placed greatest importance on attempts at gaining a non-metallurgical, non-chemical and therefore also a partly phenomenological approach to material transforming technologies. This is the approach I will follow in my descriptions of Pangwa and Fipa iron smelting practices. A ‘non-metallurgical’ approach is most likely closer to a pre-motorised, pre-mechanised reality (e.g. Haudricourt 1968). As Lemonnier has illustrated, technologies even very sophisticated ones, involve many arbitrary decisions and choices that are not directly determined by material or physical constraints. Rather they reflect higher systems of meaning within societies (Lemonnier 1986, 1989, 1992, 1993). This is in line with Heidegger’s discussion of technology (Heidegger 1977; Hood 1983).

3.8 Towards a framework for technologies

I have in the previous pages argued for technology and particularly metallurgy to be understood within a frame that accounts for embodiment and experience. I believe, as the ancient Greeks did, that technology is indeed a process that is both poetic and bringing-forth, and perhaps for the actors in Pangwa and Fipa iron working activities, it is where the truth happens, to use Heidegger’s words (1977). As David argues, I believe that Lemonnier’s chaîne opératoire approach forms our best theoretical and methodological starting point when doing ethnographic or ethnoarchaeological studies of iron working (David 2001). Nevertheless, Lemonnier’s approach may be refined if the material aspect is further developed according to how technologies were perceived and transmitted by their practitioners. Spatial and decorative (both mental and specifically material) aspects in technologies should be incorporated within the approach not only be taken as external or representations. I will therefore discuss location, operation and decoration as constituting relevant concepts or reference points when comparing iron working technologies. Location, operation and decoration as stylistic dimensions in the smelting process form a conceptual tool for my own view on how we can explore and describe African iron working technologies.

A shift of focus towards an all-inclusive mapping of the ritual or symbolic sequences in the smelting process is important for a discussion of how the components of technology and magic were working and how agency and experience were integrated. Such a shift of focus may even inform us about where to look for the collective and individual repertoires and multiple realities and levels of both technology and magic in historic and past contexts. It may also inform about the most critical steps in the process. Each set of actions, such as ore collection, charcoal production, tuyere making and the collection of magical substances are in need of similar step by step observation and descriptions, since all of the sequences becomes a (cultural) component in the process. Finally, it was important to produce a success, the masters had to extract not only iron from the ores but admiration from the audience. Discussing iron making as meaning making has to be based on notions and perception of the body and the role of the master smelters as technicians and magicians. Participation in technology is an experience (Barndon 1996a). It is based on experience, individual and common memory, and is recreated through this experience (Bourdieu 1977; Connerton 1989; Merleau-Ponty 1989; Strathern 1996).51 Why is this the case, and how were the

It encapsulates material, technological and moral aspects and strives to constitute the amalgam that iron smelting practices once were for the Fipa and Pangwa technicians. It also suggests some of the paths to follow into the description of past and prehistoric iron smelting practices, the beliefs that surrounded them and how these beliefs were used in and between social boundaries. Adding dimensions more specifically concerned with space and

51

Both Bourdieu (1977), Connerton (1989) and Merleau-Ponty (1962) discuss this in their use of Mauss’ habitus/ habitat/habitude concept (cf. Strathern 1996).

36

A discussion of technology, the concepts and its contents consider within a reconstructed iron smelting context. They were all relevant to those people who demonstrated their skills to me; in other words, they form part of the actors’ repertoire. As I proceed in my description of Pangwa and Fipa iron working in Chapters 5, 6 and 7 I will discuss how these three aspects, which also are exclusive contexts in themselves, interact as chaînes opératoires in a technical behaviour that is a stylistic cultural production.

style, what I have chosen to term location and decoration, may also yield new paths to follow in our observations of and transcriptions of technologies into chaînes opératoires. They provide for some of the meaning content in the iron working process and constitute valuable dimensions when material aspects are in focus. Both Lemonnier’s and Lecthman’s approach on operation are valuable, but neither of them are concerned with the locales in which the operations are taking place and how this is related to the perception of both humans and furnaces.52

Rorty has argued that through common sense we can postulate certain things, events and aspects to be true (Rorty 1979). These variables constitute such a point of departure, but since the field situation consists of encounters with other people, the situation is not identical to the context in which archaeological material culture studies are carried out. They (the people) as objects in the world are no longer seen as entirely separated from our subjective experiences of them (Merleau-Ponty 1989). The presence of peoples in my case studies has on the other hand informed me about how these variables are in themselves manipulated by the practitioners in the technological process.

Spatial and operational variations and choices are related to and connected with the third of my categories; decoration. Decoration refers to embodiment, bodily movements, conscious or unconscious mental constructions, ritual behaviour and real symbolic decoration of the material objects, being part of the process and not divided from function or operation. Therefore, a description of the decorations of the furnaces is as central as a description of combustion because through these actions metaphors were evoked and sometimes they were materially manifested or simply material metaphors. Bodily movements performed in the technological process form a ‘decoration’, a mentalmaterial relation in itself regardless of whether the action is tuned towards decorating in our common use of the term. In using these concepts, my aim is to illustrate how technology and magic are related to each other through embodiment. It is anticipated that adding to the chaîne opératoire approach these dimensions will contribute to a shift in our conceptualisation of technological agency and its material manifestations. Hence, it is suggested that these categories form important components in experiencing and practising a technology.

3.9 Summary My discussion of the concept of ‘technology’ and our approach towards technological behaviour has attempted, in line with Mauss, Haudricourt, Leroi-Gourhan and Lemonnier and Lecthman to grasp how the body is an active element in forming chaînes opératoires, but also a part of our creation of embodied metaphors (Johnson 1987; Lakoff and Johnson 1999; Merleau-Ponty 1989). The initial Greek understanding of technology, incorporating both magic and cosmology (religions) is closer to my own views on pre-mechanised pre-industrial technology than any Western based assumptions about technology and the ‘Standard view’ that has prevailed within modernism. Within modernism our conceptions of technological agency moved away from an approach that included embodied and lived experiences within what forms a technological experience. European culture with its scientific manifestations has on the other hand treated technology and metallurgy as physical and chemical processes that can be predicted and controlled by the help of materialistic theories. Of greatest importance in technical choice have been environmental and functionalist explanations. Even if this approach is applicable in order to understand technology in scientific terms and from the outside, it is less applicable on production processes in societies that not themselves use Western scientific terminology, such as prehistoric peoples (Collett 1993:499). However necessary it might be, the major problem with the scientific approach is that it leads the researcher to divide the different activities that occur when a raw material or an object is produced into separate categories. One divides the activity into sections such as technical, magical or symbolic, although for the

I will transcribe my observations in a way that can separate the different Western based categories (cf. Latour 1996:11), but hopefully also combine them and turn them into what they are; an interconnected meaningful activity. I will divide and then describe location, operation and decoration. I see these dimensions as both separate and inter-dependent aspects in the preparations for a smelt and the smelting process. Incorporating these dimensions into our notion of technology within a chaîne opératoire approach may bridge some of the gaps in our approaches towards the present with the past. In past contexts we cannot observe chaînes opératoires as Lemonnier suggests, but we have traces of both the locations in which operations were performed and the decorative elements they were incorporating within these settings. As such, these dimensions represent aspects of technological practices to be compared within groups of people and between groups of peoples. All three dimensions or aspects are relevant to 52

The term technological style was first introduced by Lecthman (1977), and was discussed in an ethnoarchaeological study of iron working by Barndon (1992).

37

Randi Barndon participants they are not necessarily separated but work at different levels. Refining and combining Lecthman’s concept technological style with Lemonnier’s chaîne opératoire approach into the most important dimensions in iron smelting, operation, location and decoration, may inform about stylistic variation and unity within diversity. Hence, careful observation of the procedures and locations of these dimensions will encapsulate technical agency as individually and collectively metaphorically based material production.

38

Re-enactments of smelting, how to describe and transcribe?

4. Re-enactments of smelting, how to describe and transcribe? After more than ten years of living with smelters at a variety of sites in central Africa and of building and operating furnaces, I have never encountered a serious characterisation or classification of materials in the Western senses (Avery et al. 1988:262). were often very detailed. Technological processes were often given numerous pages in monographs in addition to all the detailed drawings and illustrations, photographs and tables (e.g. Kuntz 1935; Smith and Dale 1920).

4.1 African iron working research The early explorers who arrived in remote parts of the African continent sent back vivid descriptions, often followed by drawings that depicted their observations and experiences with the wild beasts and the savages (cf. Appendix D, table 1). The ‘savages’ were used as examples of the difference between Europeans as modern, developed and industrialised, and the Africans as the opposite (Andrén 1997:83). For the explorers and travellers, coloured by, if not aware of the notion of western beliefs in modernity among themselves, described traditional ways of life, pre-industrialised technologies and primitive magic. One of the most spectacular African crafts ever described was iron working and the early explorers provided detailed descriptions of traditional iron smelting in their monographs, which for the most part were concerned with geographic and more general cultural aspects. Observations of African iron smelting and smithing were also recorded by missionaries, district commissioners and others associated with the colonial rulers. Additionally, we find numerous ethnographic descriptions that are more related to local belief systems than was typical for the early travellers’ accounts. Local concepts and explanation of the transformation of ore into iron is provided for in most of these documentations, although greatly coloured by their time (e.g. Brelsford 1949; Burton 1889; Fülleborn 1906; Hore 1892; KerrCross 1890; Lechaptois 1913; Livingstone 1874; Park 1899; Robert 1949; Smith and Dale 1920; Thomson 1881; Wainwright 1942; Wyckaert 1914).

Killick suggests that the interest appear to have been aroused by the evident contrast between the small handpowered furnaces and forges of Africa and the waterpowered (and later steam-powered) capital-intensive metallurgical establishments of their homelands. This gave rise to a steady trickle of written accounts of African iron working from the earliest contacts to the decolonialisation in the 50’s and 60’s (Childs and Killick 1993:317; Killick 1990:3). Within the positivism that founded New Archaeology, grounded on the Standard view within Western-based modernism the entire process of iron working was disconnected from culture and society (Pfaffenberger 1992:495).54 Technology was translated into European terms and placed within European values and an ethnocentrism that valued rationality (Gilje and Grimen 1992:254). The relation between technology and magic was seen as a primitive expression and illustrated a technological level of development that disappeared from Europe many thousands years ago (Tambiah 1993:42-45). Purely technologically or ecologically deterministic explanations were provided as explanatory models for the development and transmission of metallurgy. Presumptions of needdriven technological evolution prevailed and one was filling in details of uniformitarian models while observed non-technical activities were ignored or even ridiculed.

Andrén has argued that the colonial history has founded a homogenous archaeology in Africa since it has been intimately associated with colonialism and its ideology (Andrén 1997:82-83). The worldview of the period of the Enlightenment; civilisation and ideas about progress and development, had a strong influence on how early travellers, ethnographers and missionaries observed iron making (Schmidt 1996a:3). This is clearly seen in how they reported about their experiences with indigenous iron smelting (e.g. Robert 1949; Wyckaert 1914).53 Ever since early colonial times African iron working has been of interest to the colonial administrators, missionaries and travellers. Then it was a living practice; common among most of the ethnic groups the Europeans met on their travels (cf. Cline 1937, Kense 1983; Koponen 1988). Fortunately it was considered important to document these ‘primitive technologies’ and such documentations

It seems unavoidable to mention Binford in this context. Culture according to Binford was extrasomatic means of adaptation (Binford 1962). Thus, technological activities and artefacts were regarded as a class of ‘authentic’ artefacts and activities to fulfil ‘real’ needs (Pfaffenberger 1992:496). But as clarified in more recent anthropological studies, it has proven almost impossible to actually observe any artefacts or objects as having pure and authentic, rational needs (ibid.). Thus, Hodder has successfully demonstrated that the famous great divide; ‘the war’, between function and meaning can no longer be upheld (Hodder 1982). As discussed in the previous chapter the ‘Standard view’ prevailed as the dominant explanatory model as well as the model that we have lived by in our everyday lives (e.g. Schiffer and Skibo 1987; Spier 1973). In terms of archaeological interpretations of the initial coming of iron to Eastern and Southern Africa

53 Eliade notes that the Ngoni and Chewa had human sacrifice included in the smelting rituals (Eliade 1958:68).

54

39

Technology was described and discussed but not defined.

Randi Barndon The once abundant active practice of iron smelting that met the early explorers and ethnographers, and which was urgently recorded by the archaeologists of the seventies and eighties has vanished from sub-Saharan Africa.55 Only a few elders in East Africa know how to smelt or participated in the activities of making iron. Several researchers of iron smelting have experienced skilled iron smelters dying before their research was completed or during the writing up of fieldwork (e.g. Schmidt 1997), as I have. This has of course made the awareness of the urgency in this kind of documentation even stronger (ibid.). At a time when entire technological systems are disappearing as fast as the societies that developed them, the efforts behind fieldwork are beyond dispute. Especially since more recent theoretical discussions of what technology is all about recognise that all aspects of a people’s technology, the entire process of producing and using material culture should be taken into account (Lemonnier 1992).

the argument was that hoes produced by the new technology, by iron, made the conquering of the rainforest and the clearing of land possible or more efficient. The rituals and sacrifices in the technological practices were regarded as part of a primitive formula since they, the natives, did not as their European counterparts, understand technological matters. Of primary interest were theories and models that could establish and create models for input, of for instance, fuel to actual iron output. These variables were again considered in terms of efficiency based on our modern conceptions of technology (e.g. Goucher 1981; Schiffer and Skibo 1987). Although this perspective on technology, or ‘Standard view’, does not include more recent research in archaeology one may claim that it once was recognisable within Africanists metallurgical research (Schmidt 1996a:3). It was especially practised within processual archaeology; an approach that was well within the boundaries of the logical positivism (Gibbon 1989). This of course coloured African metallurgical studies and like within archaeology in general, one was first and foremost interested in life in the laboratory (Gilje 1993:21; Olsen 1987). Metallurgical classifications of various smelting technologies were the primary interest while ritual and magic were neglected or seen as part of a primitive expression or ‘primitive formula’ (Merwe and Avery 1987:144). This view was clearly within the ruling western objective perspective and magical practices as integrated parts of the technology were still not regarded as causal elements. Within this perspective ritual and magic could not contribute to an understanding of African iron technology, its initial introduction, spread or transmission and the subsequent conservation of technological knowledge.

4.2 What is iron working? Pre-industrial iron smelting was a human activity and technological agency may be seen as an individual and socially shared embodied experience. Studying iron working can inform us about embodiment and how metaphors were taken into use, applied and materialised through technical behaviour. I am therefore, first of all concerned with social and individual repertoires and the various levels of experience and meaning seen expressed in and used by people in a technological context. Still, for the reader to follow my description of the smelts observed, a basic metallurgical introduction to the requirements of iron smelting is provided. It must be regarded as an introduction to the descriptions and classifications of the smelts as they were observed, not as a contribution to metallurgy.

At the same time as this approach to technology was upheld in archaeology, the once vivid and detailed descriptions of technologies were left out of contemporary social anthropology. The links between culture and technology were therefore again ignored in many cultural studies. While today, partly inspired by the ‘Cambridge School’ and later French ‘Anthropology of Technology’ archaeologists incorporate attempts at understanding the social and symbolic contents of technological agency. This shift of focus must be seen in relation to the general theoretical change from a processual to a post-processual archaeology (Barndon 1996b), as well as the more associated general shift from modernism to post-modernism (Pfaffenberger 1992, but cf. Latour 1996).

Consequently, but perhaps elementary, the distinction between blast furnaces and bloomery furnaces has been based on the differences in the ready produced products from the two methods. Blast furnaces have been classified by the production of cast iron that is run from the furnace in the liquid state at high temperatures. The product from the bloomery furnace is a solid mass of iron (usually containing some entrapped slag and fuel) called a bloom (Schmidt 1997:112). This is the general distinction between the indirect method and direct method of iron extraction. But iron smelting can be achieved by a continuum of processes ranging from the bloomery furnace producing low-carbon iron at one end of the scale, to the blast furnace producing high-carbon cast iron at the other. The bloomery process for producing lowcarbon iron has been termed the direct method and the route using cast iron as the starting material have been called the indirect process (Avery et als. 1988).

However, this was not the case in studies of African metalworking and pottery, which in spite of processualism or New Archaeology have incorporated ideology and symbolism, based on ethnographic evidence (i.e. David et.al 1988; Kense 1983; Maret 1982; Schmidt 1973).

55

40

Cf. Appendix D, table 1.

Re-enactments of smelting, how to describe and transcribe? The production of high-carbon blooms in Africa has been termed ‘the African direct steel process’ (Avery and Schmidt 1996a, 1996b; Merwe and Avery 1982; Miller and Merwe 1994:7.fn.1). Schmidt claims that the key to success in the African iron smelting process is the maintenance of high temperatures in a reducing (CO) atmosphere (Schmidt 1996b:324). He points to the difference and superiority of the Tanzanian Haya process when compared with European bloomery and other bloomery processes in Africa (Schmidt 1978:243, Schmidt 1996b, 1997 and cf. bibliography). However, critics of Schmidt argue that ‘the African direct steel process’ is a normal variant of steel production in the bloomery furnace and that the term is redundant (Miller and Merwe 1994:21).56 Miller and Merwe even question the practical validity of a distinction under furnace operating conditions between the direct and indirect methods because both sets of reactions may take place in volumes with differing atmospheric conditions in the same furnace (ibid.: 21). To conclude, Miller and Merwe believe that the distinction between the direct and indirect methods stems from the historical distinction in Europe between the production of bloomery iron (produced directly) and wrought iron (produced indirectly), through a decarburization of cast iron (ibid.: 22). The bloomery process is far from simple and the general distinction between indirect and direct smelting methods in the African context is as suggested not to be sustainable (ibid.: 21-22).

than smaller ones. Sutton has argued for a difference in furnace construction and especially the size of the shaft as corresponding to the quality of the ore smelted (Sutton 1985:116). Killick suggests that smelting iron in small furnaces was more difficult than in bigger ones (Killick 1990, 1996). Perhaps then the quality of the ore, as suggested by Sutton, actually points at this variation (Sutton 1985:116). The ore in Ufipa and adjacent regions such as Ulungu and Unyiha all seem to have the same low iron content, basically a limonite or bog ore. The Pangwa and their neighbours in the Southern Highlands Region reduced a mixture of hematite and magnetite in their small furnaces. Both these views of size and quality of ores seem relevant to consider in my Pangwa and Fipa case studies of low and high shaft furnaces. Although some bloomery furnaces were especially adapted to low grade iron ores others were constructed to take into account the flexibility in the iron that was produced to make specific iron objects and tools (Miller and Merwe 1994:23). Leaving aside the possible faulty distinction between bloomery and blast furnaces in the African context blast furnaces were developed in China about the fourth century BC and thereafter steel and wrought iron were made in China by decarburized cast iron, as in modern industrial steel making. In general terms all bloomery furnace technologies were developed before blast furnaces while the bloomery process is an ancient technique for producing metallic iron from iron ores (Schmidt 1997:110). Blast furnaces did not reach Europe until the twelfth or thirteenth century AD, but thereafter they rapidly displaced the less productive bloomery furnaces (Needham 1980 in Killick 1990:19-20). In 1900 the last bloomery furnace within the industrialised world stopped producing iron from furnaces that had been in operation in the Adirondack Mountains in New York (ibid.). By the 1930’s bloomery iron smelting was confined to sub-Saharan Africa (ibid.: 19). Most subSaharan bloomeries were abandoned in the late 1950’s or early 1960’s.

Schmidt, on the other hand, has demonstrated that our Western based assumptions and divisions between metallurgy as a science and indigenous African metallurgical traditions are still far from completely understood (Schmidt 1996b, 1997). Again, I believe we are reminded of some of the limitations in our European based model and ‘Standard view’ or way of thinking about technology and specifically pre-industrial metallurgy. Yet only continued research will illuminate its weaknesses and possibilities. Within an archaeological context furnace morphology might be a better indicator for variability than technological or chemical descriptions and explanations. This issue, in terms of stylistic aspects will be further dealt with below, but there are problems related to the classification of furnace types. First of all, at archaeological sites little of the furnace super-structure is available for analysis, therefore, it is difficult to determine whether the furnace was originally a bowl, a shaft with a bowl or a natural draft furnace is difficult to determine (cf. Appendix D, figure 1).

In Tanzania, the Pangwa most likely stopped regular iron smelting in the mid 1950’s (Stirnimann 1976:264), while among the Fipa there have been recorded active smelters until early and mid 1960’s (Willis 1981:278).57 In 1983, as announced in the newspaper Fraternité Matin, African iron smelting came to a complete stop when the villagers of Koni in northern Côte d’Ivoire abandoned their bloomeries (Eckert 1976, 1986 in Killick 1990). Bloomery iron was typically wrought iron or low-grade steel, in Africa commonly used for axes and hoes (Schmidt 1997:112), and numerous prestige objects (Kriger 1999:234). The multiple metallurgical variations

An interesting point made by Tholander is the possible correlation between the size of iron ore stones (ores) and furnace super-structure, i.e. the height of the furnace chamber (cf. Tholander in Childs 1991b). If this is the case, tall furnaces would refine larger pieces of iron ore 56

57 Mgina suggests that the last furnace in Upangwa were in operation in 1949 (Mgina nd.) In Ufipa two of my informants claimed they gave up regular smelting in the late 1940’s and in the mid 1950’s (cf. Wrigth 1982).

Cf. also Eggert 1987; Killick 1996; Rheder 1987, 1996.

41

Randi Barndon in how to actually produce a bloom are questions beyond this study. Schmidt notes that ‘in the bloomery process the iron ore is heated with charcoal in a furnace. As the ore descends slowly through the furnace stack, it encounters temperatures and levels of carbon monoxide that allow a series of chemical reactions to occur which reduce (remove the oxygen from) the iron oxides. Eventually some of the iron oxides combine with other minerals to form a viscous, glassy material called slag’ (Schmidt 1997:112). The iron oxides are reduced in the presence of carbon to make metallic iron that settles out of the slag, forming spongy masses of iron interlarded with slag, called a bloom’ (ibid.: 110).

Most members in iron using and iron producing communities had (in the past) no means to explain iron technology in metallurgical or chemical terminology. Their notions of the metallurgical process in terms of the Western based scientific language as it developed over time had nothing to do with their work per se (Collett 1993:499). The master smelters were masters of metallurgy, but they were working by memory and operation, never by formal scientific experimentation. However, since the practice of iron smelting in preindustrialised societies expressed one of the most advanced and spectacular technologies and hence mysteries within their settings, the technology and those who mastered it were given special attention. They were persons with power and prestige although I will argue that their status was often ambiguous and ambivalent. What could be a more impressive technology than metallurgy, than the knowledge of iron smelting, among all the processes of material transformation?

First of all it is important to emphasise that smelting iron is a delicate and difficult matter. It is not something that one can easily go about copying after mere observations, something experimental archaeology has indeed illustrated (e.g. Childs and Schmidt 1985; Reid and MacLean 1995). If seen with the eyes of a metallurgist an iron smelting furnace has two main functions; it must both produce and contain high temperatures sufficient to render the slag fluid. Furnace temperatures are raised by admitting more air into the furnace, thereby increasing the rate of combustion to the fuel. Too much air will produce more CO2 than can be reduced to CO. Without an atmosphere rich in CO metallic iron cannot form and the smelter will be left with useless slag (Killick 1990:16). By merely looking through a hole in the furnace wall, as in the Fipa peephole, ntaanda, or the Pangwa’s ‘windmaking hole’, kwibepo, the master smelter would recognise the colour of the fire in the chamber. That was essential.58 His experience would tell him if the needed colours and thus temperatures had been achieved. If he had managed to keep the temperatures controlled he would succeed in producing iron, if not the smelt was a failure.59 The craftsmen had no notion of phosphorus; or the relevance of the amount of it, something first discovered in the early nineteenth century (Rostoker and Bronson 1990:2). Nor did the craftsman need to know the content of carbon within the furnace chamber as measured and described by metallurgist.

4.3 Description of iron smelting technologies How are we to describe technology or metallurgy and what is the nature in the fundamental qualities of iron working technology? What is actually happening when smelting iron ore to a bloomery iron, in a shaft furnace? How can we depict it and how can we understand it? Similarly to cultural and social aspects of technology, the natural has several levels of understanding and ways of being transcribed. The chemical process, reducing iron ore to iron, as depicted in: Fe2O3 + 3CO → 2Fe + 3CO2 is general and uniform. It is regardless of context (e.g. Kense 1983; Rostoker and Bronson 1990; Schmidt 1983, 1996b; Wertime 1980) and the formula itself represents a truth because of scientific results. It is like the discovery of DNA that Lakoff and Johnson at least at the moment, assure us is unquestionable (Lakoff and Johnson 1999). But what is the relevance of a formula that describes certain chemical reactions taking place within the furnace chamber if we are concerned with a non-western or prehistoric context in which iron was also produced?

In the past failures to smelt iron were most likely partly a neglect in maintaining this delicate balance and could happen even to skilful and experienced masters (cf. Barndon 1992; Killick 1990; MacLean 1996; Schmidt 1997).

In its widest sense a Standard view of metallurgy may of course assist us in our classifications of furnaces, their operation and efficiency in iron smelting. Additionally, final products from various smelting activities such as bloom, bloomery iron, steel and slag can be classified in accordance with a sound metallurgical knowledge. Final products from a smelt can be classified in terms of how the smelters’ use of magic during a smelt imposed on and effected the chemical reactions themselves or not. For instance, the Fipa smelters use of ingailo, medical and magical barks from various trees always cut from both the east

58

Aremu suggests that the holes at the back of the furnaces at Yankari in Nigeria allowed fresh air into the furnace while carbon dioxide escaped through the top of the furnace (Aremu 1999:206). This of course also serves as a plausible explanation for especially the kwibepo on the Pangwa furnaces. 59 Mgina (n.d.) reports that the Pangwa smelters frequently experienced failures in which the result was a useless mixture of iron and slag called ‘the burning of the iron’, ishuma khipye, referring to too high temperatures. As they also told me it became ‘too hot as the heat in anger or in cooking’. The master smelters claimed this never happened when they were around, they could control the fire and heat with the speed used in blowing the bellows.

42

Re-enactments of smelting, how to describe and transcribe? and the west side of the trunk of the tree was from their point of view necessary magical assistance (Barndon 1992).

processes’ (Lemonnier 1992:5). Thus, moving again from observation to transcription (and towards my case studies), one can say that each successful smelt shares a basic set of operations that at least include the following:

Equally important were the creepers and beans put in the medicine pot, ing’anjo, that was buried beneath the Pangwa furnace (Barndon 1996b, 1999a). But from a Western metallurgical baseline the same magical barks, ingailo, and perhaps even the Pangwa’s ing’anjo may be seen, if thoroughly investigated, to actually establish an effect that contributes to a higher carbon content in the combustion chamber during ore reduction. A more accurate understanding of this can only be established with the use of scientific methods developed within metallurgy in a combination with stronger efforts in observations and documentation of magical aspects within a smelting process. Again, my point is that in our context, for the craftsmen as for us, the Western scientific baseline that we have had towards technology and more specifically metallurgy is not really essential. Of significance is how metallurgy was indeed practised, conceptualised and finally materially manifested. One way of understanding this is through a focus on magic, on symbolism and the pre-eminence of the master smelter. Lemonnier argues that every technique (and technology) has five related components.

(1) selecting a smelting sites in relation to resources and villages; (2) selecting and processing clay and possibly other materials to build a furnace; (3) selecting, mining, processing, and transporting iron ore to the smelting site; (4) choosing trees, making charcoal fuel from them and transporting the charcoal to the smelting site; (5) performing rituals prior to a smelt, often while building the furnace structure; (6) providing a means to bring air into the furnace to stimulate combustion, often tuyeres and bellows (7) conducting the smelt, often with accompanying rituals; (8) concluding the smelt and distributing the bloom at the end of the smelt (reuse or not of smelting site) (Childs 1991b:340). MacLean suggests that the forging process can be summarised as follows:

(1) matter (the material and ones own body of which a technique acts, (iron or clay); (2) energy (the forces which move objects and transform matter); (3) objects (or artefacts, tools or rather means of work that are used to act upon matter); (4) gestures (which move the object involved on a technological action);60 (5) specific knowledge (may be expressed or not by the actors, and may be conscious or unconscious) (Lemonnier 1992:5).

(1) choice of trees and manufacturing of charcoal; (2) provision of means to bring air in to the forge; fire, tuyeres and bellows and; (3) forging tools, such as hammer stone and stone anvil and tongs (MacLean 1996:30). In all sequences; minor or major, such as the suboperations within a minor sequence or major tasks, such as charcoal burning, specific knowledge, unconscious or conscious, was the know-how and manual skills and experiences that each participant contributed to through the choices and possibilities that lay within the technological activity. The above listed steps in the processes of smelting and forging iron provides a useful reference point for my Pangwa and Fipa descriptions and transcriptions of smelting technology. Each of the operations or sequences may if following Lemonnier’s chaîne opératoire approach be subdivided into suboperations and aggregated into the amalgam that constitutes iron smelting as a technology (Lemonnier 1992:6).

According to Childs the technical requirements other than the basics of the nature of iron technology necessary for a successful smelt are raw materials; iron ore, fuel (charcoal or green wood)61, clay (for furnaces and tuyeres) and skins, clay, bark and wood for bellows if used, a reducing atmosphere, provided by the furnace construction, fuel and temperatures around a minimum of eleven hundred degrees and finally labour, the workers (Childs 1991b:340). In his anthropology of technology Lemonnier has demonstrated that gestures ‘are organised into sequences which may be subdivided into sub-operations or aggregated into operation and then into technological

As discussed above iron working can be divided into two sections, smelting and smithing (Barndon 1992:68). I will however emphasis the distinction between the smelting and the forging process and the significance of ready used objects deriving from the raw material transformation. In my opinion there is a difference in smelting and forging in terms of ritual and magical elaboration and significance that should not be ignored. Blacksmithing follows naturally from a smelt and is a less complex procedure. It

60 ‘Gestures are organised into sequences which, for analytical purposes, may either be subdivided into ‘sub-operations’ or aggregated into operations and then into technological processes’ (Lemonnier 1992:5). 61 cf. Schmidt for a description of the Haya smelters significant use of ishanga grasses in addition to charcoal (Schmidt 1997:67).

43

Randi Barndon a hoe (Wembah-Rashid 1966:68). During my stay in Ufipa we discussed old forging practices and visited several old abandoned forges and my informants confirmed preliminary rituals when a new forge was built and that women were not allowed entry. In Upangwa I conducted only preliminary surveys of old forges and was not able to get information about past forging practices. Stirnimann seems to have experienced, like me, that forging was less ‘ritualised’ (cf. Stirnimann 1976:288).

does not involve the same degree of decision and is less likely to require a specific choice of time. In addition the forge is generally placed at a permanent and recognised location within the community (Barndon 1992; Brown 1995; Herbert 1993; MacLean 1995). It is not reconstructed preparatory to each smelt. It may be forbidden for uninitiated to enter the forge or as commonly reported it may be a place for people to meet, exchange news and gossip (e.g. Herbert 1993, Kriger 1999).

Specific technological knowledge is made up of ‘knowhow’ or manual skills, and associated with the two most important components; possibilities and choices. In sum this is what Lemonnier calls social representations of technology (ibid.: 6). Technology as a social phenomenon exhibits many systemic aspects (ibid.: 17). The features of technological systems are not the simple result of physical constraints, either constraints internal to the technologies themselves (the metallurgy itself), or constraints arising from the natural environment (one may for instance chose to use or not use a raw material available in the environment) (ibid.: 6). But why? The question of the influence of social choices has to be seriously raised (ibid.), and the social is as relevant as the environmental, the natural.

The emphasis on preliminary rituals in the smelting process was possibly connected to the fact that smelting was far more difficult than forging. If something goes wrong with a forged object it can merely be reforged without too much physical struggle. Forging was seldom ritualised and seldom in secret, it was more an everyday practice that was open and available to observe for all people in the village. However, the forging process with its integrated preliminary rituals is far from an exhausted study. Recent studies have proven that although symbols were not explicitly evoked during forging they were part of an embodied philosophy in forging as in smelting iron (e.g. Bekaert 1998; Brown 1995). Wyckaert notes how the Fipa in the past used to sacrifice a cock before a new smithy was used. The cock was sacrificed within the forge and the blood was sprinkled on the floor. Anybody who wanted to buy a hoe was allowed to help in the smiths as long as they did not arrive ‘empty-handed’. They had to bring for instance a pot of beer or some food to the smithy (Wyckaert 1914:378). Cline commented on Fipa initiation of a new forge that ‘before using a new forge, the Fipa smith killed a cock and sprinkled its blood on the anvil and said that this forge may not spoil my iron, that it may bring me riches and fortune! He then chanted a prayer to his ancestors, cut off the chicken’s wings and hung them on the frame of the hut as a protective charm. After this, he and his professional friends ate the chicken’ (Cline 1937:122).

Even though technologies, even highly sophisticated ones, involve many arbitrary choices based on not only high level systems of meaning in the society but also within a more individual frame, one must never depart from the basic purpose of any kind of technological behaviour. What is really at stake in material production? Is it to produce meanings and symbols? Is it in material production not first of all to produce raw material? In line with Bekaert I believe that what the iron smelters or a smith above all wanted was to make iron (Bekaert 1998). It is as simple as that. But the important means to achieve iron production are not only tied to the environmental or natural aspects, nor the practitioners need for more efficient iron implements, they are equally connected to social, magical and basic mental aspects. These aspects derive both from the actual experience and practice and from the significance of the products produced, the blooms and the iron objects. The metal they produced and the iron they forged was a thing of value. Any kinds of aspects of the tasks, technical, metallurgical or decorative in its orientation were often not fully conscious for the participants. In sum, aspects both unconscious and conscious were made into expressive systems and therefore can be compared, the basis of all anthropology, including archaeology. Lemonnier points at the counter productivity of forcing technological phenomena into predetermined categories of importance. Metallurgical based categories might be misused and force our thinking in the wrong direction. What we want is to develop a theoretical and methodological approach that can combine and deepen our understanding of both technical and cognitive-symbolic behaviour in African iron working.

In 1936 when Greig observed iron smelting in Ufipa he believed many of the ceremonies and rituals once practised by the smiths as well as the economic profits of the work had disappeared (Grieg 1937). Willis records that when the iron has cooled it was taken to the forge, ipeembo, where it once again was heated on a charcoal fire (Willis 1966:25). In the forge, the master blacksmith was still called umwaami and he supervised two other men using the hammer stone (ibid.). Willis continues that the Twa chiefs or any from their family were forbidden at the place, while similarly to Wyckaert, he notes that others were allowed entry if they paid tribute to the smith in acknowledgement of his authority (ibid.: 26). Finally, he notes that while engaged in iron smelting and at the forge the smiths had to refrain from sexual activities and women were forbidden near the forge just as they were forbidden near the smelting furnaces (ibid.: 26). Wembah-Rashid does not even comment on ritual or symbolic aspects in his brief description of the forging of

44

Re-enactments of smelting, how to describe and transcribe? 4.4 Transcription of iron smelting technologies

constantly be questioned.

Although it is true that post-structuralism once was an uncertain term (Hodder 1989:68), it seems that because of its lack of novelty it has now, without being properly investigated, partly moved towards its end within much of contemporary archaeology. Its importance has been advocated by, among others, Moore (1986, 1990, 1995), Olsen (1987, 1990, 1991, 1997), Tilley (1990, 1995) and Yates (1990). It has been connected to contemporary philosophical thinking as represented by Foucault, Barthes, Derrida and Riceur who all have been closely linked to post-processual theoretical archaeology (cf. Olsen 1990, 1994, 1997). The philosophy and the ‘school’ has however suffered extensive critique, misreading and misunderstanding of the actual messages that its advocates tried to convey (i.e. Latour 1996:20, and see Trigger 1990 for an example). On the other hand I believe that, as demonstrated in the above discussion, we must remain conscious of how and what we write and who our readers might be.

Anthropology and autobiography merge in the process of knowledge production (Hastrup 1995), which implies that ethnoarchaeologists must try to be open about field methods because they lead to the actual writing. How then to fit in a scientific mode of thought, an approach that is commensurable with our personified field experiences, the multivocality in and among others? How can we produce a kind of science that takes into account our way of thinking and way of doing ethnoarchaeology, while at the same time we conduct research and continue to write prehistory? However aware we are of the problems of representations and translations we cannot escape the fact that we have to reduce the magnitude we observed in the field (Barndon 1996; Hastrup 1992, 1995; cf. also Latour 1996:147-149). This we simply have to do in order to move from experiences to analytical results. We have to make categorisations and classifications (e.g. Lemonnier 1992:32; Sigaut 1994), and further also make reductions from our empirically observed data to theories in order to produce new knowledge (Hastrup 1992:10, 16). To somehow make meaning out of what we have observed see figure 4.1 (below).

Because of the character of participant-observation as field method (Comaroff and Comaroff 1992:8), it is however difficult to move from the personal towards the analytical. The question here is of course how to translate personal experiences into accepted academic concepts. How to combine experiences with theory, interpretations with writing? How to manage to be both honest and selfreflexive and contribute with new analytical results? How can we reach a better understanding of the past through a study of contemporary cultural processes? Our situation is very similar to Barth’s early formulated inquiry into the social anthropologists’ responsibility of being practitioners of a human discipline. No actual solution on how to conduct and produce an ethnoarchaeology that is self-reflexive and provides a science that is applicable to the archaeological record has ever been presented. There is no emerging work within ethnoarchaeology that suggests how we can develop a methodology that allows us to employ contemporary peoples and their living traditions in studies of the past. This is because the search for archaeological relevance has implied that it is the observations themselves that are important, and that to include real people and their lived experiences would reduce the authentic value of the data collected. The reductions or classifications we make must however (a: explanation)

Hastrup’s model simplifies the relation between what we want to explain and the actual explanation itself. This is basic, but first of all it illustrates how ‘science’ works through a temporal distancing. Temporal distancing is a necessary tool if what we want to do is generalise about the hypothetical (Hastrup 1992:16-17), and if what we want to do is compare in order to see social processes. This model, in a direct manner, illustrates the problems we are faced with when returning from fieldwork on ‘our dream atoll’ (Tilley 1999). It illuminates the actual reduction process of distancing and writing which is necessary in order to continue to research (see also Latour 1999:38). This is when, whatever time and space we believed ourselves to be in, we reduce it all to ‘a here and there’ and ‘us and them’ writing in the present but about, or for the past. Thus, our observations and data are theory - and value laden while they should also fit into hypotheses or models to be of any value (Rorty 1991b: 13, in Hastrup 1995:13). If a study of various cultural traditions and cosmologies/worldviews associated with iron working is to have any archaeological significance or

(b:factors/phenomena that are being explained) x1 x2

x x3 x4 Figure 4. 1 The problem of scientific distancing (after Hastrup 1992). 45

Randi Barndon relevance, the empirical must be translated into terms of a more general language. According to Hastrup this kind of reduction is not positivistic rather it is creative (Hastrup 1992:17). In spite of this general awareness within the social sciences such as Appadurai’s discussion of ‘natives’ and ‘places’ (Appadurai 1988), the practice of doing ethnoarchaeology and the formula for the writing of ethnography for prehistory has still to be seriously considered.

treatments and distributions of the products. I also tried to determine the reasons underlying potters technical choice, first by questioning them about the origin of their knowledge, the way they selected raw materials, or the way they controlled the issue of the different technical operations, and second, by evaluating the performances that characterised each transformation process. To this end, a series of measurements and experiments were made’ (ibid.: 85-87).

Lemonnier suggests the study of chaînes opératoires as a method and an approach for grasping possible social representations in technologies. As Sigaut has discussed at length (and illustrated in his diagrams of for instance bread-baking) an operation is identified by its location on a production path or network (Sigaut 1994:420-430). Identifying an operation means locating it in both a physical and social space. When two operations have been found to occupy the same places in both, they are ‘homologous’. It is possible then to compare the various ways in which these operations are performed by different (or sometimes the same) group of people. On the whole the ways will be different, and it is to these different ways of carrying out homologous operations that we apply the term techniques. Our way to the systemic aspects in technologies is through a real encounter with the process.

In an illustrative diagram in the same article Gosselain depicts the main stages for the south Cameroonian potters in their completion of the main stages of the manufacturing of a pot (ibid.: 86). In these diagrams several phenomena are missing or at least given less attention than they should if indeed we accept technologies within a cultural and specifically semantic frame. Both Lemonnier’s description of the carpenters’ use of a second nail (Lemonnier 1992:33), or in a different context; the frequency and ways of stirring brine in French salt-making (ibid.: 44), as opposed to but also compared with, steps in salt production in New Guinea as depicted in another diagram (ibid.: 49). And finally, Gosselain’s transcription of the Cameroonian pottery making miss meaning constructions. These ethnographic studies of operational sequences in technological processes show less emphasis, possibly apart from Gosselain’s post firing processes, on the symbolic choices and chains that also constitute parts of the process.

In the case of African iron working, an encounter with the smelting process will reveal what is, again as Lemmonier frames it, ‘relevant to observe and document for a comparison’ (Lemonnier 1992:29), in my case of smelting processes. The data collection, the participant observation, in the real encounter, involves a detailed step by step transcription of the technological process. Therefore, as a method the step by step transcription can assist our comparison of technologies and even bridge the gap between then and now, past and present. It is by comparing transcriptions of operational sequences that anthropology of technological systems will take shape (ibid.: 30). The next step in the process, the transcription, entails a different attitude because in order to transcribe the accurate encounter one must provide the graphic descriptions. The graphic description of, or rather the diagram of chaînes opératoires among carpenters as their individual and specific use of hammer and nail, can be studied isolated or compared, to grasp social or cultural significant traits in the carpenters’ technological behaviour (ibid.: 31-33).

In Sigaut’s discussion of step-by-step methodological descriptions of technologies, such as bread-making, beermaking or parboiled rice-making meaning and technological choices are also missing (Sigaut 1994:42630). This is peculiar since Lemonnier as well as Sigaut’s starting-point is as they claim; social choices and meaning in technology or the cultural content in technology. Lemonnier’s approach as well as Gosselain’s application of it, lack two important archaeologically relevant dimensions; the use of space and an elaboration of how symbolic behaviour relates or is connected to technical performances. Adding these will deepen our understanding of spatial and material manifested dimensions in technologies. Because, as I argue, these are what is culturally meaningful and socially structuring for the iron-makers in their iron producing process. Lemonnier’s lack of spatial and material manifestations of technological behaviour might be explained by the fact that it is ‘the ethnographic present’ that he has mainly been concerned with. He does not take into account more specifically how his approach can be linked with traces of prehistoric technological activities. If incorporated into an approach that is concerned with the past but that also value technology as social and cultural, the location of the activities constitutes one aspect of the actual choices in the chaînes opératoires in itself. It is essential to establish why a certain setting has been decided upon whether it is chosen because of its vicinity to raw materials or its

Gosselain following Lemonnier’s comparative approach has developed a way of moving from encounter to transcription based on his ethnoarchaeological study of West African pottery making (Gosselain 1998). He describes his own research methods on pottery production as follows: ‘In the field my first aim was to carefully observe the techniques the potters used at the different chaînes opératoires, from resource procurement to post-firing

46

Re-enactments of smelting, how to describe and transcribe? distance from village life, or both.62 These aspects should be given more attention. Not only how and when but where each chain of actions is organised (cf. Sigaut 1994).

know that there were several types of experimental processes occurring during his field studies. First of all those conducted by anthropologists and metallurgists (his collaborators and himself), concerned with unwrapping the technical ‘mysteries’ of iron production and their material correlates. Secondly those conducted by the iron smelters who as technical and ritual bricoleurs successively pieced together various combinations of material and ideological ingredients to arrive at different material results’ (ibid.). Schmidt’s argument is based on a critique of the Western baseline of making a division between science and magic. His thoughts and criticism are in line with Tambiah (1993) and he attempts to break down the division between the two domains. Another important point in Schmidt’s monograph is that; ‘his meeting with, interaction with and study of Haya and Barongo technological performance has created something different from what someone would have observed in 1880 if peeking from behind a bush’ (Schmidt 1997:12). Thus, he openly emphasises that what he reports must not be taken to represent the fabric of iron smelting as it would have appeared in 1880 or 1900 or even 1920 (ibid.). However, he adds, and I agree, that ‘these limitations should not come in our way for a belief that there is still much to learn from these old men’ (ibid.).

In Pangwa and Fipa iron working tuyeres were not made at the smelting site. A different and independent microcosmos was created and constituted another aspect of the ‘technology’s totality’ or ‘path’ to use Sigaut’s words (Sigaut 1994). Besides this it yields another material manifestation. Before my observations of the Fipa smelting process very little information about the actual sequences in the making of charcoal or in tuyere making had been published (but cf. Wembah-Rashid 1973). Apart from one photograph taken by Wyckaert (1914:373), those who had observed Fipa iron working had not commented on the actual operational sequences in this part of the process. They had not commented on other specific preparatory tasks such as charcoal productions or initial slaughtering of goats for the skin bellows (cf. Barndon 1992). The smelters never told me in interviews before our re-enactments how difficult the making of tuyeres was or that in the past specialists were required. This exemplifies the importance of actual encounters with all stages in the process (Sigaut 1994:426). Schmidt, in a similar setting, describes how the Haya did not express verbally the importance of the pre smelt/roasting practice but that he after observation, could establish what the smelters actually did was to pre-smelt the iron before the refining process (Schmidt 1997:72-73).

Even if we cannot reconstruct prehistoric or more recent African smelting contexts, what we can construct can be used to build foundations for a deeper understanding of the smelting process and aspects of associated or evoked meanings. Hence, we can refine our concepts and approaches. For instance these studies can access data on the materiality of symbolic repertoires and indicate cores in (or key symbols) in both social and individual memory as part of the process (cf. Bekaert 1998; Connerton 1989). My fieldwork, among Fipa and Pangwa smelters, is similar to Schmidt’s description of his own work in Buhaya; based on a co-operation between me, my assistants (interpreters, field assistants, drivers etc.) and the local iron workers. The smelters have demonstrated their skills and past activities to me. In their performances they have recollected images from the past and performed their rituals and practical tasks in accordance with memory. Therefore, an experimental approach has formed my setting.

Systematic recording of the chaînes opératoires in decoration would most likely give answers or at least shed light on some of our essential questions and using David’s words, how ‘technical and symbolic aspects inform upon each other (David 2001:1). When and where are rituals and use of magic, the mental and material decorative actions evoked? Are ritual performances in opposition to other tasks and their operational sequences? Do furnace decorations and rituals take place at a time crucial in terms of the metallurgical process, the chemical reactions? Does the use of magic effect the chemical process? Is the symbolism conscious or not? In addition to Lemonnier’s chaîne opératoire approach my point of departure and theoretical baseline will be combined with a stronger focus on embodiment and bodily based metaphors. I have tried to follow this baseline and this focus in my observation, description and transcription of Fipa and Pangwa iron working.

4.5 First meetings and the ‘setting’ for research David is among the few to propose a guideline of what to include in publications of observations of traditional iron working (David 2001:25).63

Schmidt grasps some of the most essential aspects of ethnoarchaeological fieldwork when he calls his investigations of Haya and Barongo iron working an experimental approach (Schmidt 1997:11, cf. also Tringham 1978:171). He clarifies aspects of the field situation and data collection by letting us (the readers)

In a discussion of the research context and of which aspects that we should inform our readers about, David emphasises location, cultural and historical context of group and groups studied. Furthermore duration,

62 A wish for secrecy or seclusion does not contradict the need of being close to raw materials.

63 Schmidt on the other hand incorporates many of these aspects of fieldwork in his work among Haya iron smelters (Schmidt 1997).

47

Randi Barndon investigator’s knowledge, including linguistic competence, of the group and its language, and also political context of research, including situation of the researcher, and funding should be reported. Within his anthropological methods and techniques, sampling of group such as settlements, households, individuals, classes, ranks, statutes, roles and gender are considered important. His views on field research on iron smelting includes the amount of participatory observation, interviews, and questionnaires. Other factors are methods including re-enactments and experiments, assistants and their qualifications and situation vis-à-vis informants, the actual informants, categories and how recompensed, recording techniques such as visual and sound recording and the written materials (ibid.), which I believe refers to diaries or notebooks. Especially the assistants’ qualifications vis-à-vis the informants is crucial for the outcome. However, so is are the initial ‘sampling of the group’ of informants. Both these aspects affect the actual information gathered and are crucial during the data collection in the field. David’s third group, scientific methods and techniques, have no aspects made explicit since this belongs within the approach itself. This has been presented in the chapters above. Below I will take into consideration my own setting and the circumstances surrounding my fieldwork in Ufipa and Upangwa.

Between 1990 and 1991 four field trips were conducted among the Fipa on the Ufipa plateau in the Rukwa region of southwestern Tanzania. Three separate field studies have been carried out among the Pangwa in the Livingstone/Kipengere Mountains during the years 1993 and 1995. Part of the material presented here has previously been treated in both published and unpublished manuscripts (Barndon 1990, 1991, 1992, 1993, 1996a, 1996b, 1996c, 1998, 1999a, 1999b). However, in recognition of the general post-modern critique of ‘writing culture’ (e.g. Clifford and Marcus 1986; Olsen 1990; Vansina 1990), I have found it worthwhile to reconsider some of the documentation and the way ‘traditional’ iron-working was then presented, especially the more idealised description of my Fipa study (Barndon 1992). Whilst much of the present work therefore builds on my previous research among the Fipa iron workers (Barndon 1992, 1996a, 1996b, 1998), only more brief accounts of my Pangwa fieldwork have been published (Barndon 1996b, 1999a). I believe an open discussion of the field situation is necessary. Moreover, frequent reconsideration of the actual work conducted may constitute an important source of information and may refine our conceptual approach and transcription of experiences. Before my fieldwork among both the Fipa and the Pangwa I had visited them and established contacts and plans for our later re-enactment of smelts. During these preliminary trips we also visited old abandoned smelting sites and discussed the entire processes at the sites. We settled how much charging (ore and charcoal), how many workers that would be included in our team and the amount of beer, pombe, and meat, nyama that we needed.65 I always paid for all expenses as well as costs in connection with sacrifices or magical consultations that were part of the task.

Lemonnier still provides the most detailed discussion of what a would-be technographer should try to do while in the field and afterwards (Lemonnier 1992:27; Sigaut 1994:423). I will add all archaeologists concerned with the study of material and tool production among living people to the list of technographers. First of all Lemonnier suggests a division of the research process into observations, description and finally transcription (Lemonnier 1992:27-29, cf. also Sigaut 1994:429-430). He discusses how difficult it is, and at times even disastrous, for the technographer, for the observer to try to perform the same technological actions that he wants to describe (Lemonnier 1992:27).64 Nonetheless, participation might allow the researcher to gain better insight into the actual process, what is actually going on. There are cases in which it is valuable to try to do the things oneself to gain a little familiarity which can help one describe the process (ibid.: 28). This is similar to Aristotle’s point that builders only become builders by building. ‘But if this participation entails breaking of things, boring the informants or seriously interrupting a crucial activity, it is perhaps as Lemonnier suggests better not to try to be a potter, a hunter, a salt-maker or whatever’ (Lemonnier 1992:27), and I can only add ironmaker. This does not mean that one should give up efforts to be accepted or to get on ‘the inside’ (cf. Barndon 1992), because becoming a familiar element in everyday village life will facilitate the observation and most certainly yield paths that lead to information on related aspects of culture and social practices. 64

In the past, the knowledge of smelting technology was never transmitted orally. It was practised and demonstrated by using different levels of embodied and lived experiences, of moral spaces and symbolic structures. Open discussions were no part of their traditional transmittance of knowledge (Schmidt 1997). I agree with Schmidt that interviews alone are not valuable in studies of technological practices (ibid.). Still, they are important because they point at the differences in various forms of transmittance of knowledge and the role of secrecy in specific handcrafts. They may demonstrate discrepancies that occur between what people say and what they do. In Ufipa, the participants and their individual skills were not discussed before the actual re-enactment while more general pre smelt interviews were many. We agreed that the master smelter should call together a group of 65 Pombe is a Swahili word for a maize brew drank all over Tanzania. The Fipa called this beer commoni while the Pangwa call it pombe. Nyama means meat in Swahili.

Italics by the author.

48

Re-enactments of smelting, how to describe and transcribe? smelters, mainly those working as blacksmiths in the village forge and others of his relatives. Later I was told that the group of smelters and smiths as well as old skilful participants were actually members of the master smelters’ clan, sinkaamba.66 During our first meeting with the master we were told that women were not allowed to participate. This had been very important in the past and although this was a re-enactment the same female exclusion was upheld.67 In Ufipa my assistant was often taken aside and informed about matters that women were not supposed to know about but he was also told to explain to me later when we were alone. Then we could sit down and discuss the matter (Barndon 1992:30). How many filters the explanations went through I will never know.

were interested in their exotic life, magic and medicines, in order to use this knowledge to suppress them in one way or the other. They were worried that their use of magic, sacrifices and medicines would be used against them. Others again were more optimistic hoping that my research into old smelting practices would result in a revitalisation of iron smelting and perhaps even a factory or paid labour that they could themselves participate in sometime in the future. In Ufipa I sensed that the smelters as well as others not directly involved in our project were moving towards a revitalisation of the old smelting practices and beliefs (Barndon 1992). In Upangwa, four years later and among a group of people not so famous for their past time smelting skills as the Fipa, a similar revitalisation occurred. Sometimes I was faced with informants reminding me of ‘Muchona the Hornet’ (Turner 1967:131, in Barndon 1992), because some were more than willing to inform me, giving me explanations and interpretations. There was a competition running in the village about who was my most trusted confidant and my most important informant, while others were not at all interested in being interviewed or questioned about the actual work. If I tried to interview them they said ‘wait and you will see as our work progress’. I was however gradually able to figure out who (others than the master smelters) were truly skilful smelters by looking at both their pre and post smelt statements, their actual performances and how they formulated or termed the work, the process or the furnaces and smelting equipment. During my first visit to Ufipa I collected the most commonly used words or lexical terms used in the smelting and forging processes (Barndon 1992 and cf. Appendix B). This method of collecting terms and searching for metaphors in the smelting process moved me closer to the way the smelters themselves conceptualised the process. It also functioned as a control of previous statements about their way of conceptualising the technology since the same persons were interviewed several times. Participants who had no knowledge of iron smelting or had not been properly initiated in to the group around the master would not know the meaning of a specific term or when to use it. These inquiries made it possible for me to cross check their statements and to establish who had actual knowledge as opposed to an outsider’s opinion about the process.

During preliminary visits to Upangwa we discussed the entire smelting process with the people that we then met at the village forge. The village forge was used, although at intervals, by the twenty blacksmith of the village. Two younger blacksmiths from this forge agreed to search for participants in re-enactments at our return. Similarly to Fipa rules women were not allowed to participate in iron smelting, but a stranger was different. Hence, I was again allowed entry. During many of our interviews with both Fipa and Pangwa smelters I was told ‘wait and you will see’. The smelters’ answer to our inquiries into how things were done points at the important fact that many tasks and performances in the smelting process were not verbalised but based on tacit knowledge (Barndon 1992). Additionally, their entire tradition was based upon secrecy and restrictions in the transmittance of smelting knowledge. Verbal descriptions of technical procedures were no part of their tradition (Schmidt 1997:47). This might even be one of the hidden reasons behind some of my informants being sceptical about handing over their knowledge to a foreigner or even about demonstrating their skills and secrets. While others found it important for the same reason because they feared their knowledge would soon disappear, since it was not practised anymore. During my first stay in the villages and my initial meetings with some of the Pangwa and Fipa I faced problems concerned with the actual legitimising of my research. To some of the ironworkers it seemed incredible that any of their traditional ways of smelting iron could shed light on other or even prehistoric iron smelting practices. Although I made great efforts to explain my purpose, indeed their questions were relevant. After our work had started I learnt that some of them feared that we were planning to use their knowledge to, as they said ‘to get hold of the iron ore in their land’. Others thought we

The major approach in the field was uninterrupted observations of the preparatory stages and smelting procedures. This involved observation from initial preparations for the actual smelts such as collecting hardwood, charcoal production, ore collection, clay preparations and tuyere making. Observations also included furnace construction as well as observations of magical practices, ceremonies and rituals performed before and during each smelt. All the stages from collection of raw materials to the final forging of tools

66 Sinkaamba is the name of the uluko, descent group of the master smelter (cf. Willis 1967 and Chapter 9). 67 My own sex seemed unproblematic and irrelevant (Barndon 1992). During all my fieldwork in Tanzania being a woman was never really discussed as a problem and to my knowledge they did not find my observation difficult; I guess mainly because I was the one initiating the entire project.

49

Randi Barndon were observed in the field.68 Hence, studying Fipa and Pangwa iron production involved several methodological approaches; such as interviews at old smelting sites, interviews at the created sites and during re-enactment of smelts as well as interviews at the sites and at the smelters’ homes. Sometimes interviews were with groups of smelters or together with their wives. We discussed the sequences of the process and I arranged interviews with one or several participants at the same time. The sum of these methods forms the fieldnotes that I have collected on Fipa and Pangwa iron working. The point has been to produce as many stories as were actually experienced; events that were observed in the field, to see how or if they can inform about variations in the iron smelting process; influences, choices and furnace attributes. In sum these descriptions form multiple stories of how iron working was demonstrated to us as well as the fact that they give a more correct version of the smelters own perception of their tradition.

informants’. During my fieldwork among Fipa ironworkers my base was a village on the Ufipa plateau, 13 kilometres south of Sumbawanga. In this village of about 300 inhabitants lived mostly Fipa but also some Nyika and descendants of the Lungu and Mambwe.71 During the 70’s, in the spirit of the Ujamaa policy this village, originally called Mkumbi, was moved and re-established as Katumba Azimio closer to the main road leading towards Sumbawanga. The old abandoned village was only ten minutes walk away and surrounded by several standing or partly standing smelting furnaces. For our third experimental smelt in 1991, one of these old smelting sites, Tupa, was taken into use. This particular village had a group of active blacksmiths and many of them were the old master’s sons or other close relatives and clan members. This group of men constituted the most important part of the crew during the three smelting processes observed in Ufipa (cf. Appendix E). We never directly paid the smelters, weekly payments/wages but they were given food and various gifts. We were formally introduced by the District Commissioner and Regional Cultural Officers who informed the inhabitants in the village about the importance of our documentation of iron working. The Regional Cultural Officers direct involvement in our work and their encouraging of the smelters to demonstrate their skills and partake in experimental smelting was important for our outcome. I believe, having knowledge of the craft today gives some prestige among the Fipa. This also facilitated my work. Many Fipa see that their traditional handicrafts are disappearing and because of this they also loose a bit of themselves, their own identity as individuals and as a collective or social group. In the past they were intimately connected to their material culture, as they explained to me; ‘they made Fipa hoes, Fipa pots, Fipa baskets and Fipa cloths and lived in Fipa houses, drinking Fipa beer, isuute’ and so on. On several occasions they pointed out to me objects or behaviour that they conceptualised as typically Fipa. They were proud of their past prestigious occupation as skilled ironworkers. They were willing to inform, even about the secrets and taboos connected to the craft because they knew iron smelting skills and knowledge had partly lost its value and importance.

These stories will be compared and discussed in relation to other sources of information on Fipa and Pangwa iron smelting, such as travellers’ accounts and ethnographers’ documentation (e.g. Stirnimann 1976, 1979; Willis 1969, 1981) as well as information on neighbouring groups of ironworkers. In sum, these dimensions of iron working might introduce new aspects to our knowledge of African iron working past and present and may refine our approach towards technology. The initial visit to the Rukwa region and my first meeting with the Fipa people was in March 1990. According to formalities and custom in Tanzania the District Commissioner and the Regional Cultural Officer in Sumbawanga were visited (Barndon 1990, 1991, 1992). We introduced ourselves and informed the Government authorities about our arrival and research plans.69 I had the impression that the staff at the Regional Cultural Office were interested in my project because they immediately found time to guide us to a particular village where a famous master-smelter, umwaami, lived. I decided to come back to this village on my next trip to Ufipa because the master smelter had agreed that he would demonstrate his iron smelting knowledge to us.70 The master smelter, whom the RGO’s staff had introduced me to, became one of my so-called ‘key

The aged men were proud of their iron smelting skills and were even proud to be able to transmit their skills and knowledge to me. Friendliness and openness is a central part of being a Fipa (Willis 1981) and became crucial for the outcome of my research. The fact that our base became this particular village was no coincidence. The master smelter was known to the RGO’ officials and admired among the Fipa people in general. He had once been a prestigious master smelter, know beyond Ufipa. He was also well known because several researchers that had visited Ufipa in earlier times had based their research

68

In this study I focus on the smelting of iron and not the forging of tools. However, it is fully recognised that iron as transformed into ready forged items, tools and weapons are relevant both for social and cultural (contextual) understandings of traditional iron working and the role of iron objects made by the smelters. 69 By ‘us’ and ‘we’ I mean my assistant(s)/interpreter(s) and me. Cf. Appendix E for participants and members of smelting crews in the various re-enactments we conducted in Upangwa and Ufipa. 70 On our first visit to this village we came in a new land cruiser, we were escorted by a professional Norad driver and the staff at the Regional Cultural Office. After our survey we went back to Dar es Salaam. Several weeks later Bilham Kimati and I came back to the village. We were without a vehicle, we were walking and made the village our base.

71 The Fipa number around 100 000 while the Pangwa number approx. 60 000.

50

Re-enactments of smelting, how to describe and transcribe?

Plate 4.1 A Fipa blacksmith at work in the village forge, Katumba 1990 new umwaami.73

on his knowledge (e.g. Sutton 1985; Wembah-Rashid 1967; Wright 1980, 1982).72 When in 1990, we first arrived on the Ufipa plateau we went to his village and were invited within his house. We talked about our interest in African iron smelting and he showed us some of his iron heirlooms and blooms from when the Fipa of his clan were active iron smelters. We also visited an old furnace at a nearby smelting location. According to what he later told me; he felt that it was important that Fipa iron working as it had once been practised was carefully observed and recorded before completely forgotten. He trusted me for this recording and documentation. Since he feared that the knowledge he possessed would be lost with his death he had also been teaching one of his nephews, being among the active group of blacksmith in the village, what he called ‘the secrets of the craft’. Our wish for a reenactment of their past time smelting activities was a suitable occasion for the final transferring of these secrets. He was especially concerned with the knowledge stored in the magical basket, intaangala. As in the past it had been an important transformation of social status related to the inheritance of the intaangala. He more than anyone was aware that the times of active smelting were long gone, still he claimed that the importance of the magical basket was another thing. He wanted to hand the basket over to this nephew, a specifically skilled blacksmith, after the smelt that we were about to start. He wanted this particular man to inherit his basket of magical substances because he had always been one of the most interested in the smelting activities of the past. He was also selected because of his skills in the forge (see plate 4.1). When I returned to Ufipa the following year and the preparations for the second smelting was about to start this particular younger man was indeed the new owner of the magical basket and was henceforth the new master smelter, the

During our preliminary survey to Ufipa we looked at the present day forge and were introduced to some of the other participants in our crew. We meet another aged man that had also been an active iron smelter in the past. He was the father of the assistant, unnsole, and all his brothers that were also blacksmiths. Most of these men were living within the same area of the village and one of them opened up his home for us. This house became ‘our home’ during all our stays in Ufipa. This house was situated in a location of the village where most of the ironworkers lived, so we were at the heart of the descent clan to which they all belonged.74 The old master had however to descend a small hill and cross a little bridge on his way to us every morning since his house and his sons with their families lived on the opposite side of the village. However, the head smelter arrived early enough every morning to get one or two cups of coffee with lots of sugar and a talk before we went to the smelting site. Also some of the other elders often arrived at our house in the morning and when the assistant came, we would join the younger smelters at the site. The documentation, what Lemonnier calls description (1992:28), was at my first meeting with Fipa iron workers complicated and confusing; there I was, not ‘ready to observe the making of popcorn in a middle-class American kitchen’ (ibid.), but of a sophisticated metallurgical process. Although I had read about their way of making iron in for instance Willis’ (1981) and Wise’s (1958a, 1958b) descriptions of the smelting process and I had conducted several pre-smelt interviews and questionnaires with the crew of smelters, I was 73

For this second visit to Ufipa I brought an album of photos taken during the first re-enactment and the new master smelter decided to keep it together with the magical basket and heirlooms of iron produced in the past. 74 cf. Willis 1981 for a detailed description of the Fipa clan system.

72

I instantly felt that I could co-operate with this man, and he and his sons became my most important contacts. I admired his skills and knowledge from the first moment I talked to him and he seemed to be eager to inform me about all aspects of Fipa iron smelting.

51

Randi Barndon

Plate 4.2 A Pangwa blacksmith at work in his bicycle-wheel driven forge, Mkiu 1995 would have helped the transcription (ibid.), but for practical reasons this was apparently not possible. However, as Lemonnier also notes ‘an awareness of when and where to make close-ups etc. is also necessary for a film or video recording to make sense’ (ibid.). Again, at that time I had not any experience in observing or recording iron smelting processes at all and a video recording would have taken too much of our time. My point is that it is necessary to be aware of what is at stake in order to make photos or video recordings that can make sense when the process is to be transcribed or studied in retrospect. My advantage in the field was that I had a field assistant that knew what I was looking for and that could easily communicate with the aged men.

unprepared. I was not aware of how difficult it actually was to watch, write, take pictures (with three different cameras, slides, black and white and colour prints) and note the time of the actual smelting process step by step, as well as get an idea of the rituals or the meaning of the songs and utterances (with a tape recorder), without interrupting the people at work. I could not and would not ask them to stop or to repeat a given action that I had missed because I had decided to be mostly an observer and not participate or interfere. Interruptions would also cause problems with my observations and documentation of the sequences and their rhythm in the actual process. The re-enactments of smelts were initiated to get a picture of how iron working had once been practised in Ufipa. But my focus was not only how they actually transformed the ore into iron; the metallurgical process. I wanted to grasp which meanings they founded their work upon. During documentation I had to not only make the process comprehensible to myself, but to record it in such a way that it would be comprehensible and make sense to others working with African iron working. Lemonnier describes many of the difficulties that I experienced and certainly I faced some of ‘the practical limits of describing the material aspects of a physical action on the material world’ (Lemonnier 1992:29). In 1990 I initially planned to video record the entire Fipa smelting process. Too many practical matters made it impossible at the time such as the mere fact that I went there with only one assistant and translator, without a vehicle and our village located too far away from Sumbawanga where we could have recharged camera batteries etc. A video recording

Between 1993 and 1995 three field trips were carried out in the Iringa region, including Njombe and Ludewa (Upangwa) districts. After initial surveys in Iringa, Njombe and Ludewa, the northern part of Ludewa district became our base. My first meeting with the ironworkers in Upangwa was different from my first visit to Ufipa. I was more experienced in observing and recording iron smelting, thus I instantly knew how and what to look for, something that facilitated the work. However, giving me a difficult start was the fact that during my first visit to Upangwa I was not formally introduced to the inhabitants in the various villages by a representative from the Regional Cultural Office in Ludewa. Our village was located 80 kilometres south of Njombe, in Lugarawa, close to Mlangali. On our trip to Ludewa the road had suffered severely from the last season’s heavy rains and due to erosion it was impossible to pass beyond a few 52

Re-enactments of smelting, how to describe and transcribe? kilometres after Mlangali and we had to turn back again. In addition to this, the assistant that I had with me from Dar es Salaam turned out to be completely unfamiliar with iron working and unfortunately his English turned out to be even worse. We could not communicate in a satisfactory way about the smelting process with the people in Upangwa. We were because of the language barriers not able to get a clear picture of how much they actually knew of old practices or if they were willing to share this knowledge with us.75

smelters at a later stage, with other smelters or blacksmiths from other villages who were interviewed and women who participated in some stages in the process. In Upangwa, Amos Mgimba’s sister, Gaodensia Mgimba, helped us in the house with domestic tasks such as cooking and cleaning. She also functioned as an informant. When I wanted to see if terms, concepts or metaphors in the iron smelting practices were familiar to an uninitiated and female person she was always ready to provide answers. The same was the case in Ufipa where Marieta Mapala who helped us in the house had the same double role.76 The recorded vocabulary was checked against those terms recorded by other ethnographers such as Stirnimann in Upangwa and Willis in Ufipa (e.g. Stirnimann 1976; Willis 1966).

A person that we met by coincidence turned out to know where we could actually meet aged master smelters. This same village eventually became my base mainly because there are several family groups of active blacksmiths working at the modern village forge (see plate 4.2.). This formed a similar situation as the Fipa case because also in Upangwa some of the young blacksmiths were descendants of once active smelters and were through their familiarity with forging to some extent, also familiar with the smelting process. Fortunately for us some of the blacksmiths were themselves interested in learning about old smelting practices. They knew of old smelting sites and had been told by the elders about the old smelting technology. Many of these smelting sites we later visited together. In Upangwa the late Amos Mgimba was one of the active blacksmiths who was interested in his father’s iron working practices. Together with his brother, Sylvester Mgimba, he formed an important part of the village group of blacksmiths working on a regular basis in the village forge. Amos Mgimba was later to become an important informant and he shared his compound and house with us during both years of research in Mkiu, Upangwa. As opposed to my Fipa study, which is the most recent in a long sequence of many more recent observations of an impressive iron smelting technology, this is the first ethnoarchaeological documentation of Pangwa iron smelting. It was initiated to gain an understanding of technological, social and symbolic aspects of Pangwa iron working in line with the discussion above and a shift in our conceptualisation of technology.

David, commenting upon methodology in studies of smelting symbolism suggests that ‘in order to install models of a symbolic realm, in depth study of the symbolic manifestation among a particular group or people is required’ (David 2001:17). This involves the search for parallels in other practices in order to elucidate underlying theories, metaphors and analogies. But he adds that ‘the practice of patching interpretative holes with questionably relevant material from other people, even neighbours, must be abandoned unless candidate analogies have been subjected to source- and subject-side strategies for establishing relevance. This relevance one may gain by expanding the bases of interpretations and elaborating the fit between source and subject’ (David 2001:17, cf. also Wylie 1985:100-101). David suggests that researchers should seek validation of their work in a number of ways. For instance by showing that their interpretations in one area are consistent with and can explain behaviours in others, for example pottery making, cooking or hunting, corroborate their interpretation, and by submitting them to the criticism of intellectuals of the groups studied (David 2001:16-17). David states that apparent conflicts and inconsistencies in data should be followed up, preferable in the field (ibid.).77 4.6 Summary In sum iron smelting consists of bringing together ore, fuel and energy in a particular manner. I have argued that of greatest importance was the fact that making metal out of ores was a human labour demanding activity. The master smelters therefore more than anything needed to mobilise a group of labourers, a group of male assistants and women and children to partake in the heavy workloads and transportation of fuel and ores. The smelters did not only operate by unconscious routine

As noted in the pervious chapters, this study is in particular about the use of mental and material metaphors in the smelting process. How were the metaphors presented to me? First of all, I always inquired about why a specific location, a specific use of medicine, magical substance or a specific use of ore was preferred. Secondly, I asked my informants to define concepts and terms. Collecting lexical terms and concepts as related to items or production processes presented me with a vocabulary that I used in my conversation with the

76 These two women were two among many that were interviewed in Ufipa and Upangwa. 77 As an example David notes that ‘it would appear dangerous to transfer the concepts of the jealous wife and of the dangers from adultery from the social to the smelting arena in order to explain sexual taboos without also considering how, societies that abhor the idea of polyandry, the furnace can be the bride of several husbands’ (David 2001:17).

75

By then I managed some Kiswahili but did not feel competent enough to actually discuss iron smelting in Kiswahili. However, I did quite soon become aware of how little my assistant knew about what we were looking for by listening to his conversation with the people we encountered during our survey both in Iringa, Mbeya, Njombe and Ludewa. The name of this person is not included in Appendix E.

53

Randi Barndon observing the actual building of the smelting furnace and finally observation of reduction process until bloomery iron has been produced. It is important to include each preparatory task and each re-enactment in order to produce as many stories as were actually experienced. In sum, these descriptions constitute my account of the events that were observed in the field. Together they form multiple stories of how iron working was demonstrated to us as well as giving more correct versions of the smelters own perception of their tradition. Transcriptions of chaînes opératoires serve to illustrate variations that can be compared and they are transcriptions of what iron technology in this context seemed to be all about.

based on an embodied experience (habitus), because evoked metaphors were constantly referred to (verbally and materially). Observation of iron making, with whatever degree of participation one finds reasonable, and with whatever emphasis one chooses; metallurgical, social or symbolic involves a documentation of the entire ore reduction process. In my transcriptions of Pangwa and Fipa iron smelting Child’s suggested list of sequences in African iron smelting will be combined with Lemonnier’s chaîne opératoire approach. This implies observations of initial preparatory stages, such as charcoal production, ore collection and preparation, tuyere and bellow-making,

54

The Pangwa practice of iron smelting

5. The Pangwa practice of iron smelting Tambbalale ndiksavahungila nye voho, nye va dadi! Ndiksave ivutali, munjimelile aye voha, Yemwafawvile!78 the low forced draft furnace (Sutton 1971b, 1985). In a manuscript written in 1978, Matumizi ya Chuma, Mgina provides a description of Pangwa iron working of Madilu and Manga villages in Njombe district. This work is also based solely on testimonies by elders who once were active smelters. To my knowledge neither of the above mentioned authors have themselves conducted any step-by-step observation of Pangwa preparatory tasks and smelting processes.

5.1 Introduction The following two chapters depict aged iron smelters in their attempts to re-create a technological setting that once was part of their ‘lived experience’. Below I will present the major sequences in the Pangwa pre-smelt and smelting process as they were observed in the field. I will use Childs’ list of sequences of African iron smelting processes in view of my discussion in the previous chapters (and cf. Childs 1991b:340). I will attempt to transcribe what I have observed according to Lemonnier’s suggested step by step description of chaînes opératoires. Nevertheless, I will highlight only what I found most significant in my observations of the process in relation to my approach towards technology and comparison with the following description of the Fipa smelting procedures in Chapter 6.

During my three visits to Upangwa in the LivingstoneKipengere Mountains in 1993 and 1995, two entire smelting processes were observed. The construction of the smelting furnace, liteende, was recorded twice. Additionally, a re-enactment was initiated the first year, but not completed in time for me to observe the actual smelting process. As mentioned above numerous interviews with the crew as well as others were conducted before, in between and after our experimental smelts. Before our smelts started I had, because of these interviews, as well as studies of the earlier idealised documentation of the process by Stirnimann (1969, 1979), a fairly good picture of the sequences of the smelting and forging processes. I never in any of the observed smelts interfered with the umkoyo, the master smelter’s solutions, methods and procedures although I was tempted to do so in order to speed things up and maybe the smelters sensed this.

Bornhardt, a geologists and geographer passed through Upangwa in 1899 and commented on their iron smelting furnace, liteende (Bornhardt 1900:80). Dantz, another geologist who went through Upangwa during the same years, commented, like Borhardt, on the abundance of iron ores and quarries in the region (Dantz 1903:110). Bornhardt and Dantz report that they observed numerous smaller smelting furnaces in their geological surveys in the region between 1896 and 1900. Stirnimann is the only ethnographer who describes Pangwa iron technology in detail and provides a schematic illustration of the Pangwa furnace with inserted tuyeres, tubes and drum-bellows (Stirnimann 1969, 1976:269). Stirnimann reports that when he did ethnographic work in Upangwa in the mid 1960’s the ironworkers had completely abandoned smelting (Stirnimann 1969, 1976:263-286). His information was collected in 1969 and completed by information gathered in Mapanga and Matevele in the vicinity of Mlangali. Stirnimann’s description of smelting procedures is based on testimonies from former smelters but his description seems in general to be very similar to my own observations conducted within the same area of Lugarawa (cf. Stirnimann 1979:16).

My informants told me that if a master smelter wanted to make iron he first had to sacrifice a goat in a special place for ritual activities. This place was called ‘the place for prayers’, kulitekelo, and was restricted to members of the same clan. In the past kulitekelo was always used when a master smelter or other magicians were conducting prayers and sacrifices such as prayers for rain, a good harvest or against epidemics or famine. I was told that before the smelting started the master smelter, two boys and two young girls would go to ‘the place for prayers’. The children attended the prayers and sacrifices because they were regarded as clean and cold, awalele, and because they would learn by attending how to make a sacrifice and pray. A he-goat was slaughtered and parts of the liver and pieces of the meat were placed in a pot and given to the ancestors and gods. The meat and the liver were mixed with salt from Ubena and some local flour were given.79 Stirnimann explains the importance of salt in Pangwa sacrifices and medicines due to its whiteness and thus similarity with semen and

Sutton has conducted surveys of iron working in the Southern Highlands region and commented upon the similarities in iron smelting furnaces among the Pangwa, the Kinga and the Matengo, all of them using 78 When our second experimental smelt was about to begin one of the aged smelters announced the above quoted words. He offered the unseen spirits of the place and his forefathers food and beer and said ’Come towards me, you my forefathers, all of you! I greet you and beg you stand by me when I make my iron. Stay with me, all of you that have died!’

79 In the past salt from Ubena was exchanged with iron blooms and hoes and especially the Kinga had a well-established exchange in iron and salt (Culwick and Culwick 1935, 1967; Koizumi 1995).

55

Randi Barndon breast milk (Stirnimann 1976).80 The flour was placed in a basket, kalilo. Then the master would say: ’Dinjirika mnange ndipate uvutali’, simply meaning ‘help me to get some iron!’ They said this was no time for celebrations, it was a quiet place for quiet prayers, and there were no drums, no dancing or singing. The master smelter brought home the rest of the meat from the goat and the children helped him carry it and eat it. Later the master would select his crew for the coming smelting season and the preparations for smelting would start.

or the members of the lutanana, since today blacksmiths, mponzi, forge from scrap metal. Some of the blacksmiths in the village had never even participated in traditional smelting and did not care to participate in our experiments. Instead our research itself was what initiated the work and paid the workers. In the days of active smelting I would most likely have been considered ‘ritually impure’, but in this setting I was allowed entry. All my assistants (driver and interpreters etc.), equally ‘potentially impure’ young men were allowed entry, while women, children and other young men of the village, members of the lutanana, did not participate.

One of the aged smelter that I interviewed reported that in the past sacrifices to the ancestors were conducted before the smelt, but he was uncertain whether the sacrifices had been at the smelting camp or at ‘the place for prayers’, kulitekelo. He could not remember if the sacrifice was conducted before the building of the furnace or before the actual smelting. Stirnimann does not mention if the master smelter sacrificed to the ancestors at ‘the place for prayers’, but he notes that the master smelter would make a sacrifice and set out food and beer offerings at his own ‘houseoffering place’ the night before the smelting started. The master would call upon his ancestors for assistance in the coming smelts (Stirnimann 1976:275). I never observed and was never told that the master smelter or any other person in our crew went to ‘the place for prayers’ to conduct a sacrifice before our smelts, but then again as I was told, it was a place for quiet prayers.

The master smelters never announced the beginning of the work in the village, and nor did we, although we had discussed our plans with the village headman.81 We did not move out of the village and we never established a living area in the bush, or went on hunting expeditions in order to provide for our own food while living there. According to master smelters interviewed in Ufipa the Fipa master used to have a group of hunters with him that would provide them with food during the months at the smelting camp (cf. also Wyckaert 1914). In Upangwa the master would send some of the members of his umkovi out on hunting trips while others were working on the smelting site (cf. Stirnimann 1976:281). We had a vehicle taking us to and from the village every morning and afternoon and we used our vehicle for transportation of ores and charcoal. In sum, our work was not economically, socially or morally significant for the entire village community, as it had been in the past. It was not directed towards any needs or demands for iron hoes.

5.2 Mobilising a labour force During my visits to Upangwa in 1993 and 1995 the same aged smelters, vakolo vayango, constituted our crew, umkovi, during the entire process (cf. Appendix E). They were also my most important informants. This group of people for each smelt observed participated in various tasks connected to the operational sequences in the process. For the two smelts observed in 1995, a third old man was included in our crew. Therefore, three elders and two young and active blacksmiths from the village constituted the main group of participants in the reenactment of smelts (cf. Appendix E). Others, especially three other aged men and an elderly woman, the wife of one of our masters, were called in to assist during the actual smelting, while members of the masters’ traditional work group, lutanana, were not called in to assist in heavier tasks.

Some of our smelters were at first suspicious and believed we were planning on some lucrative iron smelting business. Still, over time our research became both socially and culturally important, but in a different way from the days of active smelting. As I had experienced in Ufipa previously, we gradually noticed a revitalisation in the significance of the master smelters (as ‘cultural heroes’) and a new interest in old pagan beliefs and traditions connected to the smelting process. The first step in any smelting process is the labour recruitment, the mobilisation of apprentices, assistants and the otherwise necessary labour force to conduct a smelt. A master smelter cannot alone manage the entire process. This essential point is not specifically highlighted in either Childs’ (1991b) or MacLean’s (1995) discussion of general sequences in African smelting processes. I believe, rather, that both the beginning of the work process by the masters’ mobilisation of a labour force and the ending of the work, by the evaluation of the outcome, are critical points to include.

The constitution of the group resembled the old day’s practice but our ‘setting’ was entirely different. The mechanisms that were at work in the past when a master smelter mobilised a working force were not present in our setting. Nobody really needed the iron we were about to make, as they would have been in the past. Local made iron was no longer ‘a thing of value’ among the villagers 80

The eating of salt had also been observed in Cewa celebrations of furnaces (Avery et. al. 1988:273). Davison gives an elaborate description of symbolism and taboos related to Cewa salt making and use of salt in food (Davison 1993:34-35,, see Chapter 9).

81 In the past the entire group of smelters and the master’s (first) wife and children left the village with their utensils, medicines and a goat for the sacrifice while singing and dancing (Stirnimann 1976:276).

56

The Pangwa practice of iron smelting manifestations have rarely been documented.82

The mobilisation of a work force, keeping the group under social control and finally the evaluation of the iron produced, are the three most crucial steps in the process. In our interviews I was told that in the past a group of six to nine smelters would work on the furnace during the actual smelting process. Women, children and younger men of the master smelter’s lutanana would assist during various preparatory tasks. These people would assist in transportation of fuel and ore to the smelting site and were paid in beer and food. Our informants told us this worked like an ordinary communal work group, umkovi. Once the smelting begun some of the men went on hunting expeditions or children would bring beer and food to the smelters. Only the professional smelters were allowed direct contact with the ores, the furnace, the medicines, the tuyeres, the bellows and the blooms and they were the only participants who were given iron at the end of each smelting season.

For the first smelt observed in 1993 we decided together with the master smelters to buy charcoal from one of the many local producers in Mlangali, the centre for the former leading clan of Mtweve. A local charcoal maker produced the needed amount of charcoal out of the correct wood for the second and third experiments carried out in 1995. Thus, the charcoal, makara, was made several months before our return to Upangwa. In line with the elders wishes my contact person among the smelters arranged for a local charcoal maker to produce the fuel from specific types of wood, locally named mahato, mtelea and mdolbole. Stirnimann reports that the species mtela, mdolbolo and mkufwa were the most common in use by the iron makers he interviewed (Stirnimann 1976:266). Our master smelters and blacksmiths told us that charcoal from the mkufwa tree, a hard kind of wood, was excellent in forging, but not wanted in iron smelting.

Of all the observed preparatory tasks those in most need of labour were the collection and transportation of raw materials, such as wood for charcoal, ore and water and to a lesser degree collecting termite soil for the building of the furnace. Stirnimann mentions that the entire lutanana, would participate and prepare the raw materials such as the production of charcoal and assist the kneading of the furnace soil. He does not however give, any numbers on the group of participants. In our setting we had a vehicle that facilitated transportation of raw materials. Other heavy tasks, apart from blowing the bellows that the elders were exclusively allowed to do, the young boys performed under the supervision of the elders. When in the past, the furnace was ready it could easily be left for a month in order to dry properly. While the smelters were waiting for the furnace to dry, ores were extracted from the quarries, transported back to the site and prepared for the smelting.

The fact that in 1995 the fuel for the furnace was prepared several months before the actual smelting was in accordance with traditions and as reported by Stirnimann (ibid.: 266). It is also an example of how Childs’ sequences of tasks in the process may vary from one region to another and indeed alter the major sequences of the process.83 Although Pangwa iron smelting was a seasonal activity it was because of this preparatory task being staged in January, several months before the smelting, more firmly integrated into the yearly circle of activities. The participants were conscious of the activities all the year around, both the practical activities and taboos attached to them. Moreover, they were aware of the master of them both; the umkoyo. My informants, as well as Stirnimann’s, claimed that preparations for the charcoal production started in January when the master smelter organised a work group of the ironworkers within his lutanana to cut down the trees. The amount of charcoal needed was known several months in advance of the smelting season and probably adjusted to local seasonal needs and demand for new iron blooms. Before the master smelter and his crew went out in the bush to cut the wood, the master told his assistants that they were not allowed to be ‘hot’ and have sexual intercourse from two days before they were going to collect the wood. Once cut down, the wood was left in the bush until after the heavy rains. The importance of sexual abstinence was re-announced when the charcoal making started in June, a few weeks before the smelting begun. Again, two days before the charcoal making started, the participants had to avoid sexual relations. I was told that all wood working is men’s work and men will make the mortar, ilutuli, and pestle, umtwangilo, for their women who are not allowed to use the mortar when in

5.3 Making charcoal, ulweko Charcoal making, ulweko, was the first preparatory task in the Pangwa smelting process. It was not observed, as I had been able to in Ufipa (Barndon 1992:51-52). However, the process was explained to me and we visited several present day charcoal making sites in Upangwa along the Njombe-Ludewa road. Mgina (nd.) notes that present day Pangwa charcoal making process is identical with the one once practised by the smelters. Only the iron working participants taboo against sexual intercourse during regular charcoal making was not mandated. Still, further inquiries into Pangwa charcoal making as it is practised today might yield a similar sexual taboo, since sexual abstinence seems to be important in most craft production as well as specific stages in a persons life cycle such as before marriage, before birth of first born etc. Studies of iron working have indeed had a tendency to ignore the charcoal making process as an integrated part of the process. Its concepts, practices and material

82

But cf. Schmidt for a detailed description of Haya charcoal making (Shmidt 1997:66-70). 83 See Chapter 4 for Childs’ point 4 of sequences in iron working (Childs 1991b).

57

Randi Barndon menstruation. Fire making is conceptualised as wood working and therefore only men are allowed to use traditional wooden friction drills. Women on the contrary will, if they need a new fire ask their men to make it or collect a piece of charcoal from another hearth.

of wetness in it and its purity. This was possible only by avoidance of ‘hot’ and impure people. Schmidt mentions the use of six (gunny) sacks of charcoal for each Haya smelt (ibid.). This amounts to about 130-180 kg of charcoal used in one smelt and it is interesting to note the similar amount used in Pangwa furnace as in Schmidt’s Haya case, especially since the furnaces are of the same size.

According to Stirnimann a 5 to 6 metres long and 3 to 4 metres wide and oval pit, ilivuka, was made for the charcoal burning. This corresponds with Mgina’s (n.d.) statements, my own observations and my informants who claimed that the charcoal preparations for a smelt were similar to the charcoal making that we observed practised all over Upangwa. At the site, ilivuka, logs approx. 1 metre length, split in halves if necessary were, placed within a pit that was covered with earth. Holes, ifiloso, one in front and one on each side, were made in the mound to let the smoke out. On the very top wet earth was placed in order to avoid bush fires. Like in iron smelting, the use of wooden friction drills was used to make fire, umpechelo, (as opposed to modern matches or a piece of burning charcoal). The umkoyo would not bring fire drills used in his house in the village but make new ones at the charcoal site in order to control their cleanness and coldness (cf. Stirnimann 1976:267). This implied that they had not been used by any person who had just had intercourse or been in touch with a woman in menstruation.84 In the past the smelters stayed out in the bush while the charcoal was burning and a small hut was made at the site. The preparations and burning of the charcoal would take up to a week and after a week the charcoal was left another day to cool down.

In order for a Pangwa person (male or female) to remain ‘pure and clean’ and ‘cool’, awalele, washing in a white liquid, ritual water, kiwalasha, was essential Before rituals and most seasonal activities people washed themselves in kiwalasha. When charcoal preparations for a smelt were carried out in the bush all the participants in the master smelter’s lutanana had to wash in kiwalasha. From the moment when the cutting of trees for charcoal making started the master smelters announced the taboos of which sexual abstinence was the most important. In the past the charcoal making was carried out several months before the actual smelting started yet the taboo against sexual activity was specifically announced. Assistants and apprentices were told about prohibitions and precautions. If sexual abstinence was not maintained it would affect the final outcome, the quality of the bloom. The masters said they were afraid that ‘hot’ people, people who had just had intercourse, vapyio, women who had been sexually active, women who had just delivered, were breast feeding or who were menstruating should harm the charcoal if they entered the site or passed by the charcoal production mounds. Menstruation blood is in general considered dangerous for Pangwa men and will make them weak. A pregnant woman should be separated from her brothers in law and brothers because she may give them mild sickness such as toothache, headache or stomach ache (Stirnimann 1979:252). If a man has sexual intercourse with a woman in menstruation he will suffer from iseke, a disease that will make his testicles swell and make him impotent (Stirnimann 1976:285, 1979:232). The Pangwa tell that calabashes are symbols of intercourse and a married wife is not allowed to drink from a calabash given to her from another man than her husband (Stirnimann 1979:179). This act would be regarded as adulterous and dangerous and could lead to iseke. As I will return to in Chapter 7 it also combines deeply experienced notions of containers, of boundedness and openness in embodiment leading to adultery and sickness, iseke, being threats in every day life.

When the furnace, liteende, was completed, the charcoal that would support the furnace with fuel for the seasons smelt, was collected and brought to the smelting site. Women, men and children transported baskets of charcoal to the smelting site, but again those who were considered ‘hot’ or unclean (i.e. daughters and wives of the smelters) could not assist. In the past the Pangwa smelters prepared the charcoal close to the smelting site; they told me that a new smelting site was often selected during surveys for hardwood to be used in the charcoal making or while they were burning charcoal.85 For our re-enactments of smelting specifically hard woods were used that the smelters called lilongalonga, mtela and mahato. We brought six sacks of charcoal to the site; each weighed about 20 kg. Schmidt notes that the amount of humidity in the charcoal can influence the outcome (Schmidt 1997:70). The smelters claimed that one of the reasons that they prepared the charcoal just before the smelting started, and not when they first went out in the bush to cut the wood, was in order to control the amount

The same fear of iseke followed all preparatory stages and a specific care for the medicinal and magical substances that were placed in the furnace pit was observed in our experiments. Especially the medicines, ing’anzo, in the ing’anjo pot were guarded. If a pregnant woman was standing on a hill close to the smelting site looking towards the pot she could harm the smelting medicines (Stirnimann 1976:272). Like the smiths explained to me,

84 Smith and Dale report that the Ila called the lower piece of wood in the friction drill ‘the female’ while the drill was called ‘the male’ (Smith and Dale1968[1920]:143). Bekaert (1998) notes the same for the Sakata smiths and this has also been recorded among the Pangwa and Fipa. 85 I.e. while hunting and looking for food to be eaten while in the bush.

58

The Pangwa practice of iron smelting

Plate 5.1 The furnace guardian, umboocha, Ing’angitoli 1993. witchcraft accusations. Our smelters never provided any technical or metallurgical explanations for abandoning a smelting site.

a woman in menstruation could harm the fire in ordinary cooking or in charcoal making as well as in iron making. Pregnant women or women in menstruation, as well as those (men or women) that were ‘hot’ after intercourse, vapyio, or even worse, after committing adultery, could become weak by loosing their coolness. Adultery caused anger, something that could cause a fire to die out. The smelters said the ‘heat’ of fire was not compatible with the ‘heat’ of anger and ‘heat’ in sexual activity. In order to maintain and secure the correct balance of substances various ‘heats’ or energies had to be kept apart form each other.

A Pangwa smelting site, pamokoka, was often located close to and within a kilometre in distance from the charcoal production area, ilivuka. Of importance was its vicinity to water.86 Most commonly the smelting site was near to a small creek or river and it was close to the termite soil that was used for the furnace building. Of uttermost importance was its distance from settlements and villagers. A master’s furnace was always built out in the bush, on his own land but in the vicinity or close to his fields and gardens, another factor limiting the choices of smelting location. Hence, the Pangwa furnaces, like the Fipa furnaces, seem to have been built within the land of the master smelter. And similarly to the Fipa, the Pangwa furnaces were reused several seasons while new furnaces were built close to the old ones. In our surveys in the Livingstone Mountains we often observed two or more furnaces within a few metres of each other or at least within sight of one another.87 These furnaces were often operated by the same master smelter. Upangwa is a mountainous country and often surrounding fields and settlements were in sight of the smelting site or vice

5.4 Smelting in secrecy, guarded by the umboocha Selecting a new site was not made at the beginning of each new smelting season. This indicates a slight revision of Childs’ list of sequences discussed in the previous chapter. The selection of smelting sites was therefore not the first step in the process at all. But why were new sites ever selected? My informants told me that if a new master smelter arrived in the area, he would create a new location for smelting. Otherwise an old smelting site was only abandoned if for some reasons the head smelter was not able to produce good blooms or forgeable iron in this location. His forefathers (dead master smelters) would then tell him to move to a better-suited place. Perhaps, as my informants told me, the furnace was impure because it had not been properly purified in the pre-smelt rituals. Or the medicines that were supposed to keep it pure had failed and impure or ‘hot people’ had too often visited the place. In sum, these explanations were about sorcery and

86

This was also stated by Stirnimann’s informants (Stirnimann 1976:268), and confirmed in our surveys of old abandoned smelting sites in Upangwa. Carrying water is a heavy task, Mgina (nd.) notes that the tuyeres were used while my informants stated that in the past ordinary pots were used. For our smelts both pots and aluminium trays were used. 87 We were told they would then belong within one lutanana.

59

Randi Barndon versa. But it was not directly in sight; only smoke from the furnace was seen from afar, by those passing by (cf. Reid and MacLean 1995:151), or those working in the fields or at home around the houses. Other sites were far up in the mountains, again only slightly visible from the fields and compounds below or at the opposite side of the valley.

from the furnace, the white umboocha and sticks, of the kiwalasha tree at each crossroad, the secrecy in smelting was made public. Unwanted persons were kept away. Simultaneously the master smelters announced the importance of themselves, their skills and their knowledge of technology and magic. Iron smelting in Upangwa was a secret event and most smelting activities were conducted away from onlookers. But in the past the smelting season was publicly announced. Women and young men who were not a part of the crew were told to stay away and were warned by ‘the furnace guardian’, the umboocha. The smelting site was not seen from afar, it was hidden up in the mountains, down in a valley and deep in the bush. Yet the smelting activities were not carried out without being noticed. The smoke from the furnace was seen. More important, the liteende, literally ‘something beneath’ but some times referred to as ‘the one dancing’ was heard far away. In the past the men who were blowing the bellows had numerous iron bells, mangala, fastened around their feet. During the bellow working they made rhythms with iron bells hang on their feet and they danced and sang while they blew the bellows. The bellows, ing’oma, made sounds like drums. The bellow containers looked like drums and when used they made rhythms like drums. The smelters sang while they worked, to keep up their spirits and ensure a constant rhythm of their work. This could be heard all over Lugarawa, in Mkiu and in Masimbwe. It was ‘the bells of iron’ that the people all over Upangwa heard during the smelting season.

The first step taken when a smelting sited had been selected was to place ‘a furnace guardian’, umboocha at the entrance side of the camp. Umboocha is a medicinal and poisonous plant, or the root of this plant, was hung from a stick to warn off intruders (see plate 5.1.). The smelters told me the poison from the umboocha would mix with the smoke from ‘the wind maker hole’, kwibepo, on the furnace. Some of the air from the furnace would go out in the direction of the umboocha and mix with its poison. Stirnimann notes that leaves from a white plant were placed on top of a stick at each crossroad leading to a smelting location in Upangwa. This might have been pieces from the umboocha or another ixilala, ‘plant from the bush, an ilala, which literally means a plant that was taken home (cf. Stirnimann 1979:81).88 In both contexts, white symbolised cleanness and was used in order to warn off unclean or impure persons. The Pangwa ‘furnace guardian’ with its white meat and leaves was and is truly dangerous. If an intruder entered the smelting site without permission its odours would come out and give him or her pain in the stomach. The same plant, that only grows wild, would cause adulterous and ‘hot’ smelters to have pain and become sick when they were blowing the bellows (Barndon 1996b:765; Mgina nd.). The oldest master in our crew explained that ‘if a new smelter arrived at the site, the umkoyo would ask him if he was ‘pure’ or ‘cold’. This meant if he had stayed away from women. If the newcomer was lying, the umboocha, being our furnace’s guardian would pull out poison and cut his arms’.89 The smelters also explained that if young men blew the bellows their fingers would become hard since they were ‘hot’ people and that it would be impossible for them to work.

5.5 Building a smelting furnace at Ing'angitoli In 1993 the Pangwa smelters selected a smelting site within a small valley located between Mkiu and Masimbwe village. The site was approximately one kilometre off the road; on the right hand side from the main road leading towards Mlangali and Mavengi. Mkiu village had a modern forge and functioned as our base during our stay in Upangwa, while Masimbwe was the present home village of two of the aged smelters. The area was called Ing’angitoli and was close to a small river and a clay deposit.90 The clay at this particular site was not from the reddish termite soil that I had expected they would use for building the furnace. At Ing’angitoli the clay was of a fat greyish quality very much like clay used in pottery making. From the smelting location we could see towards Masimbwe but we could not be seen from outside the valley. We were not seen from anywhere, not from the road climbing down the mountain from Masimbwe or the main road leading towards Mlangali. The site was in the midst of tall trees and scrubs and hidden from sight unless one walked really close. We never observed anyone passing by the site while we were there.

The umboocha was feared and was used by one of the smelters who was a herbalist and medicine man in order to control stomach problems in his daily practice as a doctor. But like one of the masters told me, ‘if a person in sorcerer the umboocha medicine will scare away problems and his bad feelings’. The fact that the plant was used in other contexts made everybody aware of the dangers it contained and caused people to stay away from the smelting sites. Because of the signs such as the smoke 88

This refers to its use as a medicine. The Fipa’s umpakasi, a hen’s wing hanging at the furnace front was called ‘the furnace caretaker’. It hang in front of the main tuyere inlet, palinyina, to warn off witches and unclean persons (Barndon 1992 and cf. Chapter 6). 89

90 Two old smelting sites we surveyed were also called Ing’angitoli (cf. Appendix D).

60

The Pangwa practice of iron smelting

Plate 5.2 Making the furnace foundation trench, luutu, and pit, umlindi, at Ing’angitoli 1993 In the past, as one of the old masters assured us, the building of a new furnace started at the end of May (cf. also Stirnimann 1976:271). This seems likely because the furnace itself needed the heat from the sun in order to dry. A Pangwa furnace was built in three major sequences, first the furnace pit, umlindi, and foundation, luutu, were dug out. Following this the furnace superstructure, the shaft, was built in two separate sequences. The building of the furnace normally took three to four days. The entire shaft was not built in one sequence because the furnace was made of wet clay. The interval between the building of the two sections of the shaft depended on weather conditions and how fast the clay or termite soil would dry. A crew of five workers was sufficient for this task as for the other preparatory tasks, such as kneading the soil, making clay tuyeres, blowpipes, goatskins for the bellows, the drum bellows, the washing of the iron ore and the numerous magical preparation.

preparations for the furnace foundation. In the past a hut for storage of charcoal and ore (for the multiple smelts) as well as another hut in which the smelters would cook and sleep were part of the site. If it was a new site, these huts were built before the work started. The clearing of a site was in the past as for our re-enactment, carried out by all the participants. The entire site sloped downwards on its western side towards where a small river was located. The kneading of termite soil, hutwanga, for a Pangwa furnace was not a task that needed the entire working force, lutanana. It was not a public event and not like the co-operation of the twenty people that the master in Ufipa to mobilise for his ifo making (cf. Barndon 1992 and Chapter 6). The four smelters, young and old, started preparing the clay to build the furnace. A 50 cm diameter circle with an opening in the centre was formed of the clay on the ground. Water was added, they were as they said, ‘making something wet’, kulaopila, by gradually homogenising the wet clay with a stick, umtwangilo. Women normally use umtwangilo when they prepare clay for their pots. Umtwangilo also means pestle. When our smelters prepared the furnace soil, lidope, they joked about how they were like women and potters. But they were also worried if women would pass by the site and see them performing feminine activities.

The site, pamokoka, was from then on the centre of all activities for a minimum of four days. The local brew that was offered by the master smelter to his crew once the furnace was completed had to stay three days before they could drink it. Hence, as the old men told me with a laugh, the construction of a furnace would take three days as well. No specific elaborate rituals or sacrifices were performed at the furnace foundation or during the building of a liteende. But as mentioned above taboos were important from two days before the work started and through the entire smelting season. The master smelter and his assistants, avayangu, had to remain cool and pure, in balance.

Before the building of the furnace started in 1993 the umkoyo from Masimbwe used a wooden stick and his body in order to get the correct measures of his furnace and the furnace pit.92 The old umwaami in Ufipa had used a wooden stick called entoa literally a ceremonial wedding stick (Barndon 1992). He also used his body, his arms and legs in order to get the size of his furnaces (ibid.: 54, see also Chapter 6). As commented by another researcher, the Fipa umwaami was a particularly tall and strong man (Wembah-Rashid 1973), something that resulted in all his furnaces being rather big. This does not

When we arrived at Ing’angitoli the entire site was first cleared of trees and scrubs and partly flattened out. This site was smaller that those used in the past active days of iron smelting.91 Hence, one hour later we could start the 91 Our site was smaller since we did not build a hut for sleeping or a hut for the storing of raw materials (cf. also Stirnimann 1976:272).

92

61

We were never given any term for this stick.

Randi Barndon

Plate 5.3 Building a furnace at Ing’angitoli in 1993

Plate 5.4 The furnace, liteende, when ready built, Ing’angitoli 1993

mean that the furnaces would vary significantly but if one were able to observe and follow several smelting crews and master smelters simultaneously, details like this might be recognised as ‘the personal fingerprint’ of the masters onto their furnaces.

at the base the inner diameter measured 48 cm. At the base the wall was 20 cm narrowing down to 10 cm towards the top of the furnace. The furnace from ground level to mouth measures 75 cm (cf. Appendix D, tables 2, 3 and 4).

The umkoyo of Masimbwe used the upper part of his arm (which measured 48 cm) to decide the measurements of the furnace pit. Based on these measurements the master drew a line on the ground and excavated a 50 x 50 cm wide pit. Then the master again used his measuring stick and the length of his fingers formed an outer circle around the furnace. After one hour, the furnace pit, umlindi, was ready. The pit and foundation trench then partly sloped from east to west, the difference in levels was 15 cm. At this stage the total diameter was 91 cm. Umlindi also the generic Pangwa term for a grave and in this context it referred to the ancestors and spirits, maoka, of the dead. During the actual building of the furnace the master only trusted the elders to touch the termite soil, while the younger men were allowed to prepare it (i.e. kneading it or mixing it with water). When they had placed approximately 15 cm of wet clay for the furnace chamber they prepared the furnace for the tuyere inlets.

Both the inner and outer side of the shaft was smoothened with water as if it was washed in ritual water, kiwalasha. The master smelter kneaded clay in front of the furnace and made two breasts, amawele, above the opening. The furnace looked like a woman. The smelters told me this was a young woman, umdala, and not yet anyone’s bride, mwinja. They said they made the liteende, ‘something beneath’ with female breasts. When I asked them why the furnace was like a woman one of the old smelters said if the furnace was like a man it would never make iron, a man never cools down. And he explained to me in other words ‘a woman cooks food, a woman gives birth, a man never cooks or gives birth to a child’. The other one said ‘the furnace woman’ is still unmarried, she is still looking for a husband. We say the iron ore will be her husband but we are the fathers of the iron ore’. Two of my informants told me ‘a Pangwa furnace, liteende, referring to ‘something beneath’, panyi.93

Back at the smelting site the next day the furnace was completed (see plate 5.4). The stick the master had placed in the furnace the day before was taken out. The doors were polished as well as the entire furnace. At its completion the furnace measured 130 cm from the base to the mouth. At the top the inner diameter was 20 cm, while

93

Stirnimann notes that a woman in menstruation is called umudalu panyi which means a woman who sleeps down on the ground (and not in her husbands bed) because of her hotness (1969 unpublished ).

62

The Pangwa practice of iron smelting

Plate 5.5 Making the furnace pit and trench for the second re-enactment, Kuking’ande 1995 The main opening on the furnace was placed towards the river and was 29 cm wide at the inside of the chamber while it was 40 cm wide at the outside. The smaller tuyere inlets at the sides were 11 cm and 13 cm measured from the inside, and 18 and 20 cm from the outside (cf. Appendix D, tables 2 and 4). The smelters called the front door, pointing towards the river for the ’mother opening’, libepo, or ’mother mouth’, umlomo, while the two others tuyere inlets were called doors, amajango, or kilulepo (cf. also Stirnimann 1976:269, 283).94 During the first day the smelters built the furnace shaft up to 40 cm above ground level.

fertility and iron smelting, They were visual metaphors and thus material metaphors. The building of the furnace, or the chaînes opératoires of building, strengthened the metaphorical link because the aged smelters worked with a special body rhythm throughout the entire construction of the furnace. The aged men did not move in full circle around the furnace. They only moved from the front and the sides towards the back. During the entire construction phase they were not allowed to walk or work a full circle around the furnace shaft. It was as if the furnace was a woman that they should not turn their back to. It was clearly a sign of correct behaviour by the smelters as magicians.

A stick was placed inside the eastern side of the wall. This stick, libi, would be taken out when the furnace had dried. They said the hole told them where the back of the furnace was. During smelting the hole worked as a peephole through which the master smelter controlled the colour of the fire and thus the temperature during the smelt. It was called ‘the wind-maker hole’, kwibepo, and the smelters said that by inserting a wooden stick or removing it they were able to control the fire in the furnace chamber.95

5.6 Building the furnace at Kuking’ande For our re-enactment in 1995 another location was selected. I had been away for almost two years and apart from correspondence with my contact person in Mkiu we had not discussed details for the remaining research. When I arrived in Mkiu in June 1995 the two aged smelters and the young blacksmiths from our previous crew had decided upon a new smelting site. The place was called Kuking’ande and was within walking distance of the new master’s home, at Magasiwalale. Another soil of a different quality, from a termite mound at the site, was used for the building of this furnace. The termite soil was red and less fat and wet than the one they had used in 1993. In my opinion this seemed a better choice and the smelters assured us this was the correct kind of building material?

The master smelter decorated the furnace front with female breasts, amawele, placed just above the front opening, libepo or umlomo. Female attributes are easily recognised on Pangwa furnaces and were seen on old furnaces during our surveys (cf. Chapter 8, plate 8.1). The anthropomorphic decoration of the furnaces is similar to Matengo, Bena and Kinga furnace decorations documented by Sutton (Sutton 1971b, 1985). They were materially confirming and metaphorically visualising the imaginative link between people and furnaces and female 94 Cf. Appendix A and D for Pangwa terms given to the various parts of furnaces during re-enactment of smelts and old furnaces. 95 The old men explained to me that mulindi or kwibepo was the back, anus, on the female furnace but they would not call it by its correct name because that would only lead to village gossip.

63

Randi Barndon

Plate 5.6 Making the furnace shaft, Kuking’ande 1995

Plate 5.7 Washing the furnace, liteende, in ritual water, kiwalasha, Kuking’ande 1995

A new master and magician, umkoyo, had been included into our group.96 The new umkoyo had been included in the group because of his knowledge of magic and medicines. He was both respected and feared, hence, he had an ambivalent and ambiguous position. Moreover, his father had been among the most respected mkololo’s in Mkiu. The three old masters decided that our new smelting site should be close to the new Mzee’s home at Magasiwalale. Magasiwalale is a part of Mkiu village and the smelters decided that they would all sleep in his compound until the smelting started. The two other smelters from the Mtweve clan who lived rather far away at Masimbwe village agreed and since our vehicle consumed more petrol than we were able to fill every second day at Mlangali we settled for this arrangement. The newcomers’ wife was to bring the elders food and beer during the smelting and if she was needed she agreed to assist in the preparations for the smelt. The newcomer assured us that she would be safe to bring to the smelting site because she was not ‘hot’, mooto. But I was also told that even old women could not always be trusted on these matters, they had to take care and only include women who knew the importance of sexual abstinence and purity. There was a long discussion among the smelters before

they agreed with the new smelter on this matter. The Mtweve men from Masimbwe had not been happy with the location that we had used in 1993. They believed that potters collecting clay had visited the place. They were also afraid the place had been bewitched and that this area outside Magasiwalale was unfamiliar to them. They were not certain if they would find the necessary herbs and magical substances in the surrounding bush. At first the two masters from Masimbwe were not happy about this idea but finally agreed. The Magasiwalale master said he would take care of the medicines, after all he was a respected magician and knew the plants and herbs from his land. Our new site, locally called Kuking’ande, literally meaning ‘where the unseen live’, was finally located on the new umkoyo’s land, at Magasiwalale within the western part of Mkui village. The smelters said the ’unseen ones’ at Kuking’ande were their gods, mamlungu and mangulowo. When we returned to Upangwa in 1995, a stick that was almost 6 metres tall replaced the small stick with its umboocha used in our 1993 experiment. Nevertheless, we could see the enormous umboocha hanging from its stick. It was seen from practically everywhere and from any direction leading to the smelting camp. In this way the smelters secured the smelting camp and the umboocha forcefully announced that iron working was soon to be carried out. The smelters had made even greater efforts to avoid disturbances and problems from ‘hot’ people. Only once did I observe an uninvited person approaching the site. A young man came close to the site; we could see

96

The entire group was growing. In 1993 we had been only two foreigners, my interpreter and myself on our two trips to Upangwa. In 1995 two more persons were added. Thanks to Professor Felix Chami at Archaeology Unit we were able to use a vehicle from the University of Dar es Salaam. With the car came a driver, Ali Hamza. In addition to my interpreter and field assistant E. M. Kessy from Archaeology Unit, UDSM, Amin Msuya came with us.

64

The Pangwa practice of iron smelting

Plate 5.8 Making tuyeres, Ing’angitoli 1993 him while he could not see the site. The smelters quickly told him that he was not allowed to approach any closer and asked if he could not see their need for privacy.

the first smelt in 1993 the crew used the same clay as for the building of the furnace. For the second re-enactment in 1995 they used the same clay from Ing’angitoli after trying to make them with the termite soil found at Kuking’ande. No tempering material was added to the clay and water mixture in any of the observed cases. Specific sticks, libihi, that were used for the tuyere making were old ones that had been in use during active smelting days. The sticks were coated with ash to keep them free of clay. The old masters worked the clay and made lumps of clay that were worked out around the sticks. The clay was mixed with water and softened before it was formed around the tuyere stick. Each tuyere was slowly removed from the stick while still wet. The tuyeres were 40 to 45 cm long with an inner diameter of 4.5 cm and an outer diameter of 10 cm (cf. Appendix D, table 3). The tuyeres were left on the ground to dry in open air only covered with some grass to prevent them from cracking because of the strong sun.

The clearing of the smelting site started at 11 a.m. on our first day at Kuking’ande. The three aged smelters told the young men that from this moment they had to be especially careful with the taboos they had told them about. Hence, they were not allowed sexual intercourse. Women and people ‘hot’ after intercourse were not allowed to enter the site. As previously observed the young men cleared the site for trees and bushes and made a semi-flat platform. The aged men surveyed the termite soil south of the smelting site before they started the preparations of the furnace pit, umlindi or luuto, and made the trench, huwaki luuto (see plate 5.5). The masters decided that the preparations for the trench had to wait until the following day. The following day when we arrived the young men kneaded the termite soil, lidope, while the elders started the digging of the furnace pit and foundation. Again, the oldest master smelter in our crew used his arm as a measure. Once the pit was ready the building of the furnace started. The furnace pit was 44 cm deep and 50 cm in diameter. The foundation pit was 16 cm deep and 84 cm in diameter. The total height of the furnace was 98 cm from the ground level to its top (cf. Appendix D, table 2). Below was the umlindi, pit, for the medicine pot (see Chapter 1, figure 1.3). Only the aged smelters built the furnace. The masters washed and smoothened the furnace shaft with water, kiwalasha, and left it to dry in the sun.

Similar to recorded observations of the Fipa smelters (Barndon 1992; Wembah-Rashid 1973), the Pangwa smelters struggled with the tuyere making. This was observed in 1993 and again in 1995. In 1995 the first tuyeres they made had severe cracks and could not be used. Only after making the tuyeres with clay that we collected at Ing’angitoli did they manage to make enough tuyeres. Hence, this preparatory task seemed to be one of the stages in the iron smelting process that demanded special skills. As we observed, the aged men as well as the young men struggled to complete enough tuyeres for one or two smelts. When I asked them if in the active days of smelting specialist made the tuyeres or even women who were skilled potters and more familiar with working clay, they denied it. In particular women would never make tuyeres. It is remarkable that one of the tasks that indeed seemed to be among the most difficult was not prepared with the same care in terms of medicinal or magical preparations as were other preparatory tasks. No

5.7 Making tuyeres, bellows and beans Once the furnace was ready and left to dry in the sun the men started making the tuyeres, hence the next preparatory stage was the making of ing’elu.97 A Pangwa furnace needs three tuyeres, one in each tuyere inlet. For 97 Ing’elwa is another term given the tuyeres is probably of Kinga origin.

65

Randi Barndon

Plate 5.9 Drum-bellows, n’goma, Ing’angitoli 1991

Plate 5.10 Tuyeres, tubes and bellow arrangements, Kuking’ande 1995 rituals, sacrifices or prayers were made before, during or after the making of tuyeres. For our re-enactment the three sets of tuyeres were eventually made, in addition to a few more as a reserve if they broke during smelting. The work was confined to the young men who were embarrassed because working with clay is a female task.

The diameter of the drum chamber varied from 12 to 15 cm and they were 15 cm deep. The long hollow sticks connected to the drum bellows were collected in the vicinity of our sites and cut to correct length before they were inserted into the drum bellows that were sometimes called iwefia.99 The goatskins, sonda, were softened with the bean oil, imono, before they were cut into correct size and fastened around the drums with ropes made from hemp. The drumsticks were inserted into the bellow skins with ropes. They made six drumsticks, umbihi, one pair for each pair of drums. These sticks were 160 cm long and the bellow workers stood while blowing the bellows. Before the smelting started the tuyeres were connected with wet clay to the hollow sticks, iwefia, that at the other end were inserted into the drum bellows (see plate 5.10).

The Pangwa bellow and tuyere system is rather complicated. It consists of five major parts; tuyeres, tubes of bamboo, wooden drums with skin bellows and drumsticks. During smelting the tuyeres were inserted into the furnace chamber, varying from 4 to 5 cm at each opening. Six drum bellows and six drumsticks were connecting the drums and the skin bellows. The drum bellows, ing’oma, were made by hollowing out trunks of the liheve tree, a softwood also used for making ordinary drums that are also called ing’oma.98 This was a task the young smelters would do in-between other tasks.

While the young men were making the drums for the bellows the elders prepared the oil, imono, made of beans

98 This term is given to drums over a wide area of Bantu-speaking Africa and commonly ngoma refers to healing cults (Janzen 1992)

99

66

Iwefia was used but may be a Kinga term.

The Pangwa practice of iron smelting

Plate 5.11 Preparing imono at Ing’angitoli 1993 were closer to Manga and Lushito. In Madoko there had been four master smelters and when he moved to Masimbwe he continued as a master smelter until his sons took over his umkoyo title. They had used one kind of ore collected from Manga mixed with another kind from Nyaluva.

that they would use in the medicine pot and on the bellows. Constantly one of the aged men was stirring into the pot with imono oil (see plate 5.11). The fire was always looked after and the content in the pot was examined. During the entire morning the men stirred the bean oil. At this stage in the process the aged men started collecting some of the medicines, herbs and barks for the medicine pot. Their purpose and effect was discussed and examined between the three aged men.

The Pangwa use iron rich magnetite in a mixture with iron poor limonite as ores. In the past Pangwa smelters mobilised a workforce, umkovi, and had members of the lutanana help them carry the ore from the mine to the smelting site. The iron ore, umdapo, that was used for our smelts was collected from an old quarry at Manga near Madilu village, in Lushoto towards Njombe.

In the past the master smelter’s lutanana would carry the ore walking up to 24 hours with heavy loads in baskets, kihivi, on their heads from Manga to Mkiu, and from Nyaluva to Mkiu in half a day. We were told that up to ten people would help them. Stirnimann gives an explanation for why two sorts of ore, umdapo was used. The one collected at Nyaluva his informants called, umyeveyeve or umunyaluhanga (Stirnimann 1979:267). This ore gave according to his informants more iron and less slag (see plate 5.12). The other, the one my informants claimed was the most important; the reddish clay like ore we collected at Manga, called umdapo,

This ore was one of two; the other was collected at Nyaluva close to Masimbwe village. The aged smelters had participated with different crews when they were active smelters. The father of one of our masters had been a master smelter in Ukinga before he moved to Magasiwalale in Upangwa. At Magasiwalale he continued as a master smelter and master blacksmith. Close to his compound was a small iron ore source that his son said had been used in his smelting furnaces at Kamaanga and Kunking’ande. The two Mtweve men had been active smelters in Masimbwe and Mlangali. Their father had also been a master smelter in Madoko but his furnaces

Stirnimann’s informants called it ungabuna. According to Stirnimann umdapo is the generic term for ore (ibid.). The smelters in our group insisted that both materials were iron ores and umdapo. They said the black one was inferior, but to me this one looked the most promising. In Upangwa the ores from Manga were famous and most people we interviewed said this ore was the most important to use for a successful outcome. Mgina who interviewed some aged smelters living close to Manga at Madole had been of the same opinion about the Manga ores as my informants. Unfortunately, his documentation lacks any discussion of two sorts of ores.

5.8 Collecting and preparing the ores.

67

Randi Barndon

Plate 5.12 Collecting ore at Manga

Plate 5.13 Ores prepared at the smelting site, Ing’angitoli 1993 smelting. One had to mix two sorts of ore, umdapo, in order to as they said: ‘make them be as one, like a man and a woman’. They used the generic term umdapo for both ores but added that it was a mixture of male and female ores that was needed for the furnace to be able to produce a bloom. The ore from Manga was washed several times in water in order to rinse off impurities and sand. After rinsing the ore several times it was left on green leaves for soaking and drying. This task was conducted behind the kwibepo hole of the furnace. Contrary to the Fipa they did not, to my knowledge give them different grades of qualities such as ‘much iron’,

Manga is famous for its iron ore source but the extension of the mine has been estimated to only 6m³. One liteende furnace was located 200 metres southwest of the mine. Mgina reports that this mine was used until the end of the 1950`s. The second ore was a black-brown surface ore collected at a small hill outside Masimbwe village. The place was called Nyaluva after a master smelter who had once lived in the area. These two sources of raw material were transported in our vehicle to the smelting sites, both to Ing’angitoli in 1993 and to Kuking’ande in 1995. At the site the ores were soaked and washed several times before they were left to dry in the sun until the day of 68

The Pangwa practice of iron smelting A conflict was evolving between the new master and the two aged men from Masimbwe. They claimed the Sanga master from Magasiwalale was not a real expert (fundi), of iron making because they did not think he knew which herbs and plants to put in the medicine pot for the furnace.100 On the other hand, this old man had demonstrated his power as a magician (and sorcerer) when an old man in Mkiu suddenly died. I was told the aged smelter was the man’s father-in-law. His daughter had divorced him and our master claimed back some cattle from the son-in-law. Some of the other men in our group believed he had bewitched the son-in-law since he was angry with him.

‘some iron’ or ‘little iron’. Thus they did not classify ores according to colours or iron content. By now only the tuyeres needed to dry and the smelters had to soften the goatskins for the bellows. They used some of the imono they had been preparing in the old pot at the site. The goatskins were oiled and tied with ropes. The following day they planned to start softening them with their feet and with stones. This was hard work and took almost one entire working day. When this was completed it remained to connect the skins to the drums and the drumsticks. 5.9 Smelting begins by magical preparations and sacrifices

One afternoon one of the Mteweve men had walked around the bush at Kuking’ande together with the Sanga smelter in order to collect the needed medicines. As the Mtweve man told me in secrecy, the other master had not seen the herbs and barks that they were supposed to collect even if they were just in front of him. Was he blind? Did he not know? The Mtweve man said he probably did not know them or their use and he had suggested to him that maybe the medicines they needed no longer grew near Kuking’ande. He proposed to walk back to Masimbwe and collect the medicines there and then bring them to Kuking’ande. This was of course a careful way of telling the Sanga smelter that he would not manage to collect the correct contents because he did not know them. From this confession followed that the two Mtweve men doubted the third man’s knowledge of iron smelting medicines. But he was a very respected and even feared man. As they said ‘look at the man in Mkiu, he is dead!’ They were afraid he would get angry if they suggested that they were the only true master smelters. Eventually the smelters managed to collect and combine the correct substances needed for a smelt.

I was told that successful smelting in a liteende furnace would produce enough bloom for two or even three hoes, which is also stated by Stirnimann (1976). The master smelters’ fathers had been respected men as were our aged men today. They had made the once admired iron bracelets that only healers, medicine men and smelters could wear and they used to make bells, spears, knives, hoes and axes. The use of magical substances, medicines and sacrifices were common among ironworkers in this region. When we first arrived in Upangwa in 1993 and talked to the aged men from Masimbwe who said they needed payments in order to demonstrate their skills to me because they had to sacrifice a goat before the smelting started. We agreed on payments, but since we had not completed the experiment in 1993 the sacrifice was never conducted. When I returned in 1995 the same issue was taken up again, as we were about to settle payments and plans for our work. The amount of money I had paid in advance last time we were in Mkiu was intact but they said the price for a goat had increased. We agreed that one of the young men should start looking for a suitable goat in Mkiu village. It was, as they argued, important to sacrifice a goat to the furnace itself and the masters and his lutanana’s ancestors. Because when the sacrifice had been given the master smelter could ask the spirits of the smelting site and his own ancestors to assist him in producing a good iron bloom.

The afternoon before the smelting was carried out in July 1995 a black he-goat was sacrificed. The sacrifice twitekhela, was carried out at the outskirts of the smelting site, north of the furnace. We had discussed the sacrifice of the goat well in advance. I had agreed upon the price presented to me for the goat that we bought from a distant relative of one of the younger blacksmiths in Mkiu village. When I inquired, the masters said ’of course it must be a he-goat, why give away more than necessary? Offerings were made to the ancestors, gods and the spirits whom I was told inhabited this particular place in the bush. The goat was not slaughtered but slowly strangled before it was bled. The importance lied in the actual bleeding of the goat, because the act of sacrificing was the bleeding itself. For a Pangwa sacrifice a goat is never slaughtered immediately, as they explained to me ‘then it is not a sacrifice but a killing’. No special announcements were made and it was almost as a mundane routine. One of the young men held the goat’s head with his two hands and while another one carefully strangled it and shot an

During all the preparatory steps in the process the old pot with beans, imono, was kept boiling over the fire at the smelting camp (see plate 5.11). While building the furnace and preparing the bellow system we had discussed the procedures in the process. We were told that magical preparations started a few days before the smelting while the goat was sacrificed the evening before the smelting started. We had discussed the pot, ing’anjo and its content and the smelters brought in more and more medicines from the surrounding area of Magasiwalale and their home area at Masimbwe.

100

69

Fundi is a Swahili term for expert.

Randi Barndon

Plate 5.14 A goat sacrifice, Kuking’ande 1995

Plate 5.15 The medicine pot, ing’anjo, in front of the furnace, Kuking’ande 1945

arrow into its neck in order to make it bleed. Then a calabash was put to the wound for the tapping of the blood.

slept at the smelting site while the rest of us went back to the village. Before we left the smelting site the young men were reminded about sexual abstinence in order to avoid getting hot. I was told that the special care taken the night before smelting started was in order for the master to enter into his role of a transformer, a magician. As a food taboo a Pangwa master who slept at the smelting site would only eat some of the meat from the sacrificed goat and the beer they had made at the beginning of the work.

Not all of the smelters cared to observe the sacrifice and those busy with the medicine pot were waiting for the blood and meat that they wanted to mix with the other ingredients. Once the strangling and bleeding of the goat was completed it was immediately skinned and butchered into pieces. Some pieces of the meat and liver were brought over to the smelting camp. One of the old masters said ‘Tambalale ndiksavahungila nye voho, nye va dadi! Ndiksave ivutali, munjimelile nye voha, Yemwafawvile!’ This I was told means ’Come towards me, you my forefathers, all of you! I greet you and beg you stand by me when I make my iron. Stay with me, all of you that have died!’ He offered them beer, beans, ugarali (maize porridge) and liver and meat from the slaughtered goat. He announced this were the smelters offerings to the unseen spirits of Kunking’ande and all their forefathers, mothers and fathers and their mothers’ and fathers’ grandfathers and grandmothers. The Pangwa masters prayed for the forthcoming smelt to be a success and that the ancestors and their forefathers would help them. Stirnimann (Stirnimann 1969, 1976:271) recorded similar prayers as those recorded by me in Upangwa 1995. The offerings of meat and beer were left beneath a musolo tree for the ancestors to collect. In the past the master smelter and his assistants slept around the furnace (not in the hut) the following night and protected the furnace against strangers or evil spirits. In our case the three aged men

Early the following morning the preparations for the actual smelt started when the medicine pot was filled with medicines. It was placed in front of the main tuyere inlet, umlomo, and the old men cut up the various medicines they had collected. These were mixed in the pot with the oil from the beans. The mixture, ing’anzo, was stirred with a stick that had the remains of its root at the end. The stick was then used to sprinkle some of the liquid medicines outside each tuyere inlet, on the furnace shaft, the bellow drums and the tubes. The three tuyeres were placed outside the doors and sprinkled with some of the medicines. The pot was then covered with a lid and placed within the furnace pit. The pot was packed and covered with termite soil. A new layer of medicines, especially kiwalasha was put on top before the entire pit was covered with more soil up to the ground level and the furnace shaft. The inside of the furnace chamber now measured 91 cm from the bottom to the mouth. Prayers were made to the gods and spirits as well as the master and his old assistants’ forefathers. Pieces of flour were set out for the ancestors together with some more calabashes 70

The Pangwa practice of iron smelting

Plate 5.16 Smelting in liteende at Kuking’ande 1995, note slices of kiwalahsa plant on top of ore

Plate 5.17 Smelting in liteende, Kuking’ande 1995, note the ‘windmaker-hole’, kwibepo

of beer and parts of the liquid content taken from the medicine pot. The master smelter’s wife was often a skilled potter. The master would instruct his wife on how to make the pot to be used during iron smelting. She would make it while the furnace itself was left to dry, but she had to keep the pot in secret and away from unclean or hot people. Even curious children could not touch the pot. She herself was not considered unclean; she was the master’s wife. As in Ufipa the master’s wife was trusted with the care of the most central medicinal and magical components in the smelting equipment. A master smelter’s first wife most certainly experienced the prestige her husband was given in village life.

The contents for the pot were carefully selected before being mixed with the oil that they had prepared from the beans. Long soft creepers, mukurumuko, were cut into about 15 cm long sticks and placed horizontally within the pot pointing towards each tuyere port. These were seen as phallic symbols (cf. also Stirnimann 1976:277). Hot oil, imono, and more herbs were added to the pot before the lid was placed on top. Stirnimann reports that the imono oil was seen as a fertility medicine (ibid.: 278). More medicines (representing semen and cures against male infertility and menstrual blood (ibid. 277-278), and charcoal was placed on top of the lid before the pot was placed inside the furnace pit through the front tuyere inlet (see Appendix F). Charcoal was packed around the pot. The hot steam from the medicines and oil in the pot should according to the smelters rise and sieve into the charcoal and ore in the furnace. Stirnimann was told that the tuyeres and tubes connected to the bellows were called the penis, ixibolo, of the bellow workers. The medicine pot was the ’womb’ of their ’furnace wife’ and called umudala. The white juices from the various plants mixed with the oil were the ’seed’, imbeyu, of the ’smelter husband’ (Stirnimann 1969:8). Stirnimann suggests that real semen was added to the pot but he was not able to confirm this and I was never confronted with this either. In addition to herbs and medicines the smelters added some handfuls of earth and some stones. Stirnimann explains that ’these were taken from where to paths were crossing each other and places in the bush where people had once had sexual intercourse’ (Stirnimann 1976:278). In this we see how the smelters combined the use of

The medicine pot used in our experimental smelt was an old pot one of the men had collected from his home in Masimbwe. This pot was named ing’anjo. It was an ordinary undecorated pot but its contents were special. A lid from an old broken pot covered the pot. The perforated lid consisted of eight holes. Just before the smelting started the pot was placed in the furnace pit. The lid was perforated with holes because they allowed the ingredients to escape into the furnace chamber with a correct timing and amount. The small holes (1.0-1.5 cm in diameter) on the lid allowed the medicines out of the pot and into the furnace womb during smelting. As I was explained the power of the pot and its content was the forces that divided the impurity of the ores; the slag from the iron.

71

Randi Barndon At 10.00 a.m. all tuyeres inlets and bellows were fastened and the smelting started. One of the old smelters placed a long wooden stick, also called kiwalasha, in the middle of the furnace chamber. This stick they had used for stirring in the pot of oil. The furnace was filled to the top with charcoal. Then eight handfuls of ore were placed on top of the pile of charcoal, four from the red ore and four from the black one. A few pieces of the umboocha were put on top of this arrangement. The pieces of umboocha were all pointing towards the tuyere inlets.

medicines with magical substances, a theme I shall return to in Chapters 6 and 7. The first smelt On the day of our first smelting we arrived at the site at 8.30 a.m. The aged smelters who had slept at the furnace site complained that we were late. Once we were there they started arranging the tuyere ports. They placed a broken tuyere underneath the one inserted into the main tuyere port in the front of the furnace. The tuyere, ing’elu was inserted into the furnace chamber in a vertical position at 10 cm within the furnace chamber. Wet clay was prepared and used to fasten the tuyeres and to cover the tuyere ports. At the other end, wet clay was again used to fasten the tuyere to the wooden tube, ivefi to a small hole, iwefia, in the wooden drum bellow All the three sets of tuyeres and bellow-drum arrangements were fastened with wet clay. Towards the tuyere there was open air, approx. 6 cm between the two wooden tubes and the tuyere.

Some of the aged men dressed in grass skirts, mtwele, or sometimes called msambi after the tree it was made from, normally worn by women (cf. Stirnimann 1976:300-301), and placed themselves in front of the sets of drum bellows (see plate 5.15). The master smelter started blowing the bellows. They then walked around the furnace from north to south but stopped at each door to blow the others’ bellows and they repeated this three times before the actual work started. I was told this was a common way to mark the beginning of the work. At 10.40 am when the lock, a wooden stick, was removed from ‘the wind making hole’ a blue flame started coming out. 10.50 a.m. one of the smiths used the wooden stick in the furnace to pull the ore partly down inside the furnace.

The goatskins were placed on top of the bellows. In order to fasten the goatskin-bellows to the wooden drums and again the sticks to the skin-bellows the smelters said they used a rope made from hemp. They claimed that only this rope would give them the strength they needed. A wooden stick was placed between the two drum-bellows to divide the goatskin with the rope into the two blowing sticks and bellows. While the final arranging of the bellows was carried out one of the old smelters made a fire with traditional fire sticks. As in other cases observed (e.g. Herbert 1993; Schmidt 1997) the smelters were not allowed to use modern matches and insisted on using the old fire sticks and dry grass.

There was a constant rotation in the bellow working. It was hard work and the smelters drank a lot of local brew during their breaks. While working the bellows they sang. At 11.00 a.m. the outside of the furnace chamber was slowly getting warm. Still, the smelters complained and thought the reason why we could still not see a blue flame coming up at the top was because of the cold weather during the last few days which had made the furnace cold and not dry enough. However, the smelters did not think this would be a problem. Fifteen minutes later, at 11.15 a.m. the charcoal had moved 15 cm down the furnace chamber and the ore was invisible because the smelters had pressed it downwards. Finally at 11.45 a.m. fire started to show from the top and the smelters were happy and thought this was a good sign. At 12.15 p.m. more charcoal was added and the furnace was refilled to the top and another eight hands of ore were added. 12.45 p.m. the fuel had moved downwards 5 cm within the chamber. Two of the bellow sticks were loose but quickly refastened. The back control hole was closed to speed up the fire again. The tuyere inlets were constantly plastered with wet clay. At 1.05 p.m. again a fire came out through the top and the wooden stick in the control hole was removed. The furnace wall was still not entirely warm at the outside; again the smelters assured me it was no problem. One hour later at 2.05 p.m. more charcoal was added and again the furnace was filled to the top. By now they had used two sacks of charcoal. Another ten handfuls of ore were added to the top. In total the smelters had now used 15 kilograms of ore. The master from Masimbwe told me he thought they had added enough ore by these three loads but he said they needed to add more charcoal. Ten minutes later a strong blue flame came out through

A Fipa master would often leave the bellow-work to the younger smelters or even give the boys a try, as would the Haya masters in Northern Tanzania, reported by Schmidt (Schmidt 1997:70). The Pangwa stressed the opposite, the older the better because they were afraid of the umboocha; poison from ‘the furnace guardian’. They stated that it was more likely that younger men were sexually active than old men. Some of my informants told me that only the elders were allowed to blow the bellows because in this way they prevented ‘the guardian’, from ‘cutting up stomachs’. At 9.00 a.m. one of the bellows was fastened and charcoal and fire was put to the furnace in order to test the combustion and force of the bellows. The bellow working gave only a weak sound of drums but the elders were satisfied and the work continued. While the rest of the bellows were fastened another pile of charcoal was put in to the chamber and at 9.20 a.m. two bellows were working. These two were blown while the third was prepared. The smelters told us this was done to speed up the heating of the furnace chamber before the actual smelting started.

72

The Pangwa practice of iron smelting

Plate 5.18 The front tuyere port, umlomo, with its tuyere inserted the peephole at the back. The tuyere inlets were replastered with wet clay. At 2.25 p.m. ’the wind-maker hole’, kwibepo, was closed to speed up the fuel from the top. At intervals the stick was placed and replaced in the kwibepo and the bellows were refastened. Below the tuyeres, outside the furnace wall the ground was by then packed with pieces of clay from the constant replastering. Some of the young men took over the drumsticks. At 3.20 p.m. the stick at the back door was once again removed and five minutes later a new blue fire came out through the top. The stick at the back was constantly removed to control the ventilation within the furnace chamber. The smelters constantly replaced each other at the bellows. At 3.25 p.m. one of the elders removed some of the clay around the front tuyere inlet, umlomo, to look for the slag, but he decided to wait for the charcoal to burn down to the level of the base. At 4.20 p.m. there was still no sign of the slag pouring out but no more charcoal was added. In total 2.5 sacks were used. Fifteen minutes later the charcoal had reached the top of the main tuyere inlet.

difficult to forge. Looking at the ‘bloom’ and discussing its possibilities for forging, the young men told me it would be almost impossible to use this in their new village forge. As we discussed the failure of the smelt the smelters said the slag had not properly parted from the iron. They thought this was because they had not added enough iron ore. They said the furnace should have been dryer. The aged smelters said ‘hot’ persons had participated in our smelt. They blamed the young men and said they would not have them as their apprentices any longer. They also complained that we had arrived late in the morning. They said ‘since you were late we did not find time to add more ore to the furnace!’ We discussed the possibilities of another smelt. They agreed but said the young smiths were not allowed to participate. Most certainly they were too hot and because of this they would not let the young men blow the bellows or come close to the furnace. We were told to arrive earlier and that they needed to find some other old men to assist in the bellow working.

The smelters decided to stop blowing the bellows and open the front tuyere inlet, umlomo. No slag had been removed through the main opening as they had said in our interviews. The ‘iron bloom’ was taken out in big lumps from the front door. One big lump of iron was attached to a tuyere. It contained several smaller pieces of iron, slag and charcoal. One of the old men put water inside the furnace and another big lump of ‘iron’ was taken out. The smelters were disappointed and they were discussing the final results from the smelt. The product from the furnace that they said was iron, ivutali, would certainly be

We are rarely are we informed about how smelters confront failures. The Pangwa smelters had however faced a failure. The young men in the crew were accused of breaking the taboo against sexual intercourse. The elders complained and threatened them. A very hostile atmosphere was the result. The aged men were magicians and one of them was especially feared in the village. The old master really gave them a bad time. But we agreed to try another smelting the following day. The young men were not allowed to participate. They were not totally excluded, but were made inactive and had to sit at the 73

Randi Barndon was added and the back door was closed with a stick. At the front the tuyere at the umlomo/libepo was partly closed because of slag. The old men said it was because this morning the young men had arrived at the site from this direction. The old smelters were complaining a lot about the behaviour of the young men. The said they had not respected the prohibitions. As one put it ‘in the past when they, the Pangwa, had guests they would offer them food and beer, while today when guests arrive they are crying’. At 12.15 p.m. another load of charcoal was added and filled to the top. They said that after one last load they would not add more charcoal.

outskirts of the camp the entire day. They were not allowed to assist. The master was convinced the reason behind the failure was the fact that the young men had assisted them in blowing the bellows. This had never been the custom in the past. Again they said at least one of them had slept with one of his wives or mistresses. The second smelt The aged men had once again slept around the furnace. When we arrived the furnace was cleaned from slag, iron and charcoal from the first smelt before we started. Most of the tuyeres were intact, only one was replaced. The furnace was once again filled with charcoal and a fire was made with the traditional fire sticks.

Again we were reminded of the possible breaking of taboos by the young men. The master said ‘you young Pangwa men never follow custom anymore, but you will see that in the future, we will not care whatever happens to you’. This was a threat. It became a rather uncomfortable situation, because we had arranged for the smelting in the first place this was setting the elders up against were against the young. On the other hand it illustrated how the masters tried to prepare for the second failure they sensed coming. The old men were getting tired. The sun was strong and they had worked the bellows for two days. At 13.05 p.m. the last load of charcoal was added and two hours later the smelting stopped. Again we were not able to make the slag run out of the tuyere at the front door. The smelters loosened all the doors before the ‘bloom-like iron’ was taken out. On close examination the smelters saw it as another failure. I was not able to establish whether it was a failure from a metallurgical point of view. The importance is that in this context the smelters classified the iron as not forgeable and blamed the young men for the failure.

The young men were not allowed to participate. They were not excluded, but were made inactive; they were sitting at the outskirts of the site the entire day. They were at the smelting site but not allowed assisting the old masters and the other old men in the blowing of the bellows. One of the old men was convinced the reason behind the failure the previous day was the fact that the young men had assisted them in blowing the bellows. This had never been the custom in the past. Moreover, he believed that at least one of them had slept with one of his wives or mistresses. He said ‘ as I see it, at least they had been walking around the village after sunset, drinking beer and talking with other villagers. Some of those people you met were at least not ‘cold’ enough!’ Three aged men from Magasiwalale joined the crew on this second day of experimentation with the past. The masters started the process by fastening the bellows in front of the main tuyere as they had done the first day. At 8.20 a.m. the smelting started by filling the furnace with charcoal. The tuyeres and bellows were arranged beginning with one, and blowing this one while the others were fastened. The wooden stick, kiwalasha, was once again placed in the middle of the furnace. Iron ore of both kinds was placed on top of the furnace (approximately 4 kilos). Ten minutes later, at 8.30 a.m. a flame came out through ‘the wind-making hole’, kwibepo. The smelters were pleased and said the furnace was dryer now. At 9.20 a.m. a flame came out from the top of the furnace. At 9.35 a.m. a new load of charcoal was added. The smelters said the charcoal they were using in the past was of a better quality. A strong flame was still coming out from the top. Four more handfuls of ore were added on top of the charcoal. The fire penetrated the ores and charcoal. Ever since the beginning the fire was seen at the back from the ‘wind making hole, kwibepo. This was not observed in the first smelt. The men were constantly blowing the bellows. At intervals they changed with the new comers.

Smelting ends Most experiments of prehistoric and pre-industrial iron smelting conducted outside Africa but based on African technologies of iron smelting technology have resulted in failures to produce forgeable iron. Espelund has pointed at how most experimental smelts conducted in Norway have resulted in failures. He suggests that this must be seen in regard of the fact that although one has knowledge of the furnace structures and operations in large those conducting the experiments have not been familiar with the rhythms and procedures of the process (Espelund, personal communication 1998). They have not possessed the ‘lived experience’ or the chaînes opératiores in making iron. I believe this forms part of what I witnessed in Upangwa. Iron smelting is a tacit, non-verbal knowledge (Barndon 1992; Schmidt 1997). Not only are the ‘secrets’ or magic and techniques hidden and stored within a small segment of society it is also tacit knowledge learned through many years of apprenticeship. Most of the smelters I interviewed had been with their fathers from ten or twelve years of ages and gradually learned the work. As an experienced adult smelter one cannot fully explain the

At 10.30 a.m. more charcoal was again added together with additional handfuls of ore ore. The master said that from now on they would only add charcoal. At 10.45 a.m. one of the smelters checked the tuyeres to see if air was still getting through. At 11.30 a.m. a new load of charcoal 74

The Pangwa practice of iron smelting combinations of maleness and femaleness as essential in procreation and conceiving. The furnace being a body needed male and female substances or powers in order to transform and produce. These were the charcoal, the ores and magical substances. The furnace was a body, a liteende, ‘something beneath’ and not only ’the one that smelted’. It was the one that contained the substances necessary for ore transformations. According to Stirnimann the smelters slept close to the furnace the night before the smelting started and the master masturbated and placed some semen in the medicine pot (Stirnimann 1969). The pot with its content was placed inside the furnace womb through the ‘mother opening’ umlomo (see plate 5.16).

process verbally and the chaînes opératoires can only be demonstrated through actual behaviour, like bicycling. The technical operations carried out by individuals and groups of ironworkers were like ritual songs and dances informing on the correct rhythm and balance of technical behaviour. This is fundamental in the bellow working and charging of the furnaces, incorrect rhythm and balance will not provide the correct temperatures and combustion. It will not provide correct air draft in the furnace and will result in unforgeable end products. Many things may have caused the failures we observed. The smelting process with all its preparatory stages was not identical to past chaînes opératoires. We had speeded things up, made new chains or changed operations within sequences and hence created new sequences. Whether we ‘truly’ faced failures will have to await metallurgical analysis of the products.

The chaîne opératoire diagram in figure 5.1 illustrates that the entire process was conceptualised in terms of metaphorical constructions based on the body. The magical preparations, sacrifices and mandated sexual taboos guided the participants’ behaviour and informed on hoe the sub-sequences followed each other. Further, the diagram illustrates the importance of kiwalasha. People, charcoal, ores and the furnace were washed in purifying water and medicines were sprinkled on the furnace’s orifices.

5.10 The chaînes opératoires in the smelting process Once a Pangwa furnace was built it was decorated with female breasts, mawele, and the smelters told me it was a young woman, fertile and ready to deliver. The smelters’ decoration of female breasts on the furnace followed their conceptions of gender powers and the gender

75

Randi Barndon

Figure 5.1 Chaînes opératoires in Pangwa iron smelting 76

The Fipa practice of iron smelting

6. The Fipa practice of iron smelting Kalyango na kama, kalyango na kama!101 6.1 Introduction101

he placed on the termite mound overlooking the smelting site (Wise 1958:234).

Contrary to Pangwa iron smelting, the Fipa smelting procedure is well known among Africanist archaeologists and have been documented numerous times. Primary sources on Fipa iron working such as Grieg (1937), Lechaptois (1913) Wembah-Rashid (1969, 1973), Wise (1958a, 1958b), Willis (1969, 1981), Wyckaert (1914) and the most recent Barndon (1992), have been constantly referred to (e.g. Cline 1937; Collett 1993; David 1999; Childs 1991b; Childs and Killick 1993; Herbert 1993; Mapunda 1995; MacLean 1995; Merwe and Avery 1987; Schmidt 1996a, 1996b 1997; Schmidt and Mapunda 1997; Sutton 1985, 1990; Wright 1980, 1982). Lechaptois and Wyckaert conducted the first documented observations that specifically focused on the smelting and ritual practices (Lechaptois 1913; Wyckaert 1914 and see Appendix D, table 1). It is uncertain if Lechaptois ever observed the sequences of the entire process or rather went to and from the smelting camp during the smelting season and also partly relied on informants’ testimonies. Lechaptois reports that the wives of the smelters were allowed to participate in the building of a furnace by smoothening it with wet clay (Lechaptois 1913:242). Wyckaert on the other hand gives a detailed description of the entire process and includes the secondary ichinteengwe that in most later reports has been underestimated.

Wise and Wembah-Rashid, opposed to Greig and Wyckaert, did not observe the smelting procedures in a standing natural draft furnace but the procedures following the building of a new furnace. Mapunda has conducted archaeological research on iron technology in the Rukwa region and discovered a different smelting practice from the plateau iluungu furnaces. Along Lake Tanganyika smaller shaft furnaces, kataruka, were used prior to the maluungu (Mapunda 1995). I have previously presented the Fipa smelting process in Traditional iron working among the Fipa (Barndon 1992). This study was based on three re-enactments of Fipa iron working, one in 1990 and two in 1991 in addition to numerous interviews and surveys. As discussed above the present work is a comparative study of Pangwa and Fipa iron working and in order to embark on this comparison the two ’practices’ must be equally described and transcribed. In the following pages I will provide a step by step description of the smelts that I observed in Ufipa in line with my description of the Pangwa process in Chapter 5. 6.2 Mobilising a labour force In Upangwa the smelts that I observed were initiated because of my research, something which also formed the setting in Ufipa. Hence, what we observed was not identical to the past active days of smelting, because the smelters were not in need of iron as they once were. In the past a Fipa master smelter, umwaami, would carefully select his group of assistants and helpers when he first inherited the master smelter title together with the basket of magical substances, intaangala. Over the years, as a master, he included new members or excluded unwanted persons from his crew. Prior to each smelt he called upon his assistants and included in his labour force only those who had not eaten forbidden food or committed adultery (i.e.Wyckaert 1914:371).102 Among the women he selected only those who were not pregnant, breast-feeding or in menstruation. Social recognition was important when making a furnace, which was a communal task among relatives and led by the master smelter, but if people did not have confidence in the knowledge and forces of the master smelter they would not assist him (Greig 1937:77).

Grieg observed iron smelting in Ufipa in 1936 and he then recorded that iron working was still extensively practised on the plateau (Greig 1937:77). Grieg claimed that the smelters were an important occupational group and that master smelters were highly admired (ibid.). To my knowledge neither Wyckaert nor Greig observed smelting in new furnaces and therefore they did not record the celebrations of the new built furnace. In 1956 a smelting furnace was raised in front of the District Commissioner’s office in Sumbawanga and the smelting procedures were recorded by Wise (Wise 1958a, 1958b). Wembah-Rashid made a re-enactment of Fipa iron working at the Village Museum in Dar es Salaam in 1967 (Wembah-Rashid 1969, 1973:65). Wise describes how he was presented with another practice that has not been recorded in any of my reenactments. In his account which otherwise is similar to my observations the first day when the work was about to start the master smelter made a clay model of a bull which 101

These words were said when the sacrifice of a cockerel was conducted at the furnace foundation. It means as the master reluctantly told me ’give us openings and ways out!’ and it was repeated several times.

102 Or he would walk around the village asking people to join him (Greig 1937:78).

77

Randi Barndon During our brief trip to Ufipa in March 1990 we talked with the aged master smelter who agreed upon the importance of conducting a smelt and, as he said it would be a record for future generations of Fipa. He told us he would not have problems organising a group of iron smelters and a workforce as long as they were given enough beer and food. For our first re-enactment the old master called together a group of men that had once been his assistants and these smelters came along with their sons and nephews. What we were about to embark on was an important social event because we were told that the master smelter planned to hand over his basket, intaangala to a younger and active blacksmith. During the last years this man had demonstrated his skills in the village forge. In the past a master smelter and his crew of a minimum of five smelters, but often seven or nine, formed a group of iron smelters, asiluungu.103 The mobilised labour force for some of the most demanding tasks in the process was, contrary to among the Pangwa, a lot larger.104 Groups of fifteen to twenty young men, for some tasks including young women, were not uncommon (Barndon 1992:125). This crew expected to be provided with beer and food in return for their labour, very much like a general threshing party. The smelters, asiluungu, were given one load of iron each from the smelting season’s production. A master smelter could, as they expressed it, merely ’call back his hoes’ because all iron ever produced under an umwaami would always belong to him. This resulted in the smelters being tied to a life long dependency relation with their master smelter who was their ’king of the furnace’, mweene iluungu. I think that for our experimental smelts the old master smelter used this old dependency relation that had once been active in order to mobilise the other elders for our re-enactment. He ‘called back his hoes’ and the elders who once were his assistants brought with them their sons and nephews because of the heavy workloads.

6.3 Smelting in secrecy, guarded by the umpakasi During pre-smelt interviews I was told that in the past once the master had selected his crew they would walk together out of the village and settle at the smelting camp. The master’s eldest wife and children would come along while the rest of his wives and children and the wives of the other smelters remained in the village. The master’s wife would cook for the crew during the smelting season. The smelting camp was always located close to the furnaces while the furnaces were not far from the village (cf. Wyckaert 1914). In the past iron making was carried out in the dry season from June until August (Barndon 1992:50). We were in Ufipa during these months in 1990 and again in 1991. Both Pangwa and Fipa marked the beginning of their work by walking in a parade out of the village making everybody aware of the beginning of the smelting season. In the active days of iron smelting the entire crew would mark the beginning of the work by colouring their foreheads with a red powder, unnkolo (cf. Appendix B). They left the village in a parade carrying with them the magical basket and the medicinal barks selected from various trees and a cockerel, nkoko, that two children would sacrifice to the masters forefathers at the furnace foundation. In the bush they would build a small hut and remain at the smelting site until sufficient iron had been produced and the smelting season had ended. In Ufipa in 1990 I witnessed how the master smelter with his apprentice and old assistants coloured their foreheads with the red unnkolo powder and walked through the village announcing the first day of smelting. Behind them followed the two ‘blind’ children who were going to perform the sacrifice.105 The parade carried the intaangala basket, the ifingila magical substances and baskets of ingailo, medical barks that they had collected the night before. They carried with them a cockerel, nkoko, which they would sacrifice to the furnace and their forefathers.

Never during our visits to Ufipa did any of the members of the crew ask for money or other means of compensation for their loss of labour investments in their own interest; their fields, gardens or the forge. They never complained about the heavy work and long working hours. Before the work started a cow was slaughtered and the meat was hung from the roof in one of the rooms in our house until, when the furnace was ready built, it was prepared and cooked. It was a lot of meat, as the master smelter happily informed his smelters and the larger workgroup that we would offer them on the day of celebrations of the furnace. The women in our compound, the wives of two of the sons of one of the old apprentices, prepared the food and the beer for the party.

Once a production site, unsakwe, had been selected and medicines and magical treatments prepared, the raw materials for the building of the smelting furnaces started or the smelters conducted the repairs on standing furnaces. The various raw materials needed to produce iron, such as termite soil, ifo, for the furnaces, clay for the tuyeres, wachelu, (sing. inchelu), iron ore, inyiimbo, firewood, kateembwa and charcoal, ikala, as in the Pangwa case, did not guide the smelters’ selection of a smelting site.106 The furnaces were never purposely located close to mines or quarries in the surrounding landscape, rather young men and women were mobilised to carry the heavy loads of water, iron ore, charcoal and tuyeres to the smelting camp.

103

The group of smelters, asiluungu, included the master, umwaami and his two aged assistant, aluumba, the master’s assistant, unnsole, and his two assistants, kampaeenga in addition to other smelters (Barndon 1992:123-124). 104 Cf. Appendix D.

105

Ictifiti literary means black but refers to blindness, which is a way of saying a person is sexually immature. 106 Cf. discussion of Childs’ list of sequences in Chapter 4.

78

The Fipa practice of iron smelting substances it contained and why they had to be applied. It was believed that without the contents of the basket iron smelting was impossible because the medicines, ingailo, would not work and the furnace would not remain standing. But the Fipa smelters applied two major components of magical and homeopathic substances in general as a cure to prevent misfortunes and provide success. The most important component was the application of a collection of barks, ingailo, cut from various trees (Barndon 1992:87).107 The ingailo were the true medicines but without magical assistance from the intaangala they would not work (cf. Cory 1949; Willis 1978b)

A Fipa smelting site was most commonly selected in relation to a termite mound, ulusuumba, which are abundant on the Ufipa plateau. It was located on the western edge of the mound the soil of which was used for the furnace constructions. It was common to select the smelting site in relation to one of the sources of raw material (Childs 1991b; Schmidt 1997:69), while the importance of the master being able to mobilise a large working force to assist in ore and fuel preparations and transportation was more crucial. Our first smelting site was within the village, visible from everywhere and not far from the village mill that was visited daily by women. Next to the mill was the village forge. During our first visit to Ufipa in March 1990 we had told the master smelter and the active blacksmiths that we were coming back in May. But a student’ revolt that culminated in the University of Dar es Salaam being closed delayed us for several weeks. While we were away a disagreement occurred between the old master and the staff from the Regional Cultural office about the location of the smelting site. The head smelter was worried about the location of the furnace within the village because, as he later told us, too many ‘hot’ people would be around, they would even enter the site. Nevertheless, the furnace that was built for our first observation was placed within the village, not far from the mill and the blacksmith’s village forge. How we managed to make iron in spite of ‘hot’ women passing by the site was never really questioned.

When the Fipa smelters had established them selves at the camp they first went out in the bush to collect the primary medicinal bark, ingailo. These are known in general among the Fipa for their curing qualities, but in other contexts when curing people they are called ileembo (cf. Willis 1978b:139, 144). I was told that in the past the master smelter and his assistant walked out in the bush together with one or two of the elder advisors, aluumba. All the pieces of barks were kept in a basket until the day they started the building of the furnace. The barks were put in the foundation trench or inside the furnace chamber. A few pieces were made into ‘the furnace caretaker’; a bundle together with some magical substances and the head of a cockerel, inside the furnace chamber (see plate 6.2). Wise explains how the use of these barks was related to the Fipa concepts of trees as spirits (Wise 1958b). In Fipa language the major god is called kaleesa or amaleesa, the god of large trees (cf. Willis 1981). The smelters told me that the barks were used because of their medical qualities and this was related to the colours of the berries, fruits or juices they yielded, where they were growing or if they were growing together, like creepers around another tree. When we collected the bark in the bush before our reenactments the effect they had on various diseases and the colour of fruits and berries were given as the reason for them to be collected. Some were collected because they had many fruits and the smelters said ‘then our furnace will yield many blooms’. Other kinds of bark were collected because their juices were white, resembling semen or breast milk. Other ingredients were medicines known to treat infertility, stomach problems or venereal and invisible diseases (Barndon 1992:96-97). The pieces of bark were always cut from both the western and eastern side of the tree and according to my informants fruits or leaves were never used in smelting because it was difficult to determine which side of the tree they were from (ibid.: 96).

In the days of active smelting, sticks with white leaves fastened on the top or two sticks with a rope tied between them were placed at every track or crossroad leading to the furnaces (cf. Barndon 1992; Roberts 1949). My Fipa informants claimed this made it possible for them to locate their furnaces close to villages and tracks in the bush leading to gardens and fields (Barndon 1992:116). In addition, the Fipa had a ‘furnace caretaker’, umpakasi, placed inside the furnace chamber; the head of the sacrificed cockerel tied in a bundle of medical barks with small pieces of magical substances in its mouth. One of the wings of the sacrificed cockerel was also called an umpakasi for the maternal opening and major tuyere inlet, palinyina, pointing towards the setting sun. It was placed above the main tuyere port and furnace entrance and called pamweela palinyina which literally means ‘the loin cloth for the furnace maternal opening’ (ibid.: 105). 6.4 Performing magical skills A master smelter legitimised his right to build a smelting furnace and lead a crew of ironworkers through his possession and knowledge of the basket of magical substances, intaangala, which was owned by the master smelter and transmitted from father to son or other close relatives, such as a nephew. The magical substances of this basket were considered vital for successful smelts and the master smelter was the only person who knew which

107 Pére Robert at the White Fathers Mission reported that a Fipa head master had told him that ingailo (mashina) samples should be taken from the eastern and western sides of the trees. He stated that the species were of lesser importance than the part of the tree the sample was taken from (Robert 1949:243).

79

Randi Barndon

Plate 6.1 The furnace caretaker, umpakasi, hanging over the main tuyere port on the high shaft furnace, Katumba 1991

Plate 6.2 The furnace caretaker, umpakasi, inside the furnace chamber, Katumba 1990. Note the bundle of medical barks, ingailo

Willis describes how a regular Fipa doctor, sing’aanga, uses the same sorts of medicinal herbs and barks as primary treatments in curing sick persons.108 The barks and herbs are used in a combination with magical substances from the medicine man’s intaangala basket (Willis 1978b and see Chapter 7). Willis notes that the colours red and white are symbolically significant for the Fipa (Willis 1967:526). Redness, ukasuke, is associated with a fruit called isuuke, that the intruding mythical Twaci women ate. Red may be a symbol associated with blood and women but also war and violence (ibid.).109 White on the other hand, is symbolic of spiritual power and peace and associated with male authority (ibid.). The duality of Fipa colour symbolism is linked to human physiology (cf. Beidelman 1968:309). As I will demonstrate below this duality was incorporated into iron smelting symbolism and part of the medicines used in curing the furnace before smelting.

out for a smelt. Before the building of a new furnace started the entire contents of the intaangala basket were taken out. Small pieces were cut off before the item was returned into the basket. Items that the master smelter identified as purposeful and particularly important for the smelt were explicitly given a meaning (by use of analogy or metaphors). Other items were used additionally for a more general strengthening of the furnace foundation. This explanation relates to the fact that the master smelter did not identify all the items. The master of our crew had inherited his intaangala basket from his father. It had been in his family group, sinkaamba, for five generations. The master knew only some of the contents, because as he claimed some ifingila had been there from before he was born while others e had added himself while he was younger and he could not remember them exactly anymore. The night before the smelting started the master opened the magical basket, intaangala, and cut off a small piece from each item he considered important for the proceeding smelt, the rest was returned to the basket. In our re-enactment the entire group of smelters observed the cutting and chopping of the medicines and the master’s assistant, unnsole, was told the significance of each object. The old advisers, aluumba, suggested items to be added or removed from the selection. The selection of magical substances contained different pieces of animal bones, skins and hair, birds and feathers, stones, slag, bark from various trees, shells, powders (see plate

The ironworkers were in the past organised in a hierarchy of skills, knowledge and duties during a smelt. This hierarchy was related to an unevenly distributed knowledge of magical substances. The master smelter and his assistant, unnsole, and the two elders, aluumba, were the key figures also when magical preparations were carried 108

Ing’aanga means a body of knowledge, handed to a practitioner (sing’aanga) by a similarly qualified teacher (Willis 1987b:139). 109 Isuuke was one of the ingredients of the collection of bark and herbs placed in the foundation trench (Barndon 1992).

80

The Fipa practice of iron smelting

Plate 6.3 Sorting magical substances, ifingila, in front of the furnace, Tupa 1991 substances would procure benefits and prevent harm during the smelting (Cory 1949:16; Willis 1978b:145 and see Appendix E). If some of the substances were missing the master would add or replace them with other items or substances of equal importance. According to observation and statements by informants, nothing new was added for the smelts I observed and the master smelter claimed he was pleased with the ifingila as it was (see Appendix E).

6.3). Not only herbs, but also bark and animal bones were parts of the magical substances as were pieces of old cooking pots, pieces of old hearth stones, charcoal and pieces of slag from previous smelts (ibid.: 94-93). The magical substances selected from the basket were when applied in the smelting, called ‘things which enter’, ifingila, or ‘latent things’, ifisiimba. These substances were used when a new furnace was built and placed in the foundation trench. When the smelters re-used a furnace the substances were mixed with ingailo and placed inside the furnace chamber.

For our re-enactment the master poured out the magical substances on a goatskin. We were told it had to be the skin of a black he-goat but the outside was turned downwards facing the ground. The skin was called imbuusa, which literally means ‘not to be seen by men’ and by this they meant the power of the ifingila was invisible. I was informed that the black skin told the smelters that there were things in the basket that they could not see. The various substances were sorted out and the master smelter told observances or facts about the animal or bird it represented (see Appendix E). One of old advisors and unnsole were then appointed to cut up the substances and they coloured their foreheads with red powder, unnkolo. The two appointed men took one at a time an item from the goatskin and chopped off a little piece of each before they returned the rest into the basket. The magical substances were chopped with iron axes on a wooden pole, which was put in front of the goatskin. The pieces selected were put into two palm leaf containers. The entire process took approximately 2 hours. When they had completed the selection of substances the aged master smelter announced ‘take this ifingila and keep it as it is’. By this expression he meant it should be kept as they said

Cory notes that this division between medicines and magic is widely used and often the actual magical substances are secondary applied in order for the medicines to acquire magic quality (Cory 1949:16). In Sukuma language the term chingira is according to Cory developed from the root of the word kwingira – to enter, hence ‘medicines without shingira (Sukuma) or ifingila (Fipa) is vacant, like an empty room without furniture’ (ibid.). Because without ifingila, ‘the primary ingredients would be nothing but roots, twigs leaves or barks’ (ibid.). During my fieldwork in Ufipa in 1990 and 1991 the contents of the intaangala basket were discussed with the master smelter and his closest associates as well as younger smiths. Local exegesis indicated a strong use of analogies. The actual item itself was not curative but the quality ascribed the animal or stone the item represented would give the wanted effect. Thus by use of analogies, or as Cory would call it ‘the law of similarity’, the magic 81

Randi Barndon ‘at a safe place beside the bed of a man who is ‘unthirsty’ and considered too old to be tempted at having a woman beside him during the night’. In our case the basket, the goatskin and the two green containers of ifingila, together with the ingailo were given to the master’s assistant. He was told to keep the magical substances under his bed until the master came to collect him the following morning. The basket was kept open and in the lid they left the red powder, unnkolo, a piece of rock crystal, ntaanda, and some iron, ululu, from a previous smelt to chase away sorcerers. Before each smelt observed the bark was collected in the bush and items from the intaangala were carefully selected before they were chopped and cut up. Prior to the first smelt the magical basket was brought out in front of the new umwaami’s house and prepared the night before the building of the furnaces started. For the second and third smelts the basket was opened in front of the furnace. During this work only the smelters were present.

soil was dug out and left in a large mound close to the foundation trench. As commented by Wembah-Rashid several tons were used, resulting in a large hole in the termite mound (Wembah-Rashid 1973). The mobilised work group cleaned the soil for stones and hard lumps of clay before they paddled it with their feet and mixed it with water. In the midst of the large pile of soil a hole was dug out and water added gradually. They sang while working and one of the aged smelters led the singing and the work. While the mobilised work force prepared the soil the smelters cleared the smelting site for trees and scrub and laid out the measurements for the furnace trench. The master smelter dug out the furnace trench. As observed in Upangwa, the Fipa master determined the size of the furnace using his own body as reference point. A stick called a ‘wedding stick’, entoa, measured the length from his arm to his fingertips. The stick was laid down on the ground and formed the diameter and the circumference of the furnace. On the ground he placed his measuring stick and fastened a piece of rope from the stick and around his right foot. Then he dragged his leg around the horizontal stick to make an exact circle. This formed the outer diameter of the trench and furnace wall. The other smelters dug out the trench with their hoes and used the master’s foot as a measure reference. The size of the trench was exactly the size of a Fipa male hoe, ise, which is 25 cm thick and 32 cm deep (Barndon 1992:55 and cf. Appendix D). The edges on the foundation trench were as straight and sharp as possible and the smelters went into great efforts in making them neat and clean looking.

Contrary to the Pangwa who during the building of the furnace and other preparatory tasks also prepared the imono beans and went out on smaller expeditions to find suitable medicines and magical substances, the Fipa conducted all their magical preparation and sacrifices before the actual building of the furnace started. Therefore, before the furnaces were built and ores, wood, charcoal and tuyeres made, they prepared the medicines and sacrificed to the furnace (see plate 6.3). The Pangwa magical and technical preparations culminated in a sacrifice of a goat the night before the smelting started followed by offerings of beer and food to the ancestors and spirits of the site. Before the furnace was charged with fuel and ores the pot of medicines was placed in the pit beneath the furnace chamber. In Ufipa on the other hand, both the preparations of medical and magical substances as well as sacrifices to the furnace or ancestors were conducted before the building of the furnaces started or before repairs were conducted on standing furnaces.110 Once the ingailo and ifingila substances were ready to put in the foundation trenches (or inside the furnace chamber) a young girl and boy were called in to assist in the sacrifice of a cockerel.

When the soil was ready prepared and the trench, ulufu, dug, a sacrifice for the furnace, yakasike iluungu, was conducted. Two children (a boy and a girl) placed the medicines and magical substances in the foundation trench. They were told to stand inside the circle (or in front of the furnace) while the master’s assistant handed the girl the cockerel. The boy cut off its neck and sprinkled its blood in the furnace trench, on top of the medicines and magical substances (or above the tuyere inlets). While the trench was covered with blood the master smelter announced repeatedly: ‘Kalyaango na kama!’ which literary means ‘make sure all the doors are closed’. The cockerel’s head was kept aside with the intaangala basket and immediately before the furnace was filled with ores and charcoal it was placed in the centre of the furnace chamber. In its mouth the head smelter placed some bones of a lion from the intaangala basket and medicinal bark, ingailo. This little umpakasi was pointing with its mouth towards the eastern tuyere inlet, the ‘the father door’, palisi, and as I was told, the rising sun. The smelters told me the cockerel would mature and crawl like a hen once the smelting started (ibid.: 104). If the furnace was an old one no further rituals or sacrifices were conducted for the furnace or at the smelting camp. My informants told me that the children would help the magical substances to work and especially for a sacrifice

6.5 Building the smelting furnaces in Katumba The first stage in the Fipa smelting process was in the past as in our re-enactment the mobilisation of a labour force and the preparations of the magical substances. When these tasks had been accomplished the building of the natural draft furnace, iluungu, started. The furnace was built out of soil, ifo, from termite mounds. Building a new furnace, ukutwa illungva, was observed only once, for our first re-enactment carried out in August 1990 (cf. Barndon 1992:54-56). Up to twenty people participated. Preparing the soil for the furnaces was a heavy task, the 110 An aged smelter from another clan claimed that in his crew the ingailo was soaked in water and left until the next day when only the water was sprinkled inside the furnace.

82

The Fipa practice of iron smelting

Plate 6.4 Yakasike iluungu, a sacrifice for the furnace, Katumba 1990 at the foundation of a new furnace both a boy and a girl had to participate since the iron produced would be used by both men and women.

Ichinteengwe The secondary refining furnace, ichinteengwe was built immediately after the high shaft furnace had been completed. In our re-enactments this furnace was always built north of the natural draft furnace. The ground was cleaned for scrubs and stones before a big outer circle was made. This marked the workspace around the furnace and a foundation for the open hut with a grass thatched roof built to protect the bellow workers against the strong sun. A smaller circle of 40 cm was made and formed the foundation trench for the furnace. The foundation trench was dug out with a male hoe and was only approximately 10 cm deep. Inside the trench the smelters made a small pit for magical substances. Once they started building the shaft the smelters were careful not to cover the small pit with wet soil.

After the sacrifice the master smelter told his assistants and helpers to fill up the foundation trench with soil. The young smelters run to the foundation carrying the wet soil in lumps in their hands throwing them into the trench with as much force as possible. This went on until they had filled up the trench with the wet soil. Thereafter the clay was placed in its circle with more care by handing lumps over to those standing close to the trench. The smelters started forming the wall with their hands. When they had reached up to the length of 75 cm they left the furnace for approximately two hours since the wet clay needed to dry out before they could continue the work. When the furnace shaft had reached the length of one entoa, or 150 cm the smelters ended the day’s work. Long wooden poles were put up around the outer wall and connected to straightened out vertically laid poles around the furnace. These were fastened with water-soaked palm leaves. The purpose of this was to have a scaffold since the furnace was too high for them to reach the top if standing on the ground. The next day the work continued in the same order, the same people still preparing and softening the soil while the smelters constructed the furnace shaft. Two men were inside the furnace plastering the inner wall of the furnace chamber. By the end of the second day the furnace shaft was completed.

When the shaft was 40 cm high the furnace was considered big enough and the smiths made the wall smooth with their hands. The technique used in building this furnace resembled Fipa pottery making. In opposition to the doors on the natural draft furnace, the tuyere ports on the secondary furnace were made immediately after its superstructure was completed. The master took a small wooden stick and measured the length from his upper arm to the joint to get the right size. The stick was then placed in on top of the furnace. The master stood on the eastern side of the furnace, close to the termite mound and pressed the stick straight over to the western side where another smelter was standing. The purpose of this was to get the right measurements and dimensions on 83

Randi Barndon

Plate 6.5 The secondary furnace, ichinteengwe, when ready built, Tupa 1991 the tuyere ports. The middle tuyere port, palisi, had to be exactly opposite the rake hole, palinyina, facing the setting sun. The master pressed the stick downwards through the eastern furnace wall and it reached the western side of the furnace 15 cm up from the furnace base. The tuyere ports and rake hole were then opened with a knife; one round hole on the eastern side, two more of the same size north and south and the rake hole at the base facing towards west.

Opening the tuyere ports The high shaft furnace was big, normally two to three metres tall and contained a large amount of fuel and ore. Our furnace had ten air inlets evenly distributed around the base.111 The tuyere ports on the furnace were cut out once the furnace was completed with new scaffoldings and the clay had dried up. In order to position the tuyeres in the correct direction the master smelter first made the small peephole, ntaanda, on the eastern side of the furnace. The peephole was cut pointing downwards into the middle of the interior of the furnace chamber (Barndon 1992:58).

The furnace was dug out on a sloping hill and I was told this would facilitate the slag tapering through the rake hole during the smelting. Outside the rake hole the smelters made two elongated clay masses called impaamba, which literally means ‘the inside of female thighs’. This formed a clay channel that was extending out from the rake hole. Water was poured through the top of the furnace and ran through the chamber and out through the rake hole. The entire furnace was smoothed with water and covered with palm leaves and thorn bushes. The building of the secondary furnace took four smelters approximately two hours (Barndon 1992).

The location of the main tuyere inlet and furnace port was then determined and cut out. The master used his measure stick, entoa, to establish the exact location of the western tuyere inlet. Using a rope he found the outer circumference of the furnace and divided this into two. The master smelter used a knife to cut out the shape of the door while his assistant completed the work. The termite soil was still wet five cm in through the wall. The masters assistant, unnsole, took over the knife and completed the tuyere port. The smelters used a rope of palm leaves and stretched it around the base of the furnace wall, bent it in half and this was where the major tuyere inlet was opened. The rope was led from the eastern door to the western side of the kiln where umwaami again cut out the dimensions.

Prior to each building of a secondary furnace a secret use of medicinal herbs and magical substances was applied. The master told me that only the master himself, his assistant and the two advisors knew of this use of magic. Lion bones and a collection of barks were made into a small bundle that was dug into the furnace pit. The other smelters were at the smelting site but excluded from observing the act and not specifically informed about the medicines.

The other smelters made the rest of the eight ports but they used the master’s stick as a measure in between the 111

84

Cf. Appendix D. tables 5 and 8.

The Fipa practice of iron smelting

Plate 6.6 Celebrating the high shaft furnace as a bride, kanawiinga, Katumba 1990, note scaffoldings and flowers

Plate 6.7 The furnace in our re-enactment at Tupa 1991, note permanent scaffoldings

doors. The tuyere inlets were given approximately the same size and shape as the eastern tuyere port, hence they were 50 cm tall and 20 cm wide. The major tuyere inlet was 56 cm tall and 40 cm wide at the base big enough for a man to enter into the furnace. The biggest tuyere port always faced west and was called ‘the mother door’ or ‘mother opening’, palinyina. This opening was used as an entrance when the smelters arranged the tuyeres from the inside of the furnace chamber and when raw materials were loaded into furnace. Opposite and facing the rising sun was what the smelters called ‘the father door’, palisi, sometimes also referred to as ‘the beginning’ or ‘the place of burning’ (Barndon 1992). New permanent scaffoldings, mizanti, were made. Nine wooden poles, miaandi, were inserted deep into holes in the ground around the furnace wall while creepers, mikoa, were laid in two horizontal circles around the furnace shaft.

When a new furnace was completed and before the mobilised labour force was provided beer and food as payments for their efforts, the furnace was celebrated. The smelters decorated the furnace. In the past these decorations had been made with pounded charcoal, or millet flour referring to the furnace as strong but still innocent and pure. In our re-enactment it was decorated with white maize flour. The smelters and aged wives of the old asiluungu as well as the younger men and married women who had participated in the building of the furnace sang and danced around the furnace. They were abruptly stopped when the master told one of the young men to leave for the forest to chase a hare, kalulu. He ran away from the furnace site but came immediately back with yellow flowers, kalulu with which the smelters decorated the top of the furnace (see plate 6.6).112 According to Fipa wedding traditions a young woman will not become a kanawiinga, a marriageable woman before she and her bridegroom have conducted a ritual of intaakaso. Intakaaso is ritual washing and the marrying couple decorates their bodies in a mixture of clay and hot water. The relatives of the young marrying couple celebrate the coming event the night before wedding celebrations start. They drink beer in the house of the

The building of a Fipa furnace took according to my observations two working days starting with the initial collecting and preparation of the soil until the furnace was completed. This has been reported in Wise’s observations (Wise 1958a). In our re-enactments all the preparatory tasks were completed approximately a week or a week and a half later. In the past a new built furnace was probably left longer to dry (i.e. Greig 1937; Wyckaert 1914). Wembah-Rashid who observed the Fipa process in a different context, at the Village Museum in Dar es Salaam, recorded that the building and preparations took more than three weeks (Wembah-Rashid 1969, 1973).

112

Other observers of Fipa smelting rituals (i.e. Wise 1958b) have not questioned the ritualised chasing of the hare. The ritual performance may represent a local innovation or practice initially among the Mambwe or Nyamwanga (cf. Chapter 9).

85

Randi Barndon woman’s relatives and make two clay figurines; one male and one female. The following morning the young couple will take these figurines in to bowl of water, akio. The water is hot and the two figurines dissolve in the bowl of water. Then the young woman and man decorate, yaeemba, their bodies with water and red ritual earth, unnsaa, from a nearby anthill. Now the young woman is a true kanawiinga and her female relatives will once again ritually wash her before she is decorated with beads around her head and waist. Coming out from her mother’s house she is a bride, nawiinga, and will meet with her brdegroom who has also been washed in wedding water and who symbolises his status of a bridegroom, siwiinga, holding a wedding stick, entoa and a wedding spear, ilawa. The water and clay symbolises the beginning of a new life and washing in water is used in order to purify and cool down the young couple.

and accused of having sexual intercourse that had effected the outcome or caused a failure (Barndon 1992). The secondary refining furnace, ichinteengwe was built on a yearly basis, commonly located beside the large furnace or at the opposite side of the termite mound.

My informants told me the furnace was ready to make iron like a young woman would be ready to marry and have children, the furnace was a kanawiinga. This ceremony was carried out only when a new furnace was completed, never when the smelters reused a furnace. Hence, in most cases the furnace was seen as equivalent to a mature, birth-giving woman as in most recorded African smelting contexts.

The secondary refining furnace that we had used the previous year had been destroyed or pulled down. A new refining furnace was constructed out of greyish clay collected southwest of the village towards Tupa and raised at the same spot as the previous year. These furnaces were rebuilt every year but a smelting crew could also have several of these furnaces at the same site. In the re-enactments these furnaces were built within less than a day.

For the second experimental smelt the natural draft furnace that the smelters had built the previous year was re-used. It was plastered with wet clay to repair cracks and holes after the last year’s rainy season. The mud from the termite mound was paddled by the feet, thrown and packed, and smeared into cracks in the furnace wall. Repairs were made at both the inside and the outside of the furnace. New scaffolds were made. The building of these furnaces followed the same operational stages in each re-enactment and was built after the natural draft furnace.

6.6 Re-using the furnace in Katumba 6.7 Re-using and building furnaces at Tupa A new smelting site was rarely established in Ufipa, because the high shaft furnaces were solid and could be reused for several years. Mapunda has interestingly noted that by counting the thin layers of clay after the replastering of the outside and interior of a Fipa furnace wall one can also count the numbers of times a furnace was reused (Mapunda 1995). In 1991 I observed how the old furnace at Tupa outside Katumba village was indeed re-plastered with wet clay before our smelting started (see plate 6.7). However, the same year I also witnessed how a furnace that had been in use only the last year, for our 1990 experiment, was not re-plastered all over, but only where severe cracks were seen (Barndon 1992). David has questioned if one can actually count the times a furnace has been plastered because of the problem of the differential heating during the furnace fires the plaster to different degrees and subsequently causes cracking along the boundaries. This may look like different plaster coats but in fact it is not always that (David personal communication 2000).

The first stages in the Fipa iron making process were, when new furnaces were not built, the preparations of magical substances and sacrifices, clearing the smelting site and repairing the furnaces. For this preparatory tasks the master smelter did not need the large workforce that he would for making the soil, ifo. Hence, for our second and third re-enactments only blacksmiths and former smelters participated in our experiments. According to my informants several smelts were conducted in the same furnace during each smelting season and the same furnace could be used for several years. Most years the preparations for a smelt involved clearing the site of vegetation along with minor repairs on already standing furnaces. How this workload itself might have been organised was observed during the preparation for the third smelt in 1991. The third re-enactment observed in Ufipa was conducted at an old smelting site, called Tupa, in an old furnace. Tupa, that literally means ‘where the rhinos lived’, is a swampy area surrounded by marshes and smaller hills with termite mounds. Tupa is located two kilometres south of the smelters present village and the furnace was located on the valley slope east of the Tupa iron ore source.

My informants told me new furnaces were built close to the old ones and as Mapunda has observed even at the very same spot (Mapunda 1995:288-239). Nevertheless, entirely new sites were not created on a yearly basis and were rather a rare event. At least one can say that in the vicinity of reparable furnaces new sites with a new crew were not created unless for some reasons the furnace would not produce forgeable iron, if this happened one of the smelters were called twachiluungu or wachililuungu

The site was cleared from scrubs and trees. Severe cracks in the furnace superstructure were plastered with wet clay. The soil used for the plastering was collected from the termite mound behind the furnace. The inside of the furnace was only partly plastered. Inside the furnace the

86

The Fipa practice of iron smelting base was cleaned from slag and broken tuyeres from previous smelts. Repairing the old furnace and making new scaffoldings was completed in one day. Hence, although the Fipa workforce was a lot bigger than the four to five smelters needed in Pangwa smelting, the Fipa group of asiluungu managed smelting most years without a large mobilised work force, no slaughtering of a cow or large beer-parties. Women would transport charcoal and tuyeres.

tuyeres. My informants had never heard of any rituals or sacrifices for the tuyeres or tuyere makers. When we were confronted with the struggles with making the tuyeres they told me that that the old advisors, aluumba, used to be tuyere specialists (Barndon 1992). This has also been recorded by Wembah-Rashid who adds to the story that the masters had first claimed that anybody among the smelters could or at least were allowed to make the tuyeres (Wembah-Rashid 1973). Wyckaert and Cline, on the other hand, report that tuyere making, a slightly shameful task, was the work of young men (Cline 1937:41; Wyckaert 1914:372). For the second and third re-enactment of smelting the tuyeres were made out of the same clay source that they had used the previous year, close to Mkumbi village.

Before the smelting at Tupa started the young master smelter hired two medicine men, asinaanga, from the village who secretly added more magical and homeopathic substances to the furnace chamber. Additional ingailo was mixed with magical water, amaasi ileembo, and they used a cow tail to sprinkle the medicines into each tuyere inlet of the old furnace. Once the smelting started the master placed a woman’s carrying circlet, ikata, on top of the charcoal before the fire was lit (Barndon 1992:108). The secondary refining furnace, ichinteengwe, was constructed sloping downwards at the northern side of the termite mound. It was built of clay from the termite hill at the site.

Most ethnographic surveys or observations of iron smelting confirm that men made the clay tuyeres. Zimbabwean Ila-speaking women were recorded preparing the clay for the tuyeres, inchela, although they could not actually make them (Smith and Dale 1968: 205). In the past, Fipa women were forbidden to participate in tuyere making and the location of the manufacturing site was hidden from village life. The men felt embarrassed working with clay, in general a female task. Nevertheless, women were during our observation and most likely in the past, were allowed to carry the same tuyeres to the smelting site (Barndon 1992, 1999a).

6.8 Making tuyeres and bellows Childs’ third stage in the smelting process; providing a means to bring air into the furnace to stimulate combustion (often tuyeres and bellows) was in the Fipa process conducted immediately after the furnaces had been built. The reason behind these sequences in the process my informants explained to me was because the tuyeres were made of clay and like the furnaces, had to dry before the smelting could start. Making the tuyeres was one of the most time consuming tasks (Barndon 1992:59). The tuyeres were first made from the same soil as the one used for building the furnace but the smelters experienced that this was difficult and they decided to move to where they used to make tuyeres in the past. This location was closer to their old village, Mkumbi, one kilometre east of Katumba. The clay was of the ordinary fat greyish type and although the smelters still found the task difficult, using the correct material facilitated the process.

6.9 Collecting and preparing ores Iron ores, inyiimbo, reduced in the Fipa furnace were socalled ‘brown ores’ or a sedimentary limonite. The smelters of our crew distinguished high quality ores from the other ores calling them icitifiti, which literary means ‘black ores for knives’, incuunda, and ‘ores for axes’, impaasa. Before the smelters were allowed to start mining for ores the master told them once again that from now on they should all be faithful to the craft and ‘leave other things behind’ (Barndon 1992:53). This referred to the taboo against sexual intercourse. Ores for our re-enactment were collected at different localities, ores at Tupa close to the old furnace repaired for our third re-enactment and at different sources closer to some of the smelters’ old village. In the past our crew had used a source at Namatanda close to Nambogo village. In the last years before smelting was abandoned (in the late 50’s) ores were collected from Tupa (ibid.). Although this was a heavy task that took the smelters two entire working days it was a task performed only by the professional smelters, asiluungu (cf. Appendix E). When the ores had been extracted they were left in large piles to dry before the young smelters carried the baskets of ores to the smelting site. The large pieces were sorted and chopped into smaller walnut-sized pieces with pebbles, amafiinga, and left to dry in the sun until the charging of the furnace on the night of smelting. The ready prepared ore was placed

In our observations the making of tuyeres was one of the most difficult tasks. Long, thin wooden poles, misuungu, were inserted into lumps of wet clay. The hole in the lump of clay was filled with ashes in order to prevent the clay from sticking to the wooden poles. The clay was worked with water to soften its outside and formed into long clay pipes that were cut into 50 cm long pipes of which more than 100 were made for each smelt and up to 150 for each smelting season, knowing that many would break (ibid.: 60).113 The tuyeres were brought to the furnace when they were dry and the smelting was about to start. In the past only specialists called kasunga wachelu, made the 113 They made one hundred, 50 cm long tuyeres for our first reenactment in 1990.

87

Randi Barndon on the northern side of the furnace. During surveys of old production sites, as well as before our re-use of the Tupa furnace ores were placed north of the furnace indicating a pattern in organisation of space on production sites.

reports that new axes were forged for the cutting of wood. The axes were left in the chief’s or rather the master smelter’s house (or hut at the smelting camp) the night before they went out in the bush. The head smelter pronounced a curse on anyone who wanted to curse them and invoked the protection of the ancestral spirits. The smelters were told the importance of sexual abstinence on the night before leaving for the bush. The next morning the smelters smeared white earth on their foreheads (Cline 1937:122; Wyckaert 1914).

For the second re-enactment iron ore was collected at Namatanda, a swampy almost boggy valley slope close to Nambogo village, which is located north of Katumba. The ironworkers chose this iron ore source claiming the iron content was higher than in the Tupa source and that it was easier to extract than the ores at Tupa. When we were clearing the old smelting camp at Tupa the smelters found a pile of iron ore left at the site by the last smelters that had used the furnace several decades ago. Our master smelter decided to use this ore in addition to ore collected at Tupa. They claimed that the ore found at the site was from the same Tupa source but of a higher iron content than the ore they themselves had recently collected. Ores of high iron content had all been extracted from the area during the past active days of smelting.

For our second and third re-enactments firewood, kateembwa, was collected from a smaller forest in the vicinity of Mtuia village. Selecting trees, making charcoal fuel from them and transporting the fuel to the smelting site was in our re-eanctments the work of the young apprentices and a special labour force called in to assist with transportation. As we were granted a permit to cut down trees for charcoal and firewood, after surveying the forests around Katumba Azimio, Kisuumba forest, south of our village was chosen, and the charcoal was transported back to the smelting site.

6.10 Making charcoal and collecting firewood In Ufipa I observed the same fear of ‘hot’ or ‘impure’ people coming in touch with the charcoal that I discussed in Chapter 5. Like Schmidt describes in his Haya study (Schmidt 1997:70), because of this fear the Fipa smelters stored the charcoal in the forge in the village and had one person guarding it the nights before the smelt started. In the past specific species of hardwood were used for charcoal production. The master told me that specifically when forging a hard wood was needed. In my previous study of Fipa charcoal making I stated that no rituals, sacrificed or magical performances were carried out prior to the charcoal making (Barndon 1992:51-52), but the smelters were reminded about the importance of sexual abstinence. They had to use new pure axes when they were cutting down the trees and logs of wood. A new fire was made with fire drill sticks for the fire on the charcoal mound.114

6.11 Smelting begins In the three observed smelts the steps described above in building or repairing the furnaces, making tuyeres and collecting and preparing ores, charcoal and firewood followed the same procedures. Once everything was prepared the master smelter entered the furnace through the main tuyere port, palinyina and the tuyeres were given to him through the furnace tuyere ports. From inside the furnace the tuyeres were positioned and placed in the doors, plastered with clay. Two tuyeres were laid down inside the furnace chamber in front of each inlet in order to direct the tuyeres in an upright position towards the furnace centre.115 When the tuyeres were in position the floor inside the furnace was sprinkled with sieved soil before the charging of the furnace started. The master remained inside the chamber while other men stood on the scaffolding and handed down baskets of charcoal, firewood and iron ore. First charcoal, then ore and subsequently fire wood (Barndon 1992:64). The furnace chamber was filled with these alternating layers all the way up to the top (cf. also Wyckaert 1914:376). Two smelters made fire with traditional fire drills and handed the fire over to the master smelter who put the fire on top of the charcoal at the top of the furnace (cf. Lechaptois 1913:240). Fire was lit on both the eastern and western side of the furnace. Before the first smelt the old master smelter and the one becoming the new owner of the intaangala basket entered the furnace chamber. In the second and third re-enactment only the new umwaami entered the furnace ‘womb’. Before the master (and his assistant) climbed down from the furnace they moved in circles around the upper creepers of the scaffoldings. At

The most detailed observation of charcoal making is described in Wyckaert (Wyckaert 1914, see also Cline 1937). Wyckaert notes that the women burnt the charcoal but if something went wrong one of the smelters would be accused of violating the taboo on the night when the axes were blessed. If he confessed he would be excluded from the group and would have to pay a fine of twenty goats or a slave (!). If on the contrary, no one confessed poison ordeals were administered to the suspects or a diviner would declare that an evil ghost had caused the trouble. If the charcoal preparation was a success the charcoal was brought to the camp by women (ibid.). Wyckaert also 114 For the first re-enactment observed in 1990 the smelters did not prepare charcoal because of severe deforestation problems on the Ufipa plateau. Charcoal was bought in Malonje and transported to the smelting site. For the two smelts conducted in 1991 we were granted a permission by the Rukwa Regional Cultural Officer to collect firewood and make charcoal within the protected Kisuumba forest.

115 Eighteen tuyeres were placed in the main tuyere inlet while five to six were inserted into the other ports (Barndon 1992:64, see also Appendix D).

88

The Fipa practice of iron smelting

Plate 6.8 A secondary refining of the iron in the low shaft furnace, ichinteengwe, Tupa 1991 tuyeres in the western door, palinyina, were removed. The slag and ore/iron/charcoal admixture forms a cake or furnace bottom which the Fipa call ‘the iron stool’, incaango. This product needs a secondary refining. Approximately 46 hours after the reduction started pieces of iron, untare, that was considered suitable for a second refining were sorted and taken over to the smaller furnace, ichinteengwe.116 The sorting of iron for the secondary refining lasted until the early afternoon.

sunset the fire was lighted and now the smelters drank beer and ate before settling down around the furnace. At intervals during the night the master looked through the peephole, ntaanda, in order to control the colour of the fire inside the furnace chamber. I was told that the person who had put the furnace to fire was the one allowed to collect the iron from the furnace chamber after the smelting. For the process to work; to get the correct air draft in the furnace chamber and the exact temperatures needed to reduce the ore and finally obtain the desired product experience is important. As pointed out by Killick, a typical phenomenon connected to natural draft furnaces is the fact that once the furnace has been charged with the fuel and is left to burn down and reduce the ore by itself, there is not much the master smelter or his crew can do to control the outcome (Killick 1990). Therefore, explanation models were connected to witchcraft and other super-natural beliefs. The smelters have experienced that they must please the ancestors and gods. Because certain practices cannot be other than perfectly adjusted to the operation of the furnace, the spatial arrangements of the furnace, its chamber and combustion through the tuyeres (ibid.).

The crew had made a platform for the secondary furnace and three aged men arranged for seats close to the furnace. Stones were placed over the iron tubes to keep them firmly in place in front of the tuyeres during the bellow working. The furnace chamber was filled with charcoal and lighted from the top. Three men worked the bellows to achieve the correct temperatures inside the furnace chamber. Small pieces of iron collected from the natural draft furnace were placed on top of the burning charcoal. The bellows used by the Fipa smelters were bag-bellows that were made of whole inverted goats-skins, preferably from black he-goats. In Fipa thought, black both connotes fierceness and innocence as the ‘blind’, icitfiti, black children conducting the sacrifice for the furnace foundation. Goatskin bellows were made only as a demonstration, as they had bellows from the forge that actually could have been used. In the past the smelters used bamboo tubes that led the combustion from the bellows through the tuyeres and into the furnace. Three tuyeres

During my fieldwork in Ufipa I realised that iron smelting is certainly not a simple technological activity, several crucial aspects must be considered each time a smelt is to be conducted. In the Fipa’s furnace, iluungu, there is no slag tapping apart from whatever pours out of the tuyeres. After approximately twenty-four hours of smelting and additionally twenty hours of cooling of the furnace the

116 Iron from the natural draft furnace was called mwiluungu or untare inntutu which means ’iron which is still unclean’ (Barndon 1992:66).

89

Randi Barndon operated on the furnace chamber. These tuyeres were of the same size as those used in the high shaft furnace.117 We were demonstrated how the skins were stripped from the animal in one piece. Compared to other technologies, such as the Pangwa bellow system, this is a simpler way of arranging for combustion in the furnace. The bellows were worked to achieve correct temperatures before the iron, untare, was added. Three men worked the bellows while one of the elder assistants was seated close to the furnace on the northern side and he added small pieces of iron. During the entire process the two elders continuously added more charcoal and iron. The younger men worked the bellows at intervals, taking turns whenever one of them became tired. In the breaks they drank beer. Young boys were allowed to blow the bellows.

had been produced out of the wrong composition of trees and the pieces had been too small. This had prevented the correct air draught inside the furnace chamber. The hardwood was supposed to be as green and wet as possible in order to slow down the charcoal burning. The smelt was only partly successful but the ‘bloom-like’ product, incaango, was taken out through the main door at the furnace base. Slags and metals were broken into smaller pieces and went through a refining in the secondary furnace in which the smelters produced forgeable iron. For the second smelt observed in 1991, the smelters themselves selected trees for firewood and charcoal production. The furnace was charged in the same manner as for the previous smelt and the doors were opened after 46 hours of operating and cooling down of the product. Again the men claimed the smelt had been only partly successful. The smelting time was correct, they had the right kind of charcoal, but claimed the ore was of low quality and the tuyeres were not functioning properly. They had not been completely dry when inserted into the furnace chamber. Showing tuyeres that had been twisted or broken under the smelting and therefore prevented the air from passing through and the slag form pouring out demonstrated this. The product was still good enough for a refining in the secondary furnace, which once again produced forgeable iron, ululu.

Although forging is hard work and they use the same bellows, smelting is harder because the bellows must be pressed harder. Nevertheless, boys were allowed to blow the bellows and in the past this would often be an introduction to the craft. The smelters said this process was like cooking and the furnace was also called inteendo, which literally means a cooking pot or cleanser (Barndon 1992:110). They told me that inside the furnace, like a pot, the substances were hot, amapiso, like female and male forces. The master smelter, or at intervals his assistant, sat in front of the rake hole. He controlled that no air was let through by closing the opening with soil or by throwing in pieces of charcoal. When the slag started to pour out after 15 minutes of intensive bellow-working it was raked away with an iron stick, inchiinde. The men continued the bellow working and added more charcoal and iron continuously. Once they heard a sparkling fire they knew they had produced a bloom. The amount of iron wanted and space inside the furnace chamber limited the time spent by the furnace. Hence, once a lump of iron had been produced the bellow work ended. The iron was taken out of the furnace chamber by an iron stick and left beside the furnace to cool down. The bloom, ululu, was taken out of the furnace through the top and I was told that the ichinteengwe had ‘delivered iron and was not in need of revealing secrets’ because the smelt had been a success. Another refining of incaango, ‘the iron seat’, started soon afterwards. In the past this depended upon how much iron they had extracted from the primary furnace.

The third smelt observed at Tupa 1991 was according to the ironworkers themselves the most successful. The actual smelting took 26 hours, while the furnace was left to cool down for another day. The speed of the smelting was following the right tracks through the furnace chamber and the tuyeres were functioning efficiently. Slag pored out through the tuyeres after the furnace had been in operation for 11 hours. When the smelters were refining the iron, incaango, in the secondary furnace, the bellows were according to my informants’ statements easier to work than they had been during the previous smelts. The iron bloom was of a better quality, like the blooms they in the past had used for making axes. Although the smelters claimed all the re-enactments had been successful in producing iron they were not totally pleased. They had hoped for more iron, incaango and untare from the natural draft furnace. After the first smelt was observed we forged a female hoe in order for me to observe the forging process (cf. Barndon 1992 and see Chapter 4, plate 4.1). After the second re-enactment in 1991 we did not forge any of the iron produced. The smelters claimed the tuyeres had not transported the slag properly out of the furnace and therefore this iron would even after a secondary refining be difficult to forge. According to my informants the last smelt observed was the most successful and the active blacksmiths from our village planned to forge this iron in their village forge.

For the first smelt observed in Katumba in 1990 the reduction process took less than 16 hours. Two critical factors had not worked out the way the smelters expected. The amount of time the furnace was operating was too short and further the tuyeres had not functioned properly at the base. I was told this had resulted in a low air input and no slag poring out through the tuyeres. These factors were explained by the ironworkers as due to the poor quality of the charcoal used as well as to the fact that the hardwood we had used had been too dry. The charcoal

In the past one smelting season would yield the minimum 117

Cf. Appendix D for dimensional features and tuyere arrangements.

90

The Fipa practice of iron smelting really not forgeable or if it was not good enough for hoes, axes or bracelets. Maybe the smelts were failures in the aged master smelter’s opinion more than from a metallurgical point of view. Once the smelting was completed, although several smelts were conducted in the same furnaces, the crew would return to the village. Wyckaert writes that each year and for each of the asiluungu a smelt was conducted in order to them all to have a bloom to take to the forge (Wyckaert 1914). Based on informants’ testimonies and observation in the field one smelting in the natural draft furnace would yield from four to seven kilograms of iron. This fits well with the number of the crew of smelters.

of 7 blooms or enough for 12 large male hoes. It was the quality of the bloom produced in the secondary refining furnace that told the master if the smelt was a success. According to the aged smelters the iron from the two first smelts was classified as untare inntutu, which means it was unclean, not only according to the generic term for iron from the natural draft furnace but had only been half purified, intutu. After the refining in the secondary furnace this iron would still be classified as pure iron, ululu, but I was told that it would be hard to forge and would perhaps crack during the forging. Iron that gives strong axes, untare impaasa, will give the best results in the forge and was the product we achieved in the third reenactment at Tupa.

6.12 The chaînes opératoires in the smelting process I was told that if the bloom was not forgeable all the members of the crew or their wives had not maintained the taboo against human sexual intercourse. Because of this ichinteengwe revealed secrets of adultery. If the end product was of poor quality this was seen as small holes, iitoli, in the bloom. If one of the smiths had not maintained the taboo against sexual intercourse he had to admit his adulterous behaviour. The same accusations of adultery faced a woman in labour if she had difficulties during delivery. Her two midwives would ask her to tell them if she has committed adultery and give them the name of the child’s father. This illustrates that adultery is morally not accepted in real life, as it was not in iron smelting. Being in contact with a person who has committed adultery might lead to one of the most dangerous sicknesses, inchila, which is transmitted by heat and fire (Willis 1981). The sickness caused by contact with adulterous people is among the Fipa considered one of the most dangerous diseases and its invisible pains might lead to death. If one of the smelters admitted adultery he was called ‘one of those who commit adultery’, wachililuungu. Men told me it is generally believed that women are especially adulterous and wachililuungu literary means ‘feminine smelters’. A smelter who confessed adultery had to offer a white hen (not a cockerel) to the master and his family. If this was accepted he would be excused and they would re-smelt the iron, if not, he was expelled from the group. If all the smelters could prove their innocence by ordeals or alibis they were told by the master to look among their wives for the one who had destroyed their iron. If the iron was classified as impure, one of the smelters wives would be accused of adultery as they were thought to be especially adulterous and the case would be a matter of shame, insoni, for husband and wife to solve in privacy. But even this rarely happened and I was told women too knew about the mandated taboo against human sexual intercourse during the smelting season. According to our discussions with the smelters this rarely happened since the furnace, ichinteengwe, functioned as a cleaner, inteendo, a cooking pot for the iron.

The natural draft furnace of the Fipa was a human and they called it a kanawiinga, a woman ready to get married. When a new furnace had been constructed and new scaffoldings were raised around it the smelters decorated the furnace shaft with flour and placed flowers on top of it. These decorations were like decorations, yaeemba, used in marriage ceremonies and the furnace became a nawiinga, a bride. I was told that the master the night before the work on the furnace started masturbated and was not allowed to wash his hands (Barndon 1992). Both Wembah-Rashid and Wyckaert note similar testimonies made by their informants (Wembah-Rashid 1963; Wyckaert 1914), as did Strinimann among the Pangwa (Stirnimann 1969). The following morning a girl and boy placed the medicines assisted by the magical substances, powers of femaleness and maleness, in the foundation trench or within the furnace ‘womb’. The children, the performers of the ritual, themselves represented the powers in gender combinations. Like the children, the asiluungu, the smelters, entered the furnace through the western tuyere inlet that they called ‘the mother opening’, palinyina. Once the smelting started the crew slept around the fire from the furnace. The smaller Fipa refining furnace was modelled as a woman in delivery but had several names. When trying to explain to me what happened in this furnace they said that it was similar to a ‘tunnel’ or it was ‘the one who reveals secrets’. The secrets the furnace revealed were about the quality of the iron bloom. However, the quality of the iron produced depended upon the behaviour of the smelters and their relatives, in particular their wives. If nobody had committed adultery the furnace produced forgeable iron, but if someone had sexual intercourse the furnace would reveal that the iron was not forgeable. The furnace was also conceptualised as ‘the one who cleans’ because it was called a cooking pot, inteendo.118 Hence, the furnaces 118

In an ethnoarchaeological study in Cameroon David et al. (1988) have demonstrated how crucial the metaphorical link between pots and persons may be (cf. also Jacobsen-Widding and van Beek 1990:24 and Richards 1956:31 for similar results).

Recalling the results from our Pangwa re-enactments it is not completely clear if the products we produced were

91

Randi Barndon were perceived as a marriageable woman, one that can conceive and as a woman in delivery because they; as real

women were perceived in terms of embodiments and containment, a theme I shall explore in the following chapter.

Figure 6.1 Chaînes opératoires in Fipa iron smelting

92

Iron technology and symbolic reservoirs

7. Iron technology and symbolic reservoirs The mind is a metaphor of the world of objects, which is itself but an endless circle of mutually reflecting metaphors (Bourdieu 1997:91). 1973; Solheim 1998:67; Turner 1967, 1969; Weiss 1996; Willis 1967).3 Less attention has been paid to how technological activities and technical equipment are conceptualised according to the same embodied thinking.

7.1 Smelting symbolism In the following I will discuss how iron working was connected to metaphorical construction as deeply embodied experiences and how these informed about matters of social significance during the smelting to the people who practised it. I will discuss smelting symbolism related to gender and sexuality. I will address notions of the body as a closed system (Douglas 1966:115), the body as a container (Johnson 1987:21) and notions of the body as open (Solheim 1998:69). Finally, the smelters’ use of magic and varied metaphorical imagination connected to their perception of the smelting furnaces/bodies will be discussed.

Taking a semiotic approach Collett has argued for a conceptual link between smelting, cooking and procreation (Collett 1985:179).4 The foetus and iron are linked because both of them involve a transformation of one substance into another and this transformation is mediated by heat. In most Bantu-speaking communities we find a notion of heat being the agent in procreation or the idea that sexual intercourse makes a couple ‘hot’ (de Heusch 1980; Marwick 1965:66; Richards 1982 [1956]; Weiss 1998).5 Based on Lévi-Strauss’ discussion of how heat-mediated transformations of the raw into the cooked are used to conceptualise other cultural domains Collett includes pots and the irreversible transformations that occur inside pots with furnaces (Collett 1985:121). Collett argues that the heat-mediated transformation is basic to the symbolic triad of cooking, procreation and smelting and that these transformations produce irreversible results (Collett 1993:505). When meat or vegetables are heated the change that occurs is irreversible and the food cannot be converted back to a raw state. Iron cannot be converted back to ore and the foetus cannot be converted back into semen and blood. The metaphorical link between iron smelting, procreation and cooking may be represented by making the furnace look like a woman (i.e. the Pangwa and Fipa furnaces were decorated as a woman) or conceptualised as a pot (the Fipa’s inteendo). Collett concludes that this mode of thought enable humans to control the boundary between culture and nature that is commonly found in myths in order to differentiate the ‘cultured’ people from the ‘wild’ (Collett 1993:505).

Semantic oriented gender models of the smelting process have been given much attention. The focus has centred on the furnace as a woman and the iron as a foetus while the people engaged in the technology, the smelters, and in particular the ambivalent and ambiguous role of the master smelters as magicians, have been missing. Missing these aspects has resulted in a reifying of the furnace into a symbol and as representing something social or cultural. Attempts to explain human experience in purely physiological or purely psychological terms have proved to be inadequate, as have those approaches which (like Descartes’) merely sought to mix the two kinds of explanations, while leaving both fundamentally intact (Langer 1989:30; Latour 1996:103). Phenomenology on the other hand situates human existence between the two. Phenomenological insight establishes a middle ground between the ‘physiological and the psychological’ (Langer 1989:30)1 In my argumentation for a holistic and human frame on technology I have tried to emphasise that technological activity is embodied and that agency and bodily movements and conceptualisations in a technological practice are culturally defined. For Merleau-Ponty the body not only mediates all experience, it is ‘our general medium for having a world’ (MerleauPonty 1989:146). Hence, with phenomenology one may say that ‘the body is the ground or anchor by means of which we locate ourselves in the world, perceive and apprehend it’ (Tilley 1999:34).2

Collett’s pioneering use of Lévi-Strauss’ The raw and the cooked (1964) in order to demonstrate the metaphorical triad pots, furnaces and people has been recognised in numerous case studies (Barndon 1992, 1996b; Childs 1991a; 1991b; Childs and Dewey 1996; Childs and Killick 1993; Herbert 1993; Mapunda 1995; Maret 1980, 1985; Schmidt 1996a, 1996b, 1997; Schmidt and 3

While ‘we’ have tended to separate our-selves from our own embodiment, the language employed in academic discourse and in describing objects in terms of the ‘mouth’, the ‘shoulders’ etc. is a familiar terminology for archaeologists and specifically used within ceramic studies (MacLean 1996:249). Why do we describe containers in terms of body parts? Why did the Pangwa and Fipa smelters do the same? 4 Collett also argues for a link between decorated furnaces, furnace bricks and decorations on Early Iron Age ceramics (Collett 1993:505). 5 Junod reported how the Thonga of Mozambique say sexual intercourse releases ’heat’ and the child is the product of a successful fireing (Junod 1913 in Barley 1994:106; de Heusch 1980:30).

Anthropology has demonstrated how peoples in nonwestern cultures live by body symbolism and in symbolic universes based on concepts of the body wherein separation, differentiation and classifications of body related symbols have been fundamental (Douglas 1966, 1

But cf. Latour who questions this distinction or ‘middle ground’ (Latour 1996:78). 2 Italics by the author.

93

Randi Barndon Mapunda 1997).6 While the idea of fire or heat as a powerful transformer is essential I will ague that this was just one of several levels in the symbolic repertoire used by African iron smelters.126 Making the pots themselves entails concepts of the heat-mediated transformation (Gosselain 1999).7

categories which are themselves fluid through their use in social discourse. As Sanders puts it ‘by examining gender, which may be defined as the cultural construction of male and female as well as the relationship between them, and the generative powers inherent in and unleashed by gender categories, we can better understand the nature of numerous ‘sexless’ but nonetheless sexual activities’ (Sanders 1998:239). All social behaviour, including technology, is grounded in a conceptual framework that imposes order on the world and lends structure to human existence (Childs and Killick 1993:325). Hence, African iron working was explained by analogy to human physiology. Of main importance in the iron smelting context was one of the fundamentals of all societies; the female/male division.8 Gender exclusions

7.2 Gender Beauvoir’s statement that ‘women are made, not born’ launched a whole generation of feminist scholars, intent on dispelling the doctrine of ‘natural’ difference and showing that differences between the sexes were socially rather that biologically constructed (Beauvoir 1952 in Davis 1997:8). Ortner’s (1974) and Rosaldo’s (1974) influential discussion of gender, nature and culture in terms of an opposition and devaluation of women into the natural or private domestic sphere as opposed to men in a cultural and public sphere has recently been challenged. This way of thinking about gender and gender/sex has been illustrated to through ethnographic cases to be itself a Western construction (Gilchrist 1999:32; Moore 1994:10). Evaluations of for instance the house as more private and exclusively associated with women than is the case in non-western societies has challenged the strict dichotomies once used in order to understand the discriminating role of women. Still, gender has gradually become accepted as among the major structuring social principles (Moore 1988; Strathern 1988). Gender as a category crosscuts and informs distinct yet interrelate cultural domains (Sanders 1998:240).

A gender divide and female exclusion existed in almost all African iron working societies (Herbert 1993: 25). Documentation of smelting practices in Sub-Saharan Africa report on exclusively male smelters and smiths (Barndon 1999a; Celis 1989; Cline 1937; David et. al. 1989; Fowler 1990; Haaland 1980; Herbert 1993; Kense 1983; Merwe and Avery 1987; Schmidt 1996a). Part of the data collection in this study has been to arrange informal gatherings of iron smelters and present day smiths to discuss the iron working process before the actual experimental smelts. In Chapter 4 I discussed problems related to describing a non-verbal practice and how many of my informants only answered my questions with the answer ‘wait and you will see’. I was, however, able to collect data on the procedures, rules and taboos of the process. Looking back at the collected material from these early gatherings the discrepancies between prescribed and actual behaviour in terms of female exclusions are remarkable. Both Fipa and Pangwa ironworkers claimed a total absence of women during all smelting activities. Of particular significance was the denial of female participation in ore collection, fuel preparations and tuyere making as well as the actual smelting. Women were feared because it was believed that they could affect the quality of the bloom or even prevent the furnace from producing forgeable iron.

Moore argues that ‘this binary biological sex division does not provide a universal basis for the cultural categories ‘male’ and ‘female’ (Moore 1994:13). Numerous case studies from New Guinea (Meigs 1990 in Moore 1994:23.24; Strathern 1988; Strathern 1996; Strathern and Lambek 1998) but also emerging cases from African studies (i.e. Davis 2000; Weiss 1996, 1998) are informative on how people through their life histories possess substances or social identities associated with both men and women. For instance, the Mount Hagen classify men as more feminine or more masculine according to both age and agency and moreover, in terms of temperatures in the body fluids as hot or cold (Strathern 1988, Strathern 1996). People are commonly classified according to how much maleness or femaleness they possess such as after acquiring more substances from the opposite sex in everyday life (Meigs 1990 in Moore 1994:13). Therefore, the lesson to draw from this is that the biological sex differences cannot be equal to the socially defined gender

Different chaînes opératoires formed different sequences in terms of preliminary rituals and taboos. Strict sexual abstinence was mandated for the Pangwa when charcoal preparations started several months before the smelting season while for the Fipa the same taboo was mandated once the smelting crew left for the bush. Until the blooms were brought back to the village Pangwa and Fipa smelters were not allowed sexual intercourse. The smelters were in one sense the ‘husbands’ of their gendered ‘female’ furnace. It was commonly explained that breaking the taboos could cause a failure to produce forgeable iron. Moreover, both Fipa and Pangwa believe

6

Although it is correct that most furnaces are gendered and considered a female, the Sukur furnaces in Northern Mandara, Cameroon are persons while their gender and other characteristics are ambiguous (David 1996:594). 7 This can again be exemplified by the Thonga who say the child is the product of a successful fireing since he is considered to be a clay utensil that has been fired and has not cracked (Junod 1919:138 in de Heusch 1980:30, see also Barley 1994).

8

The very male/female dichotomy assumes a binary gender structure in all societies precluding the possibility of multiple genders or fluidity in gender metaphors (Gilchrist 1999:32).

94

Iron technology and symbolic reservoirs a smelt (Herbert 1993:79; Reid and MacLean 1995:149). In Ufipa the wives of the elder smelters participated in the celebration of a newly built furnace, while in Upangwa the wife of one of the aged men was allowed entry during actual smelting as Schmidt reports from his Barongo re-enactments (cf. Schmidt 1996b:118, 120).

that adulterous behaviour may cause difficulties during delivery or miscarriages in real life (Barndon 1992, 1996b; Stirnimann 1969, 1976:402). But in both my case studies women were observed preparing the food needed during the smelting season because the everyday sex and labour division in cooking was stronger than the taboo against women in iron smelting.

Most ethnographic surveys or observations of iron smelting confirm that men made the clay tuyeres. The smelters’ wives within the region under study were commonly very skilled potters but were not allowed to assist or observe the men making tuyeres. Among others, like the Zimbabwean Ila, women prepared the clay for the tuyeres, inchela, although they could not actually make them (Smith and Dale 1968:205). Among the Pangwa and Fipa a taboo against women near iron ores and where or when tuyeres were made was discussed in Chapters 5 and 6. But women were the porters of the heavy tuyeres or ores to the smelting sites and in the Fipa case they prepared the termite soil for the furnace construction together with their husbands (Barndon 1999a).

Among most ironworkers a solution to this dilemma of not being able to cook because only women cook, and that one must not let women near the furnaces, was to introduce the master’s wife. She was old and since she had passed menopause, was considered less dangerous because she was not ‘hot’ (as if pregnant or as if in menstruation). Among Ila smelters a man was assigned by the master smelter to cook for the crew, but he was called by the master ‘my wife’ (mwinangu) (Smith and Dale 1968:207). This illustrates another rewriting of the taboo or rules of female absence. In Ufipa two aged men acted as ‘midwives’ during the smelting process. They were skilled and knowledgeable aged men called aluumba. As wise men they were metaphorically linked to the skills and knowledge of elderly women, amalombwa, the midwives.9 The presence of real women could harm the furnace and its possibilities to produce iron, yet women’s knowledge was important.

When considering technological knowledge (metallurgical, magical or ritual) women were always totally excluded. Still, the Fipa master smelter’s wife took care of the intaangala basket in the house in-between the smelting seasons. The Pangwa head smelter’s wife actually made the ing’anjo pot that was buried beneath the furnace. In our Pangwa re-enactments the head smelter’s wife also slept close to the furnace together with the crew the night before the actual smelting. Early next morning she cooked the sacrificed goat for our crew. Later, she assisted the smelters when they fastened the tuyeres and closed the tuyere inlets to the furnace wall. She was also allowed to blow the drum bellows, an exclusive male task among most smelters (Barndon 1999a; Herbert 1993).10

The last smelt observed in Ufipa was carried out at a different location because the smelters feared inference from ‘hot’ people from the village. They moved from Katumba village to Tupa where they conducted several new preliminary magical rites on the furnace, calling in an ordinary medicine man to secure the purity of the place. The Pangwa also moved their smelting camp to a new location. They claimed ‘hot’ women had passed by the old site on their way to collect clay for pots. They placed a bigger furnace guardian, umboocha, at the outskirts of their new smelting site at Kuking’ande. The furnace had become embodied and the process of transformation about to take place was vulnerable to female powers.

During my fieldwork I observed an age division crosscutting the strict male-female distinction and female exclusion during the smelting season. But this was related to general gender differentiation and separations and specifically female exclusions in other contexts, even exclusive female contexts such as gardening, pounding grain, pottery making, cooking or birth giving. In his description of Pangwa crafts Stirnimann points to the fact that ritual purity and taboos against human sexual intercourse were not only mandated before iron working, they were equally important before rain making rituals, hunting, pottery making, woodcarving and basket making (Stirnimann 1976). We were assured of the same importance in these taboos when we observed pottery making among Pangwa women two decades after Stirnimann’s observations.

Both Pangwa and Fipa stressed the danger of women in menstruation and pregnant women near the smelting sites but the prescribed rules for female exclusion were in actual and practical behaviour less strict. Women in general, often young married women and wives of the smelters were allowed to participate at various preparatory stages in the process while postmenopausal women were allowed entry at the smelting site during the smelt (cf. David et. al. 1987:187-188; Schmidt 1996b:89). It is commonly stated that only postmenopausal women and girls before puberty were allowed to participate at the various preparatory stages of

In 1993 and 1995 the prohibitions against sexual intercourse and exclusion of women in menstruation or pregnant women during pottery making were explained to

9

The term refers to slag once the smelting products were evaluated (Barndon 1992:188). In oral cultures words sounding similar have power to produce similar effects (Maxwell 1993:25 in Davison 1993:35).

10 My informants assured me this was not dangerous and that aged women often participated in smelts in this way in the past.

95

Randi Barndon me by female potters when they were demonstrating their technology to us in Masimbwe village. We were told that people who were ‘hot’ from sexual activities or women menstruating would cause the pots to crack during firing and were therefore not allowed to participate. They were even not allowed to enter the area where the pots were made. Stirnimann’s informants told him that ‘hot’ people were not allowed to come in touch with the raw materials for conducting various crafts in which clay, wood or grass was involved (Stirnimann 1969:9, 1976:264). A similar taboo was in 1990 and 1991 mandated for Fipa potters. In Ufipa clay sources that were considered the best for pottery making were kept as a secret among a potter and her daughters. Fipa potters claimed that secrecy was in order to avoid ‘hot’ people from entering the compound where pots were made.11

rainmaking and everyday practices such as pounding grain, fire making, cooking and even eating. Sexual abstinence and other means of bodily control may be seen as expressions of social control in many contexts of production and use of material culture (e.g. Douglas 1970:74). As an expression of social control, gender combinations and exclusions also worked as a benefit and for the well-being of social life. In the late colonial and post-colonial era, the Pangwa and Fipa people lived by these gender distinctions and combination and created metaphors based on ‘being in the world’ cross cut by male and female powers. It was the combination of gender that was a source of power, not the female category isolated or the female body as a model for the furnace itself. Sexuality’s indivisibility from human purpose gives sexuality in the African smelting context a cosmic power (Heald 1995:497; Sanders 1998:239). Sanders points at an important shift in focus ‘that more than merely being about the act of sex or sexual abstinence we are dealing with cultural specific ideas about the combination of gender categories and the powers inherent in this combination’ (Sanders 1998:239, see also Strathern 1988). Rather, the combination of the two cultural categories ‘male’ and ‘female’ was and is important in view of the balances of substances in the human body of any gender and hence, in the physiology of the furnace.

Women in menstruation, pregnant or breastfeeding were excluded not only from iron smelting but many daily tasks when they were in these states of ‘imbalance’ being considered too hot. My point is that in Pangwa and Fipa iron working women and femaleness was both excluded and included. Gilchrist argues that gender identity in any society is maintained by ideology as habitus, common sense knowledge of how to proceed as a man or as a woman (Gilchrist 1994:14). These common sense ways of acting may be negotiated and adjusted to specific needs or tasks and even change over time due to this negotiation. Sexual celibacy or stripping away sensuality was not the point (cf. Foucault 1981 in Gilchrist 1994). As I will elaborate below, in the iron-smelting context, a context pregnant with sexuality, the point was rather to actually control the amount of body heating or bodily overheating that lay in the powers of gender oppositions and combinations.

It was combinations of ‘male’ and ‘female’ powers, or the gender complementarity (Gilchrist 1999:98), that was adapted in a structural model and acted out in the smelting process. This we see in how the contents of magical substances were based on combinations of gender powers, male and female fertility medicines and female and male powers and values.

Gender combinations Herbert has successfully demonstrated how ‘gender’ and ‘age’ categories have been used for guiding both symbolic repertoires and power structures in African iron smelting traditions (Herbert 1993). Her gendered model of transformation or ‘procreative paradigm’ gives meaning and permeates to the smelters’ notions of the human life cycle as it was expressed in the smelting rituals. Both Collett and Herbert’s analysis grasps the essential sacredness of sex and gender and the centrality accorded to reproduction in iron smelting (Collett 1985, 1993; Herbert 1993). The sacredness of sexuality and gender combinations was taken in and adapted in the model used for making iron. But gender symbolism, male and female, of fertility and procreation was and is not confined to iron working. In both Pangwa and Fipa society symbolic thinking based on gender and procreation prevails in many ritual and non-ritual contexts. Hence, not only was iron smelting and forging metaphorically linked to procreation or sexual intercourse, so were gardening, basket making, cooking,

The smelters distinguished ores into ‘male’ and ‘female’ categories (Roscoe 1923 in Childs 1991b:344; Rowlands and Warnier 1993:55).12 The description provided by Rowlands and Warnier echoes the technological treatments of the ores with that of the Pangwa. The Pangwa smelters used two ores; one was conceptualised as male because it was similar to semen while the other was female and equal to breast milk. Both breast milk and semen are claimed necessary for a foetus to mature. The two sorts of ore, one female and one male, were equivalent to Pangwa notions of the power of gender combinations and conception of how a foetus can only be nourished through a combination of the mothers milk and semen in the mothers milk. Hence, male and female forces and substances were combined. The Fipa and Pangwa named the various furnace parts such as the bellows, tuyeres, tuyere inlets and the furnace 12 Rowlands and Warnier report from the Grassfields in Cameroon that ores were classified as ‘male ore’, a ferralithic gravel while a ’female ore’ was not an iron ore at all and perhaps not even flux. The ‘female ore’ was clay collected from the banks of a local stream, made into a cake with slags and water and left to dry in the sun before it was smashed into smaller pieces (Rowlands and Warnier 1993:55-60).

11 In the early and mid 1990’s both Ufipa and Upangwa women made pots and stored and prepared food in their locally made pots while aluminium casseroles were preferred for cooking.

96

Iron technology and symbolic reservoirs create through our creativity and conceptualisation (Olsen 1997:117; Rorty 1979, 1991).

chamber in terms of body parts. The Pangwa front opening was called umlomo, which means ‘mother mouth’ (vagina), and the drum bellows were called amatongo, which literally means testicles (Stirnimann 1976:274). The Fipa called the main tuyere inlet, palinyina, ‘the mother opening’. The two elongated clay masses outside the rake hole of the secondary furnace were called impaamba which literary means ‘the inside of female thighs’, or loins. Willis accords embodiment and gender a central structuring principle in Fipa myths and identity (Willis 1967:159). The two symbols ‘head’ and ‘loins’ represent maleness and femaleness (Willis 1966:22, fn.82, and 1967:132). Female and male symbols such as the Fipa marking their foreheads with a red powder, unnkolo, normally used by women in marriage ceremonies or the Pangwa using and calling the wooden stick for softening termite soil a pestle, umtwangilo, exemplifies how gender categories worked as metaphors in order to conceptualise the process.13 Gender, or rather as Sanders formulates it ‘the cultural categories ‘male’ and ‘female’ provide particular salient categories for thinking about the world, and their combination is a powerful means of transforming it. Gender in this sense is good both to think and act’ (Sanders 1998:240). This as a principle formed a system of symbolic or metaphorical thinking in iron smelting.

Embodiment is difficult to define, but it breaks with the mind-body divide as seen in the social sciences from Durkheim to Douglas. Embodiment has to do with the body, but it implies that it is something else, other than or added to the physical body itself, that is embodied and such a ‘thing’ often turns out to be an abstract social value such as honour or bravery (Strathern 1996:198). As I will suggest for the iron smelting contexts I have observed, morality is another embodied and disembodied value taken into use. Hence, technological agency is not only based on habitus, but also a constant evaluation of actions (cf. Chapter 3). Johnson has argued that using metaphors is one of the central projective operations by which we establish semantic connections (Johnson 1987:192). Some kinds of metaphors must be regarded as irreducible primary cognitive functions by which we create and extend structure in our experience and understanding (ibid.). Inspired by Merleau-Ponty’s phenomenological approach, Johnson brings embodied experience into the study of metaphors. Hence, Merleau-Ponty claims that we are never separated from our bodies and from forces and energies acting upon us to give rise to our understanding, as our ‘being-in-the-world’ (Merleau-Ponty 1989:205). Both Fernandez (1980, 1986) and Geertz (1973) have demonstrated in various settings and from various parts of the world how metaphor is an instrument of the expression of value, both positive and negative. Metaphors provide a way of making the private public, of externalising the self by providing terms for the expression of sentiments that may otherwise be inexpressible or unintelligible to others (after Cohen 1995:140-141). It is a question of self-consciousness and of making the ‘I’ into ‘We’ (ibid.). Therefore, metaphors are the elementary form of social life and individuals create and move in their social worlds by metaphors (ibid.: 140; Fernandez 1986, 1991; Johnson 1987, 1993; Lakoff and Johnson 1980; Tilley 1999:10). This follows in Bourdieu’s statement that ‘consciousness is turned inside out and that the mind is a metaphor of the world of objects’ (Bourdieu 1977:89). Bourdieu argues that the representation of mental categories is embodied in two ways: by material constructions people make (e.g. houses or furnaces) and by positioning their own human bodies in relation to these constructions (Strathern 1996:26). Hence, Bourdieu concludes that the ‘the body is socially informed’ which implies that the physical and moral are fused (Bourdieu 1977:89, and after Strathern 1996:28). This approach entails that the social component may inform on the personal or embodied component. In the smelting context personal or individual experiences were embodiments that were simultaneously socially shared. The furnaces were not material representations of mental categories they were these categories, because the material and moral were fused.

7.3 Iron smelting and embodiment Although, as argued by Davis the body is not universal, it may be an object to which we can point, it may be humans’ substantial being, and it may be one element of what is at stake in our living and mortality (Davis 2000:39). However, since the body is so culturally conditioned, as conceptions of gender are, the body translated must be the body transfigured (ibid.). Davis’ major thesis here is that there are no concepts of the body that are ultimately universal or totally wrong (ibid.). In Chapter 3 and 4 I discussed how we might easily be mistaken if we merely transfer modern Western based approaches towards technology upon non-western technological processes. We have to be aware of the same danger when we transfer perceptions of the body or concepts in metaphorical thinking. Lakoff and Johnson have discussed the use of metaphors from a Western perspective. The same sort of embodied processes are however at work in other cultures in other parts of the world, something their most recent results have opened for (Lakoff and Johnson 1999, and cf. Moore 1994). A position that includes lived experience and a bodily foundation opens for and moves beyond the either-or thinking that at least until recently has prevailed in archaeology (Bekaert 1998:6, cf. also Lakoff and Johnson 1999:205; Latour 1996, 1999). This is founded on the assumption that knowledge is something which we can both discover from the reality and physicality that surrounds us and something that we experience and 13 Many of my Pangwa informants said that working clay or any kind of work with earth is women’s work, thus they felt embarrassed while they were making the tuyeres (see also Barley 1994:52).

97

Randi Barndon Symbolic thinking in iron making was based on deeper rule metaphors and metaphorical imagination in which notions of the body itself were the central experience. It is more likely that at a deeper level we will find a way of interpreting the connections made explicit and public in the smelting process. Therefore, it is time to modify the idea that an actual, real life ‘wedding ceremony’ (Childs 1991b:343; Childs and Killick 1993:326) or a ‘procreative paradigm’ (Herbert 1993) is all we see symbolically played out in iron smelting rituals. Rather, the two spheres both based their metaphors on a different and deeper domain or level of embodied experience. How then were individual embodied experiences expressed, used and socially shared and why were they at all taken into use in the metallurgical context of technological behaviour in the first place?

experiences of being in something, or for locating something within another thing, we find recurring organisations of structures; the experiential basis for inout orientations is that of spatial boundedness’ (ibid.). The most experientially salient sense of boundedness seems to be that of three-dimensional containment, being limited or held within some three-dimensional enclosures, such as a womb, a crib, or a room’ (ibid.). 7.5 The open body Johnson has demonstrated how the body is experienced and conceptualised in metaphors as a container. As a container the body is not closed but open for intake of food and water etc. and extracts various substances such as blood or semen. Hence, the body can also stand for an open or unlimited system (Solheim 1998:63). Johnson argues that because of containment we experience in and out orientations by our embodiment (Johnson 1987:31). Moreover, our experiences of ‘balance’ are connected to our experiences of containment and the body as open (ibid.: 74-75).

I will suggest that the triadic relation between procreation, cooking and smelting first acknowledged by Collett appears more complex if the use of the human body as a model and metaphor is more thoroughly discussed. If the metaphor the ‘body as a container’ is added to the triad these transformative processes take place within containers but are nevertheless related to the notion of heat or fire as a transformative agent.

The body is experienced as a container but has orifices and is open, something that makes it both vulnerable and powerful. Since we conceptualise the body as open the body itself may be transformed or is a transformer of substances. This is how I believe the Pangwa and Fipa conceptualised their bodies too, which therefore adds another conceptualising level to the body metaphor in the smelting process. The genderisation of the furnace is at a deep metaphorical level experienced as a woman since the female body is more open than the male body (Solheim 1989:69). Since the body is experienced as open and as a container, a womb and even a pot it is able to transform substances. In real life these substances for the well-being of a person are of uttermost importance. Human fluids, blood, semen and breast milk are the most vital but the body may also contain malevolent substances or take in nutrition or healing ones from magic and medicines that will effect the forces in the bodily substances. These substances often metaphorically conceptualised as ‘semen’, ‘breast milk’ or ‘blood’, may assist and cure a body suffering from infertility or sickness. Conceptualisations of the male body as being open and vulnerable, a transformer or transformed was equally important, something sexual abstinence, sacrifices and the use of magic in iron smelting exemplifies.

7.4 The body as a container Douglas argued that the body is a model or a closed system that may stand for any system. Using the body in meaning constructions and as a metaphor is not confined to language but also seen in artefacts and activities such as rituals (Douglas 1966:115). But why do body and gender divisions reoccur in symbol structures? Johnson provides an explanation through his argumentation for our experience of the body as a container, a metaphor we make in all sorts of contexts of orientation and construction of meaning (Johnson 1987 and see Chapter 2). Johnson suggests that our encounter with containment and boundedness is one of the most pervasive features of our bodily experience (ibid.: 20). Johnson argues that ‘we are intimately aware of our bodies as three-dimensional containers into which we put certain things (food, water, air) and out of which other things emerge (food and water wastes, air, blood etc.). From the beginning, we experience constant physical containment in our surroundings (those things that envelop us). We move in and out of rooms, clothes, vehicles, and numerous kinds of bounded spaces. We manipulate objects, placing them in containers (cups, boxes, cans, bag etc.)’ (ibid.: 20-21). In search of these cases there are repeatable spatial and temporal organisations. In other words, there are typical schemata for physical containment.14 Therefore, as Johnson concludes ‘if we look for common structure in our many

How then were deeply experienced metaphors for the body and its physiology part of Pangwa and Fipa social worlds and moral standards? How did the practical logic of body physiology work? And was its logic transferable to the logic in iron working, to the physiology of the smelting furnace? As noted by Lakoff and Johnson the similarity thesis is false because what happens in metaphorical thinking is a cross-domain mapping (Lakoff and Johnson 1999:126). This implies that all domains are equally important in the process of iron smelting as for instance in gardening, procreation, in drinking from

14

Johnson defines schemata as imaginative and non-propositional, not information rich images but only some basic features of an image. A schemata of a face has the basic features; eyes, mouth and nose, while an image of a face is more detailed (Johnson 1987:21-24).

98

Iron technology and symbolic reservoirs There is a taboo against having intercourse with a breastfeeding wife because it is also believed that semen from the husband can poison the milk. Therefore, until an infant can feed himself husband and wife are apart (ibid.: 399).18 Sometimes this sexual abstinence lasts for up to three years. Women who have just delivered are excluded from many daily activities such as cooking and cleaning. Women’s bodies are by both men and women held to have more heat than men’s bodies because they have breast milk. Their heat leads directly to blood loss in the form of menstrual flow, which is held to be the hottest of all blood. This caused women to be excluded from many daily activities such as cooking or fire making.

calabashes, when cooking or when making fire. Ordner has recently argued that in Bantu thought the essential is to uphold ‘the vital force’ (Odner 2000:69). This ultimately depends on the correct balance between ‘hot’ and ‘cold’ (de Heusch 1956, 1975, 1980).15 In describing the Haya people in Tanzania, Weiss has demonstrated how thermodynamic qualities of hotness and coolness lie at the base of conceptualisations of the equilibrium of the body (Weiss 1998:180).16 A similar division of the physiology of the body into ‘hot’ and ‘cold’ dichotomies has earlier been recognised in numerous Bantu-speaking societies and beyond (Barley 1994; Beidelman 1968; Berglund 1991; Flikke 1994; Heald 1995; Junod 1931; Sanders 1997; Strathern and Lambek 1998, Whyte 1990). Collett demonstrates the links between pots, people and furnaces but he does not explain how or why there is a logical link between them. De Heusch on the other hand has demonstrated how the Thonga lived by a physiological ‘heat code’ and a completely original thermodynamic philosophy (Heusch 1980:31, 39).

The furnace is located out in the bush and associated with the same kinds of morally correct behaviour in order to avoid iseke, hence the exclusion of impure persons such as a woman in delivery or in menstruation is logical. One can recognise the link between the heat code in everyday philosophy and how the furnace was decorated with female breasts. The task of decorating the furnace (and making a material metaphorical construction) was something the smelters performed with great care; they washed and smeared the clay body and breasts with water.

The Pangwa believe in a close procreative link between breast milk and semen. They claim that they originate or are created from semen in a combination with breast milk from their mothers (Stirnimann 1967:403,1979:181). A woman’s breasts are both male and female, the milk in the right breast contains semen, will later provide her with sons while the milk in the left breast will give her daughters. The Pangwa believe the foetus in the mother’s womb needs semen as nutrition until the seventh month of pregnancy.17 Therefore, one intercourse is not enough to conceive (ibid.). A part of the father’s semen from the multiple intercourses remains in the mother’s breasts during her pregnancy. Women often smeared their breasts with the husband’s semen after sexual activity (Stirnimann 1967:404). But from the seventh month onwards sexual abstinence was mandated and intercourse at this stage would harm the foetus. Women in Upangwa went out in the bush to deliver and a Pangwa woman who has just delivered is called umunyalisitu, ’a woman in the bush’ (Stirnimann 1979:254fn.2).

The Fipa conceptualise descent as a complex of marital and affinal relations in terms of the two composite symbols: head, unntwe, and ‘loins’, unnsana (Willis 1967:522).19 At the most abstract level the concept ‘head’ stands for the most inclusive Fipa descent group or category, uluko, with it’s elected chief and network of reciprocal duties and rights. In contrast to this stands the symbol ‘loins’ which represents the descent group as a changing entity through the exchange of women and bridewealth (ibid.). Less abstractly Willis sees the category ‘head’ as representing masculinity or maleness, patrilineality, paternity, intellect and authority while ‘loins’ are associated with femininity or femaleness, matrilineality, maternity, sexuality and reproduction. Finally, in terms of social categories and consistent with these abstract associations, ‘head’ denotes patrilateral relatives while ‘loins’ denote matrinlineal relatives (ibid.). Hence, the Fipa conceptualise descent in terms of embodiment and human physiology and the terms themselves refer to separate regions of the human body. As Willis sees it the literal meaning of unntwe is head in its physiological sense while unnsana, as loins, refers to the lower abdomen and lower back both in men and

When a boy reaches about six months he is taken with his mother out in the bush. With her goes her mother in law who brings two calabashes of water, one with female water and one with male water. The mother washes her hands and breasts in this water and afterwards presses as much breast milk as possible on the boy’s urethra. This rite is practised morning and evening as long as the mother is breastfeeding her boy. It is believed that the more milk she can ’feed’ her son, the more children he will have (Stirnimann 1967:404).

18 As Stirnimann states ’So findet es Die Frau selbstverständlich daβ der Mann sich einen Ausweg bei einer weiteren Ehefrau oder eine Witwe sucht’ (Stirnimann 1967:399). 19 Willis has analysed how opposed, complementary and associated ideas and values are contained in a basic sovereignty myth of the relations between Milansi and the intruding Twa (Willis 1967, 1981). Willis sees the binary discriminations in this myth as related to Fipa ideas about sovereignty. The kings of the two Twa dynasties (Lyangilile and Nkansi) conceptualised the leader of Milansi (the opposed and first dynasty of Ufipa) as a priest. Willis suggests that the king of Milansi had ritual sovereignty while the kings of the Twa dynasties had political sovereignty and related these to other sets of binary distinctions (Willis 1967:522).

15

Both Pangwa and Fipa, as numerous African peoples, conceptualise knowledge as a force or power (cf. Bekaert 1998). 16 Cf. also Gosselain 1999; de Heusch 1980. 17 This is common throughout sub-Saharan Africa (e.g. Barley 1994; Turner 1967).

99

Randi Barndon

Plate 7.1 A smelter wearing isuuli, Katumba 1990 women (ibid.). It adjoins but does not include the genital regions in both sexes. Still, the word unnsana has marked sexual connotations for the Fipa (ibid.: 523).20 Willis believes these Fipa categories to be what Turner called ‘dominant symbols’ (ibid. and after Turner 1964:35). They are the foci for sets of ideas from markedly different areas and levels of experience (Willis 1967:523). In his structural analysis of these symbols Willis suggests a system of oppositional pairs in which ‘head’ is associated with maleness, authority, intellects, seniority, lightness fewness, weakness and constraint as opposed to the ‘loins’ which represents femaleness, sexuality, reproduction, juniority, heaviness, numbers, strength and fellowship (ibid.).

‘loins’ its impaamba. Childs suggested that, although furnace height functioned to draw up the natural draft during smelting, it also symbolised the wisdom and control that the master smelter/husband needed to ensure the productivity of his furnace/wife (Childs 1991b:346).21 If following Willis’ distinction between sets of ideas and symbols into the oppositional but also complementary categories ‘head’ and ‘loins’, the secondary refining furnace, ichinteengwe, certainly is something also for the eye, a material metaphor. It follows Willis’ analysis and line of thought in Fipa imagery and the idea of head and loins and the secondary furnace with its impaamba, ‘inside of female thighs’ modelled in clay.22 As empirically demonstrated from various ethnographic cases there were clear visible signs of the furnace being a woman. The smelters also named the various furnace parts in terms of genitals and when actually conducting the technological process of fuelling the furnace or blowing the bellows their actions mimicked the sexual act. The sacrifice of the white cockerel and its head placed within the furnace chamber pointing towards the rising sun connotes maleness and male powers needed within the furnace womb.

Explaining how these complementary oppositions are explicit in Fipa thinking, Willis notes how a Fipa who is asked for his descent name ‘on the head side’ will reply with a name he inherits from his father and which is transmitted patrilineally (ibid.: 524). Willis concludes that these contradictions are felt rather than seen because a structure is apprehended through the ear instead of the eye. In Willis’ words ‘the reason why Lévi Strauss choose music to represent the spirit of the mythical thought is so applicable’ (ibid.: 527).

21 Childs does not comment on the obvious link between ’loins’ and the elongated clay masses on the secondary refining furnace of the Fipa, but notes that the Shona who have the same clay masses outside the furnace rake hole, in this way represents human legs open at the base of the belly (Childs 1991b:349). 22 If connecting these embodied binary distinctions to cosmology they may be recognised over a larger area; into Utabwa on the eastern side of Lake Tanganyika (Roberts 1991; Willis 1991) and southwards into Mozambique among the Shona (Childs 1991b).

The Fipa linked the tall shaft furnace with ‘head’, above or up as contrasted to the second refining furnace with its 20 Complete muscle control of this region is considered a sign of erotic maturity. It is the object of a style of dancing called imiteete, which is taught to pubescent girls and which boys imitate, to facilitate such control (Willis 1967:523).

100

Iron technology and symbolic reservoirs fuels and ores.24 The smelters said the furnace ‘bride’, nawiinga, was mature. The spells to make sure all the doors were closed, Kalinyango na Kama, Kalinyango na Kama uttered in front of each tuyere inlet or around the furnace base was repeated in order to make sure all its orifices were closed and the furnace protected from malevolent forces. The blood sprinkled around the doors, the smelters and children moving through the main door and into the body of the furnace emphasised its bodily openness and vulnerability. The magical substances, the ‘things that enter’, ifingila, were placed in the trench and at the furnace base in order to secure what went into the furnace from harming its inside. Medicines were applied to strengthen fertility and cure any sickness. Magical and homeopathic treatments were forceful in terms of gender combinations and in terms of metaphors of containers. The furnaces were containers as were the intaangala basket, the ing’ano pot, the smelters and their women.

I will argue that for many Fipa and Pangwa people the body is experienced as a container but also as open. Thus, the body is constituted of various substances and forces, vital for living. The state of the body is commonly referred to as ‘hot’ or ‘cold’. Blood is the most crucial but the character of blood is ambiguous and ambivalent. It is a life-giving force but too little blood or too much overheated blood, such as after sexual intercourse, may be dangerous. When a change in temperature occurs, body fluids change and the body is transformed (cf. de Heusch 1980; Weiss 1998). Too much coolness may turn a person weak or a woman infertile while also too much heat or overheating as after sexual intercourse weakens a man and harms the foetus in the female womb. The blood and the temperatures of the blood running through the body are vital for living, not becoming sick or even dying. The thermodynamics in one person’s body may effect in a good or dangerous way other people who come into contact with this person. In iron smelting the human body modelled in clay on the furnace shaft made the furnace into a material metaphor of the female body. It was a female body because of the properties and functions of the female body; its maturity and potential fertility, its menstruation and it’s menopause. Most importantly, like a woman it had the possibility to become pregnant and thus a container for transformations.

7.6 The naked smelter When discussing the isolation of smelting sites Childs and Killick have suggested that smelting was carried out far from villages because of the exclusion of women and the secret rituals that were carried out in order to prevent sorcery or interference from ancestors (Childs and Killick 1993:325). According to Herbert ‘the uncivilised and wild naked smelters had to conduct their work in secrecy because of the forbidden nakedness’ (Herbert 1993:9339). This view on nakedness has roots back to our more Victorian based notions of nakedness and specifically sexuality as linked to secrecy (Caplan 1995:6, 287). Can sexuality, secrecy and nakedness in iron smelting be taken as ours, based on our moral standards for behaviour?

Washing in ritual water, kiwalasha, was essential in order for a Pangwa person, male or female, to remain ‘cold and clean’, awalele, in most daily activities. Before most activities people would wash in kiwalasha, which is a white liquid made from a locally growing tree. Its juice or milk is mixed with water and used before all sorts of purification rituals (cf. Stirnimann 1976:402),23 before a marriage ceremony, sexual intercourse, birth celebrations, death rituals, offerings and sacrifices and rain making rituals. Moreover, it was used before pottery making, hunting, gardening and iron smelting. When charcoal preparations for a smelt were carried out in the bush all participants had to wash in kiwalasha. Once iron ore had been transported to the smelting site the smelters had to remain cold and pure and they washed themselves with kiwalasha. Prohibitions against sexual intercourse were once again mandated and while they were at the smelting camp they were not allowed to wash themselves, cut their hair, get angry or violent because these activities could change the balance of hot and cold in bodies.

Herbert suggests that the nakedness of the smelters during smelting corresponds to the isolation of the smelters in the bush and sees the smelters as mediating a role between the wild and the civilised (Herbert 1993:9394). She has also suggested that the nakedness in iron smelting may be further linked with special clothing as bark cloths, skin cloths or even special skirts, because they all link the smelter more closely with nature (ibid.). This may of course be one explanation, although I suppose greatly coloured by western ideas of nakedness as constituting something more natural than civilised dressing. I believe, as demonstrated in my case studies, that the importance lay in making everybody aware of the sacredness in sexuality and secrecy that surrounded times of smelting. A view that values the sacredness in sexuality (e.g. Sanders 1998), sexuality as vital life forces of both men and women has not been thoroughly discussed. The concept of sexuality and procreation was part of the smelting process and was used by the head smelters in order to stage aspects of their folk model of

In the Fipa case the smelters would dance, sing and actually move their bodies as if having sexual intercourse with the furnace. The master smelter physically entered the furnace, the furnace chamber and the furnace ‘womb’, through the western tuyere port called ‘the mother door’. When he entered the furnace he moved his body as in gestation. From the inside of the furnace he arranged the tuyeres at their inlets and finally ‘fed’ the furnace with its

24 23

During one of our re-enactments the master’s assistant, unnsole, followed the master’s actions during the fuelling of the furnace.

Cf. also Stirnimann 1976:49. fn.1.

101

Randi Barndon morality. This model was different from Victorian based notions of sexuality.

Containers of all sorts, not only bodies, but also furnaces, pots, calabashes, granaries and even houses are open and may contain powerful and malevolent forces.27 This, rather than secrecy or embarrassment because of nakedness caused the smelters to enclose the smelting sites with furnace ‘guardians’ and furnace ‘caretakers’.

Greig reports that the smelters used special clothes during smelting (Greig 1937:79). Wembah-Rashid has pointed at how magical purposes of personal decorations and dress have not been extensively explored and notes that during a re-enactment in 1967 he observed a Fipa smelter wearing a goatskin in front of his genitals (WembahRashid 1973). I observed some of the smelters wearing isuuli over their ordinary clothes during our reenactments (Barndon 1992:101 and see plate 7.1). When I asked them why they were wearing isuuli they said, not surprisingly that they did so in the past and still had to in order to smelt. But they also told me that the proper name for the main tuyere inlet on the natural draft furnace was pamweela palinyina meaning ‘the mother genitals with a waist cloth’ or ‘the loincloth of the mother opening’.25 I was told that in the past Fipa smelters had a goatskin, isuuli, around their waist during smelting while they in everyday life were naked. They were protecting the furnace against male powers and themselves against female powers.

7.7 The baby and the bloom In order to explain technological processes peoples have often adapted models drawn from their own human natural and social environment. Eliade has shown how models in metallurgy were based on analogy for the growth of the bloom in the furnace chamber to the growth of a foetus in the female womb (Eliade 1958).28 The body-oriented taboos that we find in iron smelting can be transferred back to real life experiences of procreation and giving birth. The furnace’s husband was the master smelter and because of this other women were kept away and especially pregnant or menstruating women were feared (Brelsford 1949:28; Killick 1990:123). In real life one of the several wives of a Fipa or Pangwa man, the one whose turn it was to sleep in the husbands hut would also cook his food (cf. Herbert 1993:81-82). When the furnace had been filled up with fuel it was (metaphorically) pregnant and as in real life separated from other women, excluded from many daily tasks and from sexual activity.

Before the smelting could start in Upangwa the old men made waist clothing of grass. In the past Pangwa men never dressed in everyday life, yet they did so during smelting. In our context they had the skirts on top of their ordinary clothes. As opposed to being naked the smelters rather tried to hide and veil their ‘open bodies’, their body orifices and powers as well as those of the gendered furnaces. While working the bellows the Pangwa said that if the bellow worker was ‘hot’ from sexual intercourse, poison from the furnace guardian, umboocha, would come out and give him pain in his stomach.26 This exemplifies how metaphorical thinking in terms of containment and embodiment was not confined to the female body; too much heat could harm men as well as women and furnaces.

Both Pangwa and Fipa say that men and women contribute equally to the formation of a child. Hence, the cultural categories ‘male’ and ‘female’ are complementary. The Pangwa clay furnace was modelled with attributes of the female body. The furnace shaft was carefully decorated with female breasts, mawele, and orifices, such as ‘the wind maker’s hole’, kwibepo, and front opening, umlomo. When I asked my informants why, they said, not surprisingly, but explicitly using metaphorical thinking, that the furnace was a young and fertile woman ready to give birth to a child. The Pangwa gave the furnace its identity as a woman when they decorated it with female breasts and they called the furnace chamber a female womb or a pot. They conceptualised what happened in the furnace chamber with what happens in procreation. When they made visual material metaphors they created a link between wombs and pots and placed a pot, ing’anjo, containing male and female medicines and magical substances beneath the furnace chamber. They recognised their own containment by the fears they connected to blowing the bellows. If a man was too ‘hot’ as after human sexual intercourse the furnace guardian, umboocha, would pour out poison and the bellow worker would feel pain in his stomach.

Special clothing or nakedness in iron working was directly related to specific perceptions of the body. The body, naked or dressed in bark or skin was a mediator between the transformation in the furnace and the transformative powers and magical skills held by the master smelter and in bodies in general. In the Pangwa case, the front of the furnace with its ‘mother opening’, libepo or umlomo, and its female breasts as well as the smelters with their genitals covered during smelting, indicate a perception of the body as being a container and open and the furnace as a container and open. The medicine pot, ing’anjo, with its perforated lid was treated like the Fipa’s tuyere inlets, sprinkled with blood and protected against malevolent forces and the openness of containers, in this case the furnace, was guarded.

27 Collett links pots, people and granaries because they all have the same decorations (Collett 1993, and see also Barley 1994:53). 28 Eliade has highlighted how common it has been throughout times and in different cultures to conceptualise metals as a foetus. Within the womb of ‘mother earth’ one believed that ores over the centuries had matured into metals that could be used in order to produce raw materials (Eliade 1958).

25

Cf. chapter 6, the caretaker, umpakasi, hanging in front of the main tuyere inlet was also called pamweela palinyina and believed to warn off hot people or witches. 26 Interestingly, although not specifically related to iron smelting or embodiment, Willis notes that if a Fipa person falls ill with a stomach complaint, poisoning is the immediate explanation (Willis 1968:148)

102

Iron technology and symbolic reservoirs formulates it ‘in this respect male and female together can, quite literally, make things happen’. To join the genders ‘is to generate, to create and to transform’ (Sanders 1998:243). From this Sander argues that there is an inherent paradox in this thinking because once the gender categories are combined they cannot easily be recombined to effect further transformations. Because of this the powers in gender combinations are threatened by their very development. Hence, once combined it logically follows that gender categories must once again be separated if they are to regain their original generative power. Therefore, the generative powers in gender must both be combined and separated (ibid.). Once the correct amount of bodily fluids of hotness and coolness has been combined the generative powers of gender may transform. Too much may turn a person overheated or sick while female powers without male or vice versa is useless and without transformative powers.

The Fipa tall natural draft furnace had no visible materialised markers such as the female breasts on the Pangwa furnace. Yet, once decorated with flour and flowers it became a nawiinga, a woman ready to marry. The decorating of the furnace with maize flour (or charcoal) and flowers on the furnace top resembled the body decorations of a bride and bridegroom with clay, water and beads, yaeemba. The secondary furnace, ichinteengwe had clay made thighs that echoes the metaphorical thinking related to the ‘loins’ and femaleness in their conceptualisations of descent (cf. Willis 1976:523). This furnace revealed the outcome of the process; if the bloom was pure and forgeable and the furnace womb was called a cleanser or a cooking pot. Both the Pangwa and the Fipa combined metaphors from Collett's triad of heat-mediated transformations within the same material context. This line of thought is what we see in procreation. But the container was primarily the human body itself not only the female womb that nourishes the foetus but the physiology of the human body in general. In line with the philosophy of thermodynamics and the body as a container of hot and cold substances, blood, semen and milk, the powers of men and women, were bodily fluids and generative forces. These substances are a vital force in the human body as a thermodynamic system. The conceptualisation of the forces in the furnace was of a similar kind.

There is perhaps a contradiction in women being kept away from iron working and a woman being the container for the bloom. But if we consider a real delivery, women conceptualised as impure, hot or in menstruation were equally dangerous for a real woman as real women were for the furnace woman. If a pregnant woman ate hot food its heat could cause the foetus to burn (Stirnimann 1976:231). Moving in and out of metaphorical domains, not giving any of them pre-eminence was a way the smelters used moral standards from real life as effective dynamics of the smelting technology and as parts of their chaînes opératoires of the smelting process

What went on in the furnace womb was like what went on in a woman’s body. The thermodynamic physiology of the human body was not confined to the female body or the furnace as a female but also the master and his assistants. Heat was related to the womb, to food taboos and eating taboos since hot food could not be eaten before smelting as it could cause misfortune just as hot food could cause a foetus to burn (Stirnimann 1979:231). Smelters could not eat hot food before smelting because of the same fear of becoming overheated and thus transmitting their hotness to the furnace.

7.8 The body and magic There are two interrelated ideas that interpenetrate rituals and magic in African traditional iron smelting. These are ideas of fertility and reproduction and the ideas of a protection against evil and malevolent forces. These ideas are in general seen as crucial for the well-being of the body. The forces, good or bad, affected the thermodynamics of the body and if out of balance, ‘hot’ and ‘cold’ substances, as in overheating could cause a person to be sick. However, these forces were interrelated in such a manner that it is often difficult to distinguish the boundaries between them. External factors, spirits and malevolent forces could also cause a person to become sick or die.

The thermodynamics of bloomery furnaces were equivalent to thermodynamics of human bodies. The furnace temperatures were controlled by the rate heat or generation or loss of heat (Avery et al. 1988:263), but the human body was equally experienced in terms of hotness or coldness. Heat must also pass out of the fuel at the furnace temperatures so that the ore is reduced when it has absorbed heat, both from the fuel and furnace walls (ibid.). In conclusion then, reproduction and production were equated (Davison 1993:37). The identification of the furnace with a woman and the process of smelting with fertility-related elements are widely recorded. Like the Pangwa and Fipa, the Chisinga smelters in Zambia said the furnace was ‘pregnant with iron’ (Brelsford 1949:28). The bloom was equated with a foetus and therefore sexual misconduct could kill a bloom as it could kill a child.

Willis sees the Fipa medicine men as distinguishing between two sorts of maladies, those which they regard as peculiar to the surface of the body and those, which they regard as existing inside the body. Those on the surface of the body are regarded as minor and may be treated with herbal remedies. On the other hand, serious sickness is seen as analogous to an invasion of the body by a hostile force (Willis 1978b:142). Curing the intrusive agency that has caused the sickness, a Fipa medicine man must first determine what kind of intrusive forces that caused the suffering of the sick person. These follow three ‘paths’: the path of the territorial spirits, the path of the spirits of a persons kindred and the path of

The culturally defined gender categories ‘male ‘ and ‘female’ are a source of power and thus provide a means of purposively acting upon the world. As Sanders 103

Randi Barndon sorcery (ibid.: 143-143). When a Fipa doctor has established what path the person’s sickness had worked along he goes into the bush to collect medicinal bark which will form the primary constituents of a preparation. These ingredients are called ‘the pieces of worked wood in themselves’ (ifiti fikola) (ibid.: 144), and resemble the master smelters use of a collection of barks, ingailo, collected before a smelt (cf. Cory 1949 and Chapter 6). After obtaining these medicines the Fipa doctor takes out the ingredients from his intaangala basket. Willis describes how the medicine man, in a similar way to the Fipa smelters, selects small pieces from the basket and that these ingredients do not act directly but rather symbolically. Many items described in Willis list of contents in a medicine man’s intaangala are similar to my own observations (cf. Willis 1978b:145 and Appendix F). Thus, the magical basket of an ordinary Fipa medicine man was called intaangala and his local pharmacopoeia formed part of everybody’s daily lives (Barndon 1992:9195; Willis 1978b:144-145).29 The Fipa master smelter was not an official or ordinary medicine man like the Pangwa umkoyo could be, but he was a magician.30 He was able to enter into the role of a transformer, assisted by his use of magical substances.

In Upangwa my informants told me that in the past a blacksmith who was not a medicine man could hire an umkoyo to perform rituals and assist in building his furnace and conducting a smelt. The body was an anchor and perceived as a container but embodiment was also external or extra corporeal in terms of the forces outside the body, hence the furnace body and human bodies were not only closed systems, but also open. The fact that the three Pangwa elders were also medicine men, amukanga, was made obvious to us over and over again. From the first day and during the entire period of preparations for the smelt the aged men announced their skills in various ways. They constantly collected and brought in new ingredients for the medicine pot and they cooked the beans, imono, in a pot at the smelting camp.31 This task was paid a lot of attention; the master constantly stirred in the pot or checked its contents, exactly as Mauss explained. Thus, the preliminary rites and preparations were among the most significant. A Fipa master smelter like his apprentices was known to have magical skills and was believed known to move into a trance and trace thieves, something not paid enough attention in studies of smelting symbolism, because we have been too occupied with models adapted to notions of female fertility. Among the most significant of the preliminary rites was the use of magic, which by definition may be said to include offerings and certain prohibitions such as food taboos and taboos against human sexual intercourse and violence.

The Fipa division of ingredients used by medicine men into two categories is, as Willis notes, a widespread system of thought in Bantu Africa (Willis 1978b:151.fn.19). A similar notion might be recognised in Pangwa magical and medicinal treatments of their furnaces, although both in my re-enactments and Stirnimann’s inquiries the medicines seemed to be the largest amount of ingredients and paid more attention than the magical ingredients (Stirnimann 1976:277-278 and cf. Appendix F). However, The Pangwa sacrificed a goat, not a rooster, and explicitly offered food and beer to the spirits of the place and to former smelters.

Malinowski’s (1935, 1948) discussion of magic demonstrated that magic is not a primitive science or a confusion of the natural and supernatural. Magic works together with, and in addition to, the other realms of knowledge practised in a technology. Medicinal and magical substances that were part of the smelting process, part of technological know-how and skills were placed in foundation trenches and in bundles in the ground as the Fipa did, or in pots as the Pangwa did (Barndon 1992:75; Mapunda 1995:228, Schmidt 1997:250).

The Pangwa literary meaning of furnace, liteende, was ‘the one beneath’ or ‘something beneath’ referring, I think, to the pot of medicines that was buried beneath the furnace chamber and substances beyond the body being able to transform it. The master smelters in Upangwa were medicine men and magicians but they had no specific basket of magical substances like the Fipa’s intaangala. Still, their herbs and plant medicines were of a similar kind and with similar connotations and metaphorical imagination linked to curing as the Fipa ingailo barks or medicines. Since they were practising their skills as herbalists and medicine men on a more daily basis their knowledge was not specifically symbolised or stored in a smelters magical basket. However, they carried with them their magical equipment as I had seen in use before the third smelt in Ufipa when additional use of magic was believed to secure the success of our third re-enactment.

In his discussion of magic Mauss argued that the preliminary rites or the magical preparations, were the actual centre of the rite and not the presentation or the performance of the magical materials themselves (Mauss 1950:53). Magic, Mauss continued, ‘is the art of preparing and mixing concoctions, fermentations and dishes and magical substances were chopped up, pressed into special shapes, formed into images, they were drunk, eaten, kept as amulets or used in fumigations’ (ibid.). This cuisine, this ‘pharmacy’ or ‘chemistry’ as Mauss called it, not only caused magical materials to be utilised, but served to provide them with a ritual character which contributed to the efficacy of magic (ibid.). The Fipa’s intaangala contents were chopped up and pounded in as similar way to the Pangwa ing’anzo, beans that were

29

Willis notes that a sorcerer, unsoosi, was a person who had turned his knowledge of in’gaanga and intaangala into the service of evil and anti-social ends (Willis 1978b:141, fn.8). 30 In both regions they were exclusively male.

31 Bekart notes that for a Sakata healer the performance itself; the stirring with a stick into a shell referred to sexual intercourse (Bekaert 1998:21).

104

Iron technology and symbolic reservoirs Each piece of magical substance was examined and its use was announced and explained (Barndon 1992:91-95). The substances in the Fipa head smelter’s intaangala were carefully selected, most of them were specific to the smelting context but some were used to cure and heal human bodies were part of contents in ordinary magical baskets. The selection of barks for the ingailo was also part of general medical treatments. Some of them were supposed to cure measles, vomiting or inchila, the most severe invisible disease, cased by adulterous behaviour. As the magical and medicinal substances were placed in the furnace trench or at the furnace base, they became ifingila, ’things that enter’ (ibid.).

mixed with water and oil for the kiwalasha in the medicine pot. As described above, the magician underwent special preliminary rites. The eating of a special porridge or sexual masturbation were ritual acts that involved the individual alone. The magician and who ever was associated with him, had to remain chaste and pure, washing and anointing their bodies prior to the rite and prior to smelting. They had to fast or abstain from certain foods and they were told to wear special clothes. They wore masks to disguise themselves or they were naked (i.e. Schmidt 1997). This was either in order to remove all barriers between themselves and the magical forces or perhaps in order to act through ritual (Mauss 1950:48). Among the many prescriptions Mauss mentioned some were indeed seen as a part of practices and prescriptions in the smelting processes I observed in Upangwa and Ufipa.

Killick makes an important observation when he points at that since the charge and combustion in the natural draft furnace chamber cannot in any way be controlled during the actual smelting it becomes even more crucial to please the ancestors and spirits. Sexual taboos were mandated not only because there was a metaphorical link between furnaces and humans, but also because the smelters wanted to please the spirits (Killick 1990). This might be one of the reasons for the elaborate rituals and especially the use of the two children in the ritual at the foundation trench or inside the furnace in the Fipa case. But a sacrifice was also conducted for the spirits of the place of a Pangwa furnace. In both context the furnace trench or pit was named a grave or death. The Pangwa named the foundation uluuto while the Fipa called it ulufula. This metaphorically linked the living furnace and master smelters as magicians and medicine men with the ancestors and spirits. Thermodynamic perceptions of bodies and general healing and homeopathic practices were linked to morality and the goodwill of the sprits and ancestors.

Activities that affected the body, the well-being of persons themselves in their daily lives were initiated long before the preparations and building of furnaces and smelting started. Eating rituals and food taboos followed sexual taboos and were part of everyday life. Pangwa smelters slept close to the furnace the night before the smelting started. For the masters this meant entering into the role of a transformer and a magician. A Fipa master had to eat insima, a millet porridge in which was put the head of a white cockerel the night before the smelting started. This marked how he was entering into the role of umwaami, a master and magician. The meal was prepared for him by his wife, who was the guardian of the master’s intaangala and who kept the basket in a safe place in the house during the non-smelting season. Wyckaert noted in his observations that the magical basket of Fipa smelters was a sacred object and treated with respect also in between smelts when it was left in the master smelter’s house (Wyckaert 1914:274). The same was reported in Roberts and Wembah-Rashid while Wise denies that the magical basket was given any special attention (Roberts 1949:240-242; Wembah-Rashid 1966:66; Wise 1958b: 233). In our re-enactments the magical basket was treated with care and respect and on the day of building the furnace it was in the centre of the parade of smelters walking to the smelting site. Special care was taken to perform the preparations of the intaangala, cutting pieces of selected items in a correct manner. A small black he-goat skin was laid out on the ground before the content from the basket was poured out.

The placement of medicines or magical substances defined the importance of what was beneath and above, what was there and based on embodiment but also extra corporeal and disembodied. Hence, the furnace body and smelters bodies were both containers and open systems. The master was not allowed to have sexual intercourse with any of his wives, but the night before the work started he would masturbate and not wash his hands. The two children who assisted the master had slept in his house and were told not to touch their bodies or scratch if insects bit them. Touching their bodies could interfere with the cleanness they possessed and which through the rituals would then be transformed to the furnace. The next day the smelters coloured their forehead with a red powder, unnkolo, also used in wedding celebrations. This announced to everybody that they were smelters. The children carried a white cockerel, unnkoko, and the magical basket and medicinal bark followed by a parade of smelters out of the village to the smelting camp. Under the master’s instructions the children performed the rituals and sacrifices at the furnace foundation ensuring that the acts were performed in coldness and cleanness. This links to human bodies because the ingredients used

Careful selection of items for each smelt was observed.32 32 Both in Ufipa and Upangwa the better we have come to know each other the more the smelters have revealed of local magical practices. It seems they strongly wanted to succeed and to prove their mastery of the technology. A similar observation has been reported from Cewa reenactments. The smelters were at first only partly concerned with rituals and taboos while before the last re-enactment they were all more concerned and because of this they were all wearing bark cloths (Avery

1988:268).

105

Randi Barndon use of magic. The treatment of the embodied furnaces resembles the homeopathic or magical treatment of fertility or infertility in female and male bodies. The treatment of furnaces and bodies were similar because of equivalent perceptions of thermodynamics in the two domains and similar notions of transformations within them. The treatment of furnaces was like treatments of human bodies with magic and medicines. The two domains, metallurgy and magic, were not separated, as they would have been if the point of departure has been our western based notion of technology. Furnaces were treated like bodies because they are bodies. Furnaces were conceptualised just as bodies are conceptualised. They are both containers and open systems. Therefore, furnaces and bodies were treated the same way, with the same set of homeopathic medicines and magic. They were bound by the same rules of avoidance, separations and seclusion and especially the same taboos against sexual intercourse. Finally, the same use of fertility medicines to secure a successfully delivery was used in securing a successful smelt. The body as an anchor may be seen as an entrance into iron making as meaning making.

in the magical preparations both in Ufipa and Upangwa furnace treatments were of a similar kind as those used to prevent sickness or curing of illness or infertility in human bodies. The furnaces were conceptualised as humans and because of their embodiment they were treated and cured as human bodies. The medicinal substances were in both cases placed beneath the furnaces, in the trench or in a pit. The medicines were extra corporeal, but things believed to affect the furnace body when they entered into the furnace womb. The ingredients in the magical pots or bundles were a combination of male and female powers that should strengthen fertility or cure infertility. The use of herbs, the use of magical substances, the use of sacrificed blood from the sacrificed cockerel (Fipa) or goat (Pangwa) may be seen as analogous to human blood, semen and breast milk. Many Bantu-speaking people believe that it is a combination of blood and semen that makes a woman pregnant (after Heusch 1980). The furnace charges included the magical substances that represented the body fluids and the combinations of these transformed the ore into iron similar to the way in which thermodynamics in bodies and hotness in sexual activity results in a foetus. Anger could cause a fire to die out because the ‘heat’ of fire was not compatible with the ‘heat’ of anger and heat in sexual intercourse. The various ‘heats’ or energies should be kept apart in order to maintain and secure the correct balance of substances. Magical substances, medicines for fertility or against adultery as well as actual sexual abstinence controlled the hotness or overheating.

7.9 Chaînes opératoires and symbolism The fact that the Fipa used two furnaces in order to reduce ores to iron resulted in the process being constituted of two separated procedures or sub-sequences. The first part was conducted in a furnace that once charged was left to reduce the ore by it-self. The second furnace was operated by human labour, by men constantly working the bellows in a similar way to the Pangwa technology. Childs’ observance of the performance of rituals prior to a smelt, often while building the furnace structure has proved to be correct in my Pangwa and Fipa case studies.

Commonly there is an oppositional association between heat and anger. The master smelters put taboos of heat, of anger and sexual intercourse upon the smelters. Anger is a kind of heat that is commonly seen as incompatible with the heat in sexuality. Anger may harm fertility (Herbert 1993:89; de Heusch 1980; Rowlands and Warnier 1988, 1993). The forces in anger are like menstrual blood, a kind of hotness incompatible with the heat in sexuality. Following the taboos against anger or violence a Pangwa smelter was not allowed to speak loudly or to be angry during the bellow working (see also Warnier and Rowlands 1993). When a person is angry his heat is hot in a way incompatible with sexual hotness. Stirnimann describes that there was an important taboo against anger after the offerings to the ancestors (Stirnimann 1976:276). The experience of heat were drawn from bodily experiences and connected to moral models and images. Moral order and how to socialise and organise ones life was played out in body related metaphors of copulation, procreation, hotness, heat and fire. Keeping these forces under control or in balance was tied to the knowledge of the master as a magician.

The Fipa smelters sacrificed a cockerel at the foundation trench when a new furnace was built and after the repair work on the furnace shaft when they reused a furnace. But both immediately before the Pangwa and Fipa smelting begun the master smelters once again sacrificed, offered drinks and food to the ancestors and the masters went into the role of transformers. In Upangwa the smelters danced in circles around the furnace before they started working the bellows. In Ufipa the master and his assistant moved in circles around the natural furnace before the fire was lit and they climbed down the scaffoldings. Although Child’s lists of sequences outlined in Chapter 4 is valid both for the Pangwa and Fipa smelting processes that I have described in the previous chapters, my studies indicate slight revisions of these sequences. My fieldwork has confirmed Childs’ general insight that during two separate sets of operational sequences rituals were of specific significance. The first, Childs suggested was at the actual building of the furnace or when the building was about to start while the second was before smelting begun (Childs 1991b:340, cf. also MacLean 1996:29).

The relation between technology, embodiment and disembodiment is seen in how the use of the body, its state or well-being is related to the powers of gender combinations because the body is both a closed system or a container and open. Gender combinations were powerful or were protected (against overheating) with the 106

Iron technology and symbolic reservoirs MacLean argues against Childs’ list of basic sets of operations because she reads it as a chaîne opératoire (MacLean 1996:28).33 MacLean suggests a refining of Child’s sequences and proposes the inclusion of two additional pre-smelt components; the decision to conduct a smelt and selection of an appropriate time for smelting (ibid.: 29). She also comments on Childs’ notion of performance of rituals as separate categories and argues for ritual as an integral component of the process (ibid.: 30). What MacLean means by both ritual component and integral part does however seem to be similar to Lecthman’s initial notion of specific and stylistic modes of thought in technology (Lecthman 1976:13-14). This is a perspective that was also discussed by Childs with her basic sets of operations (Childs 1991b). Lecthman’s perspective has been especially influential in studies of African iron working (e.g. Barndon 1992, 1996a; Childs 1991a, 1991b; Childs and Dewey 1996; Childs and Killick 1993; Collett 1986, 1993), although some have questioned this notion of style (e.g. Boast 1997). I partly agree with MacLean but believe the ritual component to contain ‘symbolic reservoirs’ (Sterner 1992) and ‘symbolic repertoires’ (Bekaert 1998). The metaphorical thinking was partly an evoked symbolisation that for the participants in the smelt was a materialised integral part of the technological activity. Without metaphorical imagination and embodiment there was no way of practising the technology. We should not distinguish between function and symbol implying that symbolisation starts when function ends (after Gosselain 1999:21). The smelting context was a ritual and social event as much as a technological or practical one. There was no real divide between technology and symbols or signs in the way the process was practised and understood.

Returning to my initial argumentation for an acceptance of multiple realities, Bekaert has distinguished six separated but integrated levels in the symbolic repertoire used by Sakata smiths in Congo. In between what Bekaert calls the phenomenological ‘type’ and structuralistic ‘binary opposition’ lies a host of other semantic operations that are at work (ibid.: 16). These levels of meaning are according to Bekaert (ibid.: 16-17) of:

Nevertheless, for the master smelters the way metaphorical imagination was integral to the process was a ‘tool’ ready at hand to mobilise a labour force on the one side and control entry into the occupational group on the other. The master smelters used metaphors from their symbolic repertoires in order to render meaningful the technological process. Finally, evaluation of the results of the technical and social processes was crucial for the master smelters’ reputation. Tambiah has argued that the practitioners in a technology, may shift into and out of different orderings of reality, but we cannot completely separate these realities from each other (Tambiah 1993:136). The smelters I observed moved in and out of different realities, but they were engaged in both magical and technical activities. But they were metaphorically constantly criss-crossing and bridging domains. Notwithstanding, they were aware of the multiple realities or levels in their structured models. This was seen in how they explicitly used various levels in the symbolic repertoires, such as hunting rituals or symbols of fertility during defined stages in the process.

What Bekaert calls typified experience is what would be ‘normal’ for the smelters and exemplified in answers to why women were prohibited from entering the smelting camp as ‘it has always been like this’ or ‘it is our custom’ etc. Things are ‘taken for granted’ or are seen as ‘normal events’ (ibid.: 17). This relates to the second level, of a pragmatic motive in which phenomenology situates the everyday life of working as action-oriented and with a pragmatic motive, hence ‘this is the way it works’ or ‘it would not work if women were around’ (ibid.).

(1) typified experience; (2) pragmatic motive; (3) metaphysical intervention; (4) experiential gestalt; (5) explicit metaphor; (6) codic opposition34 Bekaert’s thesis is that the metaphors of smithing, and I would add smelting, may be distinguished and that humans inhabit and draw on many worlds, from one repertoire simultaneously. Bekaert’s approach is particularly valuable since it takes into account both phenomenology and structuralism. Like Bourdieu, Bekaert demonstrates the validity of combining at the outset incompatible approaches in order to understand meaning and symbolism. Bekaert rejects the ‘one-worldism’ of either functionalism or semantics or of either structuralism or phenomenology. In his multileveled analyses Bekaert argues that phenomenology is suited for pre-reflection while structuralism holds for reflection (Bekaert 1998:25). Phenomenology describes cultures as enormous repertoires while structuralism can read it as a perfect system (ibid.). He sees the symbolism as moving from typified experience to metaphors and finally codic oppositions (ibid.: 24).

At the level of metaphysical intervention we can recognise the role of the master smelter as a magician and as a human in contact with the supernatural, former master smelters, ancestors and spirits whose assistance the master need in order to become a transformer. According to Bekaert the smith, and again I add the smelter, ‘is the ultimate sorcerer who is making iron and dealing with invisible forces which he masters or 34 Bekaert adds another level, the level of metonomy in-between experiential gestalt and explicit metaphor, that I have excluded in this discussion. The level of metonomy is according to Bekaert where we see resemblance in productive forces such as ’flowing’, ’being strong’ ’transforming’ etc which he link with a Bantu mode of though of the image of life force as a ’stream’ (Bekaert 998:20).

33 MacLean does not refer to Leroi-Gourhan, Gresswell or Lemonnier or in her use of this concept (MacLean 1996).

107

Randi Barndon male/female oppositions were also crucial when placing the furnace in its specific direction or with its specific gendered features (Barndon 1992). I then recognised them as binary oppositions, but this does not imply that they cannot also represent deeper metaphorical imagination or senses of being-in-the-world. The fact that they may be representations of a duality does not imply they may not also be seen as ‘twins’ rather than only incompatible; as have been discussed in terms of the forces assigned gender combinations. As I see it the gender oppositions in smelting were specifically used by the master smelters to communicate aspects of importance for the well-being of the furnaces. The codic oppositions became part of the symbolic repertoire the smelters used based on deeper experiences of the body.

manipulates’ (Bekaert 1998:18). Willis’ distinction between magic and medicine in Fipa culture captures how the medicine man both uses homeopathic medicine and magic in curing people (Willis 1978b). This mixture of ‘magic’ and ‘medicine’ has been recognised in how the Fipa smelters used both barks, ingailo, and magical substances, ifingila, when curing furnaces and connecting themselves to the spiritual or metaphysical world. At the level of experiential gestalt one unit (i.e. the human body or a furnace ) is not a metaphorical extrapolation of the other. They are mutually constitutive, with many other experiences because there is no abstract idea underlying the gestalt (Bekaert 1998:19). Here we see how lived experience is both pre-reflexive and seized by reflection (ibid.). The elaborate embodied symbolism and explicit use of metaphors was one level in the symbolic repertoire that was acted out as parts of the technological activity. I have discussed how embodiment and femaleness/maleness were made into material metaphors when modelled in clay on the furnaces or at the furnace foundations. The genderisation of the furnace was heard through the smelters explicit verbal metaphorical linkage between pots and wombs, pots and furnaces and their naming of furnace parts in terms of genitals. It was verbalised in songs sung in praise of the furnace as a young fertile woman or a bride. It was recognised in the spells and in the taboos, in the medicines and in the use of magic as combinations of male and female forces. The verbal use of metaphors was only one level of the symbolic repertoire used by the smelters. As I have argued, deeper rule metaphors in which the body was conceptualised as a container, but as open, made a deeper experienced foundation for the more explicit use of metaphors, including magic. The metaphors were not only verbal but also materially embodied, the furnace, being a body, was decorated as a woman and once decorated it was a woman ready to marry. The actual metaphorically based actions informed on what steps to follow in the process and on behaviour during the process. The metaphorical construction moved with the symbolic repertoire into a system where separation between different domains became important. Hence, Bekaert’s final level of codic oppositions between man and woman, hot and cold, clean or pure etc. formed as system based on both metaphorical imagination and a structural principle.

Since most studies of iron smelting have focused only on the smelting context the importance of similar taboos in other production contexts such as pottery making or house building, or in cooking, saltmaking or specific agricultural tasks, hunting or carving, has not been seriously taken into account. From a general perspective Damm has highlighted the fact that rituals consist of both unique and specific behaviour and that the contents in rituals do not always work by an inherent logic or meaning (Damm 1998:51). Initially the Kalulu ritual in Fipa iron working may seem as one of these ‘illogical elements’. In the midst of the celebration of the furnace as a young marriageable woman, kanawiinga, the dancing was interrupted when the master smelter announced that one of the smelters should go look for a hare, kalulu. One of the smelters ran into the bush to chase an imaginary hare. On his return he did not enter the site emptyhanded, but carried yellow flowers also called kalulu (cf. Willis 1978a). The smelters decorated the furnace top with these flowers. The flowers, or rather the grass from the Kalulu plant together with the Fipa fire drill sticks combine metaphorical imagination from different domains. Herbert has argued how hunting is linked to kings and ‘the immigrant stranger’ in many African oral histories (Herbert 1993:165). The link between them is made visible in the many artefacts that the smith supplies to both the king and the hunter, insignia of office in the one case and tools of the trade in the other (ibid.). If we start comparing the actual sequences in the smelting process, notions of the prey into hunter (Bloch 1992), and specifically hunting and hunting rituals seem to be included in iron smelting similar to the way wedding rituals and sexuality are. Herbert argues that both the hunter and the smelter ‘play with fire’ because killing animals releases dangerous forces. Hunters are associated with concepts of heat like smelters (Herbert 1993:177). They both operate to divide socially beneficial actions from sorcery. Hunter novices were often told to avoid sexual relations until they had killed their first animal and often women were accused of adultery after hunting accidents (ibid.). Hunters were often included in the smelt (Wyckaert 1914). The Fipa master smelter and master hunter were called umwaami or king, mweene. They were both magicians that confined their knowledge to their magical skills and the use of magical substances in their

During my first encounter with Fipa iron smelting I observed how both technological agency and material metaphors in the smelting context were structured in terms of dualistic thought and hence codic oppositions. Constantly I faced female exclusions because of their purity/impurity or hotness/coldness and simultaneously ‘real’ female inclusion because of their fertility and knowledge. The binary opposition between ‘head’ and ‘loins’ or the furnace as a ‘wife’ as opposed to the smelter as a ‘husband’ were both incompatible and compatible because it was the correct balance between the forces inherited in these categories that was important. Nevertheless, the day/night, sun/moon, east/west or 108

Iron technology and symbolic reservoirs fear of hotness or loss of energy, male and female powers, was equally stressed, as it was in the fields and vegetable gardens, during salt making, hunting expeditions, pottery making and during sexual intercourse and pregnancy. Fire was as a transformative agent and moral and sexual taboos could if not maintained, influence persons and objects coming in contact with fire or hotness. A familiar example is the idea that menstruating women are prohibited from cooking (Collett 1983; de Heusch 1980; Herbert 1993; Whyte 1990) and sickness caused by adultery is also believed to pass from one person to another by fire (Barndon 1992, 1996a, 1996b; Willis 1981). Hence, by fire a hot woman preparing food may transform her hotness to the food, which then becomes dangerous for those served the food. However, in everyday life these matters were not stressed, only practised as indicated in Bekaert’s level of typified experience and experiential gestalt.

intaangala baskets. A hunt was, like a smelt, impossible without the correct balance of substances and balances of people, something only the masters knew how to control and establish by the use of magic. Curing a seriously sick person was not possible without the combination of medicines and magic (cf. Willis 1978b). Hence, we see the crossing of domains. Davison argues that saltmaking, carried out by postfertile women and children may be interpreted in the same ideological context: as a transformation by fire of brine into salt (Davison 1993:39). In Malawian thinking he believes that salt is equated with, or takes on, the ‘heat’ of sexuality. If salt is consumed in food, especially in contexts where adultery has taken place or where a ritually dangerous condition is present, it may transmit illness (ibid.: 36). Davison explains how the correct handling of salt, not only in its making but also in its use in cooking, represents what was defined as normal moral behaviour especially between marriage partners (ibid). Stirnimann reports of similar sexual taboos mandated in saltmaking (from burning plants) in Upangwa (Stirnimann 1967:98). The Fipa, like the Pangwa, lack natural salt deposits but in Ufipa a form of substitute salt called iciluunga was made from burnt plants (Willis 1981:281.fn.68).

Above I have commented on the sexual taboos and the fear of ’hot’ people coming in touch with the charcoal, the ore, the tuyeres, the furnaces or the Pangwa and Fipa smelters. While this is generally seen as a central part of iron working in East Africa and as being based on fertility symbolism and a procreative paradigm, I have tried to move beyond this level and incorporate a more deeply experienced level of embodiment. I have argued that in Pangwa and Fipa technology the furnace womb was experienced as a container because the human body is experienced as a container. The metaphor ‘the body is a container’ was taken from one domain of experience to another. The pot is a container for food and the furnace chamber is a container for iron. In these containers, as inside the human body, substances were transformed and changed because although the body is within boundedness or a closed system it is also open and vulnerable. One may argue that this embodied experience moved between the first and final levels of Bekaert’s multiple realities in iron working.

My point is again that one has avoided seeing the totality or amalgam that living by social rules and a certain morality which sexual abstinence and fears of becoming ‘too hot’ exhibits. For instance, the Pangwa say that wood working of all sorts is men’s work. Therefore, men are the only ones allowed making fire with friction drills or allowed to make the wooden mortar and pestle that only women use for pounding grain. A woman in menstruation is not allowed to use her mortar nor to sit on it (Stirnimann 1976). Women in menstruation are not allowed to make fire or carry out charcoal burning. It makes sense that women are not allowed near places where wood is transformed into tools or into charcoal. These taboos or restrictions were rules people lived by and the fears of sickness caused by adultery (i.e. breaking the taboo) or sexual intercourse outside morally correct defined contexts were part of everyday life. One lived by the same rules and taboos in all spheres of life not only iron smelting. In most studies of African iron working the focus and explanations have been centred on the contexts as separated and divided. However, the participants in one context moved in and out of other contexts or were aware of the taboos and restrictions put upon the participants in the various contexts.

The balance or well-being of the human body as a container and as open was seen terms of the thermodynamics of substances or vital forces for living. The correct balance of hot and cold as in the state of blood, semen or breast milk was associated with anger or hotness/coldness caused by human sexual intercourse and the state of the body after committing adultery. These bodily experiences were transformed to the context of smelting iron but simultaneously the smelting context played back on how sexuality and the well-being of the body were experienced. One may therefore argue that the embodiment in technology ‘represents a return to the sensuous quality of lived experience and thus naturally bases itself largely on phenomenology’ (Lakoff and Johnson 1999:97; Strathern 1996:198). Still, this embodiment has to do with values that in some ways are also disembodied or may be thought of separately from the body itself, such as morality.

The fact that Pangwa ironworkers stressed the taboo against ’hot’ people several moths before the actual smelting explicitly illustrates that this ’mode of thought’ was not confined to specific days or weeks in the year or to the smelting contexts only. It illustrates that iron working and the associated rituals and taboos were part of everyday life. The smelters were not as we tend to se them practising an exotic way of making iron. During cooking and other production tasks sexual taboos and the

I have so far tried to demonstrate how material symbols and metaphors in technological activities take on meaning 109

Randi Barndon My description above has demonstrated a unity in conceptions of the body and furnaces as based on thermodynamics and well-being, in the use of medicines and magic and metaphors of sexuality and notions of morality. What does this tell us? It is evident that African material transforming technologies, such as iron working, pottery making, salt making, cooking and even rainmaking rituals were built on a model that used the human body as an omnipresent tool or anchor bridging the inexplicable with embodied experiences. Hence, perceptions and conceptions of the human body functioned to bridge between the known and unknown because transformative processes resemble each other and work back on each other. Not only was the human body used as a model, but also the human body was explained or further understood when other material objects or processes were used, such as the furnace or the pot. Thus, the different settings of technological human agency worked back on how the body itself was understood. The metaphor ‘the body is a container’ but also the body as open and vulnerable opens for a broader understanding of the duality in medicines and magic, gender, in hot and cold, clean and impure that was constantly played out in all spheres of life.

and enter into social discourse. This does not imply that the smelters were as Bekaert formulated it: ‘…symbolising their primitive minds’ (Bekaert 1998:7). The master smelters were ‘set apart’ or rather set above because they mastered the transformation of iron ore to iron tools by means of the ‘craft' and ‘ritual’, by means of making metaphors and using magic in numerous combinations of male and female substances and powers. In line with my previous discussion of magic and technology as one level of the embodiment in technology, the important points made by Mauss (1950) and later Lévi-Strauss (1964) have proved to still be of value. The magical operations appeared to the agent, the smelter, as having the same necessity as a sequence of natural causes as the technological (Rowlands and Warnier 1996:51). Most important is how conceptualising both ‘body’ and ‘furnace’ included a notion of morality. Rather than cosmology the smelting process was about morality and the well-being of the body and the furnace body. In line with Bekaert’s insight, the use of embodiment was played out at various levels moving from explicit metaphors to the binary oppositions (Bekaert 1998:17). As such the chaînes opératoires of the smelting process do not represent ‘something’, they are not social representations but incorporated parts or sub-sequences of the process (see figures 5.1. and 6.1).

Through and in line with this context the objects and participants of the place, producers and produced objects (blooms or hoes) were given a specific significance. Furnaces, bellows, tuyeres, and hoes, but also other assemblages of objects, such as pots and baskets are metaphorically linked. The smelting site and the forge were both representations of and structuring the activities and metaphorical imagination of the process. Rituals and taboos based on embodied conceptualisations of the technology reflected a socially structuring principle which order would be damaged if inappropriate behaviour such as for instance sexual intercourse were not maintained (cf. also Stirnimann 1976:283). This, the crew of smelters as well as intruders would easily recognise.

7.10 Summary My description of Pangwa and Fipa smelting processes has illustrated that the participants moved back and forth from one domain of experience to another. As participants in technological activity they moved in and out of metaphorical thinking combining the use of magic with practical technological operations. In sum ‘doing as one has to’ by ore processing (transforming a substance) and magic and metaphor making resulted in forgeable iron. The smelters were as I see it both aware of and unconscious of some of the symbolic steps they took and metaphors they made and used. The breasts decorated on the Pangwa furnace made the furnace into a woman able to deliver iron. This was an intentional action taken by the master smelters in order to give the furnace its identity. I will further argue that this gender identification of the furnace, of ores and of furnace parts created a social space that made it possible to create rules for correct moral behaviour.

I have tried to illustrate how, as Lemonnier argues ‘technology and technological systems are integrated into the bigger systems we call societies and become representations of these’ (Lemonnier 1992:9). But I have also attempted to move beyond symbolism in technology as mere representations because I have discussed how moral concerns were parts of the metaphorical boundedness and openness of bodies conceptualised and made meaningful in the smelting context. As MerleauPonty argued we inhabit and embody our worlds grounded in bodily experiences. The concepts of thermodynamics relates to the well-being of the body and folk models of morality. In iron smelting as conceptualised by Fipa and Pangwa we cannot deny the significance of the human body. Its physicality was as Merleau-Ponty saw it the anchor and as Johnson and Lakoff argue the base or foundation for metaphorical imagination and construction and was bridging domains of experience.

In Upangwa and Ufipa iron working provided a context in which technology and morality merged. It was therefore a context in which both materials (the material world) and metaphors (based on bodily experience) were constructed around moral imagination. It was a context in which technology cum technological agency created a ‘microcosm’ a social, ritual and technological space where iron was made and wherein metaphors about life and its moral content was staged (cf. Bourdieu 1973:104).

110

From ethnoarchaeology to archaeology, from recent history to prehistory

8. From ethnoarchaeology to archaeology, from recent history to prehistory The artefacts do speak (or faintly whisper) to us - the problem comes in the interpretation (Hodder 1986:123). the subject (i.e. Tilley 1994:12). This implies that phenomenological questions which are concerned with how we are subjects or persons ‘being-in-the world’ are given a place in our accounts (Langer 1989:32; MerleauPonty 1989). We need a ‘pre-modern’ or premetallurgical approach towards iron making in order to avoid the ‘Standard view’.

8.1 Operation, decoration and location Studying the prehistory of the African Iron Age includes an examination of relevant data from many fields such as ethnography, linguistics, traveller- and missionary reports, oral history, botanical and zoological studies, anthropology and archaeology (Kense 1983:13). But although archaeology often is considered synonymous with prehistory, it only comprises that body of techniques which relates to the recovery of material remains (Trigger 1968:3). Prehistory on the other hand, concerns the entire overview with which a particular problem is examined, and thus integrates information from several fields (Kense 1983:13). Therefore, if we perceive prehistory as such, ethnoarchaeology is but one small part of the complexity that writing prehistory is all about.157

Any direct technological relation between the African prehistory and recent iron working is now rejected among most archaeologists but there is a stronger wish to synchronise technical and symbolic aspects of the smelting process. This can however give the unwanted result that in comparative studies rituals and techniques become historically homogenised (see Barth 1993:10-11; Schmidt 1996a). The challenge is how to best describe and translate different expressions as they were through time without losing the historical and particular variation (Andah 1995:101). My approach towards technology has demonstrated that the complexity of Pangwa and Fipa smelting practices were combined by multiple levels from symbolic repertoires. These repertoires were based on metaphorical constructions and codic oppositions. This cannot uncritically be transferred to other cultural contexts.

Studies of pre-colonial and recent iron working have illustrated that technical agency involved in the smelting process was far more advanced than previously assumed. It was in all respects much more complex and varied (i.e. Childs 1991b; Childs and Killick 1993; Collett 1993; Herbert 1993; Kense 1983; Killick 1990; Schmidt 1981, 1997). This complexity is seen in descriptions of rituals and symbolism and how they worked at different levels and were assimilated or adjusted to existing symbolic repertoires (see Chapter 7.9). But, as Lemonnier argues ‘although archaeologists only recover a small part of these elements in prehistoric technological systems doesn’t lessen the fact that they are part of the general theory of material culture and technology’ (Lemonnier 1989:156). This refers to my own point, that more than anything our approach and concepts in studies of technology, past or present, should reject the long lasting ‘Standard view’. Hence, I have argued that we should reject our modern and false conceptualisation of technology as something de-humanised.

How may late colonial and recent pre-industrial iron smelting practices provide a source for understanding past societies (i.e. Misango 1996:3)? How may studies of preindustrial and recent African iron smelting focusing on symbolism be of relevance for archaeological theory building and a better understanding of human agency in iron smelting technology? As discussed in Chapter 7 iron smelting was a context in which technology, social and body control (morality) were merged into a magical and material manufacturing of raw materials. Iron smelting was a context in which humans and furnaces together formed a ‘moral stage’ and because of this iron smelting became a setting where metaphors about life and death, right and wrong, guided the chaînes opératoires of the process. How may the empirical complexity of Pangwa and Fipa iron working metaphorical constructions be of any relevance for archaeology? Is it possible to transcribe again what I observed in our re-enactments into categories that are relevant to compare beyond the reenactments?

I have argued for an approach towards the observations of technology or techniques in terms of bodily experience and Merlau-Ponty’s notion of the body as an anchor in terms of perceptions of technological transformations. I believe, as the ancient Greeks did that technology is indeed a process that is both poetic and bringing-forth and perhaps, at least for the actors in Pangwa and Fipa iron working activities, it is ‘where the truth happens’ to use Heidegger’s words (1977).

Throughout this work I have organised my descriptions or transcriptions of observed smelts in terms of Lemonnier’s chaîne opératoire approach. My object, in addition to descriptions of smelting practices, has been to contribute to our concepts being refined. Lemonnier’s approach is partly ignorant of issues that concern archaeologists occupied with producing answers to prehistoric

Our approach and transcription of technologies should start with how technological processes are experienced by 157 Among prehistorians concerned with the African Iron Age ethnographic related fields have played a crucial role and anthropology has been like the twin sister of archaeology (MacEachern 1995).

111

Randi Barndon the relational human and material components of the symbolic repertoires in East African iron smelting. Returning to my starting point, I will discuss these conceptual dimensions because I believe that they may be of use in bridging a recent historical and present context with prehistoric contexts.

technological practices. Knowledge of social choices in a technology as well as the symbolic repertoire that the technology was founded upon is essential but may be hard to grasp in contexts beyond the present. In Lemonnier’s approach there is little room for a material emphasis that is also relevant in contexts where the actors are missing. Lemonnier is not concerned with the spatially created location of technological agency in terms of where the technology was performed or the notions of specific location of technical activities necessary to reach a technological goal. Nor has he been concerned with the physical or mental decorations of the objects used in technological activities. Decoration is taken as within the other realm, as part of semantics, as stylistic and as representations or connotations of something other than the technology itself and how the practitioners conceived of a technological activity.

Operation In my introduction I considered furnace morphology and placed the Pangwa and Fipa furnaces within the existing classifications of African pre-industrial furnaces (cf. Appendix D, figure 1). Within my approach operation takes into account the form and functional aspects of the actual furnaces and the smelting sites. The operation of the technology and chaînes opératoires of the actors were linked to how they perceived and conceptualised the process in terms of human thermodynamics. Ritual behaviour, magical and specifically metaphorical thinking were part of the overall perception of the smelting process. Both metaphorical imagination and its particular focus on embodiment was the essence of technology. Indeed this embodiment and bodily thermodynamics affected many of the choices taken such as who could participate, where to conduct the smelt, which ores to use and final judgements of the blooms produced. The operational sequences were guided by a deep symbolic reservoir and conscious and unconscious movements of the body were performed within this baseline. Hnece, a Bantu philosophy of thermodynamics guided and structured the performances in the process and the way the furnaces were perceived as humans. This guided performances of preliminary magical acts and rituals as well as technical activities. In the Pangwa and Fipa cases the two most critical stages in the process (Childs stages 5 and 7) were highly ritualised. These stages in the process were also materially manifested because they were not only verbal metaphors but became material metaphors or a material manifestation of the philosophy of thermodynamics. The boundedness and openness of bodies and furnaces as containers not only determined what substances and forces to apply or avoid in the smelting process but also the localisation of the smelting activities.

I will argue that a re-definition of the approach towards iron working as composed of Lemonnier’s operational sequences as operation in combination with a focus on decoration and location may bridge some of the gaps between re-enactments of smelts in the present and past iron smelting practises. As conceptual dimensions these terms establish what variables were materially manifested. Although they cannot fully account for the actual meaning content and how meaning was structured or manipulated by the participants, they may provide us with better concepts in our search for how making iron and making meaning was connected, materially significant in making technological styles or expressing social identity. Thus, as a part of a shift in our conceptual approach these concepts might produce paths or provide inferences to be made about past or prehistoric technologies and their stylistic attributes. As an approach it may provide insight into which social and symbolic processes were at work and account for some aspects in the transmittal of technological knowledge. I have elsewhere described the ritual component in iron smelting as making the smelting context opposed to everyday and mundane activities because it was a seasonal activity (Barndon 1992). When the master smelter and his crew moved out of the village and installed themselves in the bush in order to make iron a specific ‘micro-cosmos’, a mental and material space was created. This was a specific space where rituals were played out and where metaphors were evoked and used (Barndon 1992; 1996b). This lead me into asking why embodied metaphors were played out in a setting that first and foremost was a technological activity. And why were the mental images based on moral and social order in everyday life?

Decoration I have already discussed at length the conscious and more deeply embodied experience in iron smelting in Chapter 7. In 1988 David, Sterner and Gavua asked ‘Why are pots decorated?’ More explicitly than ever before they addressed the fact that most researchers have taken decoration for granted and that decoration starts when operation (function) ends (David et. als. 1988:365). While most researchers into i.e. pottery styles were conducting more and more advanced analysis of its expressions, David’s initial statement was that material objects are decorated because the decoration in itself contains messages (ibid.). The authors argued that ‘pots

The concept technological style (Barndon 1992, 1996a; Childs 1991b; Lechtman 1976, Lechtman and Steinberg 1979) may be refined, perhaps if concepts such as ‘operation’, ‘location’ and ‘decoration’ are specifically used in our (pre-modern) approach we will better grasp

112

From ethnoarchaeology to archaeology, from recent history to prehistory consistent in a mental sense, it is also spatially embodied in the construction of the Berber architecture and specifically their houses (ibid.). Bourdieu argues that the representation of mental categories is embodied in two ways; by the material constructions people make (e.g. houses) and by the positioning of their own human bodies in relation to these constructions (ibid. see also Strathern 1996). Bourdieu’s notion of the habitus can be directly traced to Merleau-Ponty’s account of the body image as a set of dispositions of habitual actions (Merleau-Ponty 1989:98; Weiss 1996:6).

are persons’ and forwarded a line of thought in which ‘concepts of the body were closely related to and partly determinative of decorative expressions on pots (and sometimes other media also (ibid.: 365). Hence, more than a decade ago David argued that for the people of the Northern Mandara Highlands there was no ontological distinction between objects and humans because pots and people were perceived as one and the same thing. As argued by Lemonnier, among others, although the physical purpose of an object (often referred to as its function) and its communicative purpose (its style) are separate, they may be intricately linked during the process of manufacture for example when decorations are incised in the material during manufacturing (Lemonnier 1990:29. fn.4).

The interaction between the body and the house plays a key role in Bourdieu’s theory of practice (Carsten and Hugh-Jones 1996:2). In 1977 Bourdieu used the text metaphor in his explanation of Kabyle houses and argued that the house is like a book in which is inscribed a vision or metaphor of society and the world. Moving with ones body in the house, the body ‘reads’ the house though habit and by inhabiting the structural and spatial organisation of the house. To the Berber the interior structure of the house is both cosmological and social, it is dual and gendered like objects and tools are, as well as people when they use the house. Hence, activities and practices within the house (and beyond) add and construct meaning among the people in Kabylia (Bourdieu 1973, 1977). Bourdieu was among the first to incorporate structural analyses with what people do (Moore 1994:75). This is what makes his approach different from classical structuralism (ibid.: 76). It has made his approach relevant in many ethnographic archaeological studies because his emphasis is on the material and in particular how the symbolism of the material world guides and structures (cf. Preucel and Hodder 1996:309).

As a concept decoration considers both material decorations and mental decorations on and of the furnaces, the smelting site and the process itself. Hence, it depicts styles of a practice and styles of experiences. It takes into account the mental world, more specifically the embodiment in iron working. The furnaces had not always obvious materialised decorations, still they had a mental one. The furnaces were organised within a spatial code, and then the location informed practically and symbolically on the various dimensions as onto it self. The decoration assigned the furnace its identity, its gender (Herbert 1993), and the process became meaningful as the different dimensions (operation, decoration and location) influenced and imposed on each other. Location People both create and find their behaviour influenced by the built environment (Lawrence and Low 1990:454). Built forms are defined as different building types such as dwellings, temples or meeting houses, created by humans to shelter, define and protect activity. But as Lawrence and Low argue built forms also include spaces that are bounded and defined, not necessarily enclosed, and they may include landmarks and sites, such as shrines (ibid.). Spaces that are defined but which do not necessarily shelter or enclose activity are such built forms in the environment. The space where iron smelting took place was such a location or locale.

The locations of iron smelting sites have been commented upon in most descriptions of African iron working. Most observers have emphasised the importance of the site being outside villages or compounds and that the smelting activities should be in secret (e.g. Childs 1991b; Childs and Killick 1993; Herbert 1993). But values of separateness do not necessarily translate into Western understandings of remoteness (Schmidt 1997:191). In interviews the Fipa emphasised that their furnaces had to be out of sight of the villagers and thus away from villages, after surveys in the region it became clear precisely what Schmidt notes, that there are different notions about remoteness. Outside several villages or along the tracks leading to villages we Bourdieu’s discussion of the symbolism of the interior of observed smelting sites. Along the road towards the the Berber house has illustrated how embodied concepts historical and royal village Milansi three standing of space and metaphysical mappings are directly related furnaces, maluungu, were observed. Outside our crew’s to the layout of the house (Bourdieu 1973).158 Bourdieu village, several furnaces were also observed. One of these claimed that among the Kabyle knowledge is not only furnaces we used for the third smelt. When I confronted my informants with this they said that in the past a 158 Lawrence and Low and Carsten and Hugh-Jones discuss the great furnace did not necessarily have to be far away from the influence of Lévi Strauss and his ‘house societies’ (Carsten and Hugh- village as long as it was hidden and marked in some way Jones 1995:2; Lawrence and Low 1990), and how cosmologies determine or other telling people not to enter the site. the built environment (Lawrence and Low 1990:468-469). They value the influence of researchers based in Africa such as Griaule’s study of the Dogon, Kuper’s study of the Bantu homestead, and Turner’s influential discussion of the ritual process (ibid.).

113

Randi Barndon location where moral values are played out; the smelting and smithing localities were others. My argument is that the house and the smelting and forging spaces are settings in which activities are centred on the fireplace. The Fipa call the cooking place in the house, siiko. In the past they also used to call the open fire in the forge siiko. They call the secondary furnace ichinteengwe which means ‘the one who reveals the secret’ referring to the quality of the bloom (Barndon 1992). But when they explained what happened in the furnace they said it is like the cleansing and cooking in inteendo. It is not only the actors, their embodied experiences and their spatially gendered worlds but the locations or spaces themselves such as the fireplace for cooking or the ‘fireplace’ for smelting that form material and spatial components for metaphorical imagination.

Maret has suggested that the smelting site reflects not only practical or technological constraints but also ideological aspects of the iron working (Maret 1980:279)3. Herbert argues that a gendered structural and spatial ‘map’ is reflected in the use of and conception of space in African iron working (Herbert 1993:59, 96). In the previous chapter I discussed how the iron smelting was secret in order to control forces, male and female powers and substances in both the furnace and the smelters. The sites were ‘sealed’ and protected by charms such as the ‘furnace caretaker’ and ‘furnace guardian’, cockerels’ wing, white leaves, roots and sticks. The sites themselves inform not only of vicinity or distance to raw materials but also non-metallurgical factors (cf. Rodman 1992). At specific sites where the furnaces were built, forefathers and master smelters of the past were offered food and beer. A specific place was chosen because the master smelter could ask forefathers to assist with their experience and knowledge. Medical and magical substances that were part of the smelting process were placed in foundation trenches or buried in pots or bundles in the ground (Barndon 1992:75; 1999b; Mapunda 1995:228; Schmidt and Mapunda 1997; Schmidt 1997:250-251). In order to work the medicines and magic needed assistance from the supernatural. This defined the importance of embodiment but also what was beneath and above, what was there but extra corporeal. My focus on localities is concerned with the use of space within a technological context, but these technological contexts are also cultural. I will concentrate on the specific use of space in iron smelting because these localities or defined spatial areas were settings in which specific technological activities occurred and a host of social and moral issues were staged.

Embodied and material metaphors make it possible to move from one domain to another. Because of this I will postulate that the locations of both the smelting outside villages and the forging inside village boundaries were technological activities guided by the body as an anchor that were determined by and created specific spatial settings very much like the house. What is staged in the house if not cultural and social values? The house is a social space and a moral space. As I have demonstrated the smelting location was also a moral space because this setting as other daily activities were perceived through embodiment. It is commonly stated that it is important to focus on material evidence that would preserve through time and thus provide a link between studies of material culture in the present with the past. Can an ethnoarchaeological study of metaphorical imagination in technology operate as a mediator between the present and the material and the mental in past context? The multi-vocality of material culture is another challenge since no single view of meanings of things and their relationships with social categories and individual identities can ever be sufficient (Lane 1998:179). How then can we see metaphors in material culture? How may we grasp objects and places as materialised metaphors or parts of an embodied way of conceptualising technology?

I will argue that the smelting locations or the settings where iron working was carried out bear a resemblance to living spaces, compounds and houses although smelting locations differ in being narrower in the rage of tasks and activities performed within them. The most striking resemblance lies in the fact that, as within houses or living spaces, there is a fireplace or cooking place around which the activities at the site and in the house take place. Fire is the agent that is the central transformer both in the smelting furnaces and in the forge (e.g. Barndon 1992; Collett 1985; Eliade 1958). In my study of Fipa iron working I realised that in Fipa culture fire was seen as an ambivalent, powerful and dangerous agent. Fire or hotness was believed to transfer sickness, originating in fears of impurity and weakness in menstrual blood and adultery, insoni (Barndon 1992:190, cf. also Willis 1981). Through merely sitting close to a fireplace, preparing food or through eating cooked food prepared by an ‘unclean’ person, fire could indirectly transfer sickness or impurity from one to another person. Thus, fire has a central role in the wider concepts of accepted social practices in Fipa culture. The house is one setting or 159

But were any of these, mental and material decorations of the furnaces also spatially manifested? Were they seen in the actual location of the furnaces or in the use of space on the smelting sites? I have previously argued that a spatial code or dimension which may be conceptualise as a ‘technological style’ was a way of approaching African pre-industrialised iron smelting (Barndon 1992, 1996a). I also argued that the smelters by their way of ritualising the smelting context moved not only physically but also mentally out of everyday mundane life and into the bush where they created a ‘micro-cosmos’ (ibid.). Comparing my Fipa data with the Pangwa smelting practice illuminates the significance of magic and the fact that the master smelters were magicians applying medicines for

Cf. Barndon 1992:116; Childs 1991b:343.

114

From ethnoarchaeology to archaeology, from recent history to prehistory central for an understanding of the conservation and transformation and for the diversity within the unity of East African iron working practices.

the well-being of the furnace as a body. The role of the master smelters and the way they treated the furnace as a human informed me about the fact that rather than being about cosmology the smelting context was about morality. The activities in smelting may be seen as carried out on a ‘moral stage’ (after Beidelman 1980). This stage is where the masters through technical and magical skills reinforced their powers and reconfirmed their prestigious reputation as master smelters.

Johnson and Lakoff have argued that moral concepts are mostly metaphorical, based ultimately on our experience of well-being and family (Johnson and Lakoff 1999:557). Moral space is according to Beidelman a ’stage’ on which people enact and debate ideas about what is proper and good for one another (Beidelman 1993). The idea of the stage suggests both real playing and improvisations. Moral space was established at home, at the fireplace, in the fields and gardens. The smelting site was a secret place, a private space, like the stage created for the delivery of a baby out in the bush or in privacy in the house. All these locations or locales were places where the presence of persons connected to immoral behaviour such as adultery, killing or anger, or persons in certain states (e.g. women in menstruation or men ‘hot’ after sexual intercourse), affected the success or failure of the delivery of a healthy infant or of forgeable iron. Morality was based on ideas of a balance in the thermodynamics of the body. Sexual abstinence was necessary in order to maintain and secure the faithfulness of furnace ‘husbands’ because sexual intercourse during smelting would infer with the wanted balance of substances, vital forces, in the furnace womb. Securing the health of the furnaces by correct balance of temperatures and substances (ores, charcoal and magical substances) resulted in sacrifices and prayers to the ancestors and goods, the sprinkling of ritual water or blood and ritual washing and decorations. The master smelters always feared that the taboos were broken by some of the participants but were even more worried that sexually unclean persons would pass by the smelting site or otherwise come in touch with the smelting activities. Therefore sacrifices were conducted and ‘charms’ such as the furnace guardian, umboocha, and furnace caretaker, umpakasi, should protect the furnace bodies. It was because of the close metaphorical linkage between human bodies and furnaces as both being containers of substances (vital forces) thought necessary that the smelters remained ‘faithful’ to the furnace as long as it was producing iron. Because of this the smelting sites were located away from intruders.

In our re-enactments of smelting practices it was recognised how the furnaces were constructed in space in order to render meaningful the transformation of ores to iron. The building of furnaces were accompanied by bodily movements. The Fipa master smelters entered the front tuyere port, ‘the mother opening’, palinyina and their bodily movements resembled the sexual act. The Pangwa masters were not allowed to move in a full circle around the furnace trench or construct the furnace from the back. Once they were blowing the bellows the Pangwa mimicked the sexual act. The magical substances that the Pangwa and Fipa used were applied to the furnace as a body because of its containment and openness. Agency and metaphorical imagination during iron smelting were parts in their overall perception of the human body and its well-being in terms of thermodynamics. In sum then the technology was a human activity and the conceptualisation of the process were based on human concepts of body thermodynamics, hence the furnace (1) operated like a body (as a container) and was treated with magic and medicines very much like the human body, male or female; (2) was decorated as a woman. The conception of the operations occurring in the process gave the furnace its identity and its gender and the furnace became a woman. It was not like a woman it was a woman since it was decorated with female breasts and orifices, washed in ritual water or sprinkled with flour or wedding decorations; (3) was located in secrecy, out in the bush wherein the smelting location created a moral stage where sexual taboos, secrecy, separation and magic became central. The location of the smelting activities reinforced the centrality of the furnace being a body and a woman in a specific physical state; sexually hot and in delivery.

8.2 Some survey results Moral order and how to socialise and organise one’s world was played out in body-related metaphors and a desire of balance between elements or substances in order to secure a good health (cf. Strathern 1996:120). Because of these categorisations of persons and objects (furnaces) and a balance between the two oppositional states; hot or cold, of participants’ bodies and furnace bodies secured a good bloom. The dimensions operation, decoration and location affected and imposed on each other and might be seen as components in a physical and metaphorical crossing of domains.

These dimensions, or bodily based metaphors, have a certain materiality and they are both beyond time and time specific. In one way or another these three variables are all meaningful (express meanings) and they are archaeological relevant (produce material culture and deposits). Hence, they are stylistic. They may provide a model for the interpretation of the material evidence in iron smelting because in sum these dimensions connote a style and styles of chaînes opératoires. The role played by the master smelters as magicians and medicine men operating in and between these three dimensions may be

115

Randi Barndon

Plate 8.1 An old Fipa smelting furnace Western scientific metallurgical side of the smelting process behind. As I have argued, we may be walking in the wrong direction when we use more and more advanced metallurgical methods in order to understand a technology from a baseline from which the technology it self was never understood or operated within in the first place. In studies of African pre-industrial metallurgy we should turn our attention towards the overall human component in technological agency and how this relates to material culture. This implies studying the iron smelters and the culturally defined operations they performed in relation to the objects onto which they performed their actions (i.e. Latour 1999:122-123).

Unknown objects or artefacts, structures or monuments, have always been thought of as ‘ritual’ or ‘symbolic’ objects as opposed to those that were classified as of a more true practical and technological character (but cf. Lane 1986). Few archaeologists have ever looked for non-technological aspects or how body techniques and magic might be materially manifested, not only within one technological domain but several, such as for instance both iron smelting and pottery making. We have had a tendency to ignore the cultural context in which objects were produced and thus ignore the objects identity as cultural objects through their fabrication. On the contrary, like Strathern states if we look at material culture in this way it will also remind us of the viewpoint propounded by Mauss on the relation between persons and things. Mauss stated that things may themselves be seen as ‘agents’ and thus as ‘beings’, in a way comparable to the being of people’ (Strathern 1996:200). This is an essential observation because objects are then given a value through the context in which they are made, and of greater importance, may structure the making of objects or materials.

An argument for a certain uniformity in meaning (i.e. the embodiment in metaphorical imagination) has been recognised in the use of metaphors in iron smelting among the Pangwa and Fipa and may be linked to principles in the location, operation and decoration of the technology. Based on Pangwa and Fipa iron smelting practices I have found that these terms may grasp some of the materialised meanings or metaphors as part of the process. These aspects may form technological styles or culturally specific chaînes opératoires. It is sufficient to recall the most obvious use of material metaphors as they were seen in the visible anatomy of the furnaces being decorated with female attributes. But since complete furnace constructions normally do not last longer than maximum 80-100 years, we have to look for other traces of thoughts or metaphor making at the smelting sites. These are to be found at the furnace foundations and at the floors of the smelting sites (see Schmidt 1997). It has

According to Trigger ‘the most far-reaching contribution of post-processual archaeology has been Hodder’s (1982) irrefutable ethnoarchaeological demonstration that material culture inverts and distorts, as well as reflects, social organisation and hence plays a more active role in social processes that was hitherto believed’ (Trigger 1995:449). Although this is correct we still struggle with the great divide between subjects and objects. In this work I have decided to leave the Standard view with its 116

From ethnoarchaeology to archaeology, from recent history to prehistory been suggested that the organisation of space during smelting and the traces of activities that are left there by the smelters when they move to another site can yield information beyond the metallurgical and technological aspects (Barndon 1992, 1996a; Schmidt and Mapunda 1997; Schmidt 1997).

of slag heaps and number of tuyeres). In addition to the dominance of surface registrations, a few test pit excavations were carried out at some of these sites. These test pits were excavated together with my assistants from the University of Dar es Salaam and our Pangwa or Fipa informants.

During my fieldwork in Ufipa old abandoned smelting sites were located and the workspace was documented. In Upangwa in 1993 and 1995 old abandoned smelting camps were again surveyed. In both cases some of the smelting sites we located were known to my Pangwa and Fipa informants and some of the aged smelters had themselves participated in smelting at these camps. The smelters commented upon specific circumstances or events that had occurred there, but other sites were too old for them to recall who had once worked there.

I have previously presented the Fipa smelting workspace and shown that that my surveys combined by the reenactments pointed to a specific Fipa ‘technological style’ that also had its material spatial manifestations (Barndon 1992:115, 1996a). Although the Fipa smelting sites were outside village areas they were not hidden. The Ufipa plateau being almost treeless (which it has been for quite a long time), the three metres tall furnaces were visible from far away. They were most commonly located outside but close to villages, fields and gardens. Most smelting sites were a day’s walk from a village or less. In Ufipa the smelting sites were often seen to be located close to a termite mound. Most commonly the furnaces were built of the soil from these mounds and we were able to locate where the soil had been taken from. The natural draft furnace was seen to be located on the western side of the mound. At most sites we found that on the high shaft furnaces the major tuyere inlet, palinyina, faced towards the setting sun and west while the next largest tuyere inlet, palisi, faced towards east. At most sites we were able to

Surveys included documentation of the environment of the smelting sites, such as the vicinity to raw materials and water supply but also other significant factors such as the vicinity to other smelting furnaces, villages, ancestors’ graves or the spirits of the place. At each site that we surveyed the distance to clay sources or termite soils were documented as well as eating and resting places. The extension of use of a site was documented whenever possible (i.e. amount of potsherds, fireplaces or extension

Cooking and resting place

Ichinteengwe

Ichinteengwe Crushing Ore

Iron Charcoal

Clay for furnace Preparing clay Termite Mound Sorting Iron Slag

Slag

Iluungu

N

N Firewood

Iluungu

Clarcoal

Tegn.E.M.Hoff

Figure 8.1 Sketch plan of Fipa workspaces at the smelting site Tupa

Tegn.E.Hoff

Figure 8.2 Sketch plan of Fipa workspace at Tupa imagined as an ‘archaeological’ site

117

Randi Barndon

Plate 8.2 An old Pangwa furnace, liteende in a small valley. It was always close to a small river or creek.160 Water supply seems to have been one of the determinants for location of the smelting site as opposed to for instance ore that was transported from Manga and mixed with more local ores, also transported to the site. According to informants charcoal was transported from the vicinity to the site. The building material (clay or soil from termite mounds) was not always located at the sites we surveyed, hence, it must have been transported to the sites.

localise slag heaps and these were always outside the main tuyere inlet. The rake hole on the secondary refining furnace, ichinteengwe, also faced west, but at many sites the smaller secondary furnace was located at the other side of the termite mound while at other sites it was not located at all. My Fipa informants told me that in the past the secondary furnace was often destroyed after each smelting season. Often a new secondary furnace was built over the old one. The vicinity to raw materials was not consistent, some sites were close to ores while others were close to water supply. In previous discussions of Fipa iron smelting practices I have stated that the cosmology and differentiation between ’head’ and ‘loins’ as concepts was recognised on the smelting sites. The east/west and male/female codic opposition that was part of Fipa mode of thought was seen in the chaîne opératoires of the process and materially manifested in the workspace (Barndon 1992, 1996a).

The sites we surveyed indicate a specific Pangwa pattern or style in terms of choices in selecting a proper place for smelting.161 More important than vicinity to raw materials was the quality of the site itself. Sites were selected if ancestors or spirits were known to inhabit the place and if they were partly inaccessible and invisible for people passing by. Most sites were certainly seen from afar once the fire was lit and smoke came out of the furnace mouth but in Upangwa the sites were not actually along roads or tracks leading to fields, gardens or villages. Like I observed in Ufipa, the sites were always at most a day’s walking distance from the smelters’ villages but the smelters made provisional huts for sleeping and cooking while engaged in smelting activities.

Below I will, in a very preliminary manner, investigate whether the Pangwa smelting practice exhibits a materialised workspace that can be seen as related to their way of conceptualising the smelting process and whether it relates or not to the Fipa workspace. A Pangwa smelting site was traditionally located away from villages. Often it was quite far up in the mountains and according to our surveys always on a sloping hill or

160

Stirnimann reports the same importance of smelting sites being located close to water (Stirnimann 1969). 161 Recall Childs’ list of sequences in the smelting process discussed in Chapter 4.

118

From ethnoarchaeology to archaeology, from recent history to prehistory diameter. Additionally when possible to observe the position of tuyeres, rake hole for slag and slag deposits were localised (cf. Appendix D). The main tuyere port, umlomo, and opening for slag removal was on old furnaces as well as those made for our re-enactments, placed slightly sloping downwards and pointing towards a river or small creek. On surveyed sites it was possible to observe large slag heaps in front of the main tuyere inlet and slag removal port. In general these slag heaps were estimated to represent five or more smelts. However, as stated by Schmidt, archaeologists should not assume that slag heaps are pristine (Schmidt 1997:62). Slag is, as he has demonstrated, often recycled or re-used in ritual or domestic activities (Schmidt 1996b, see also Mapunda 1995; Schmidt and Mapunda 1997). We never observed that more than one furnace had been raised at exactly the same spot and most informants also denied that this ever happened in the past. At some sites two or three furnaces were observed within sight of each other, similar to my observations in Ufipa.

The use of space in Pangwa smelting was different from the Fipa way of spatially organising the smelting activities but their overall conceptions of the place as inhabited by ancestors and spirits that assisted the masters in their work was similar to the Fipa. The workspace within the smelting sites yielded a specific pattern or style although the use of space at the site was not as clear as in the Fipa case. The Pangwa, like the Fipa, physically made the area into ‘a built environment’ by marking the outer edges of the site with the furnace guardian, the poisonous umboocha. The area was cleared of bushes and the furnace was built in the centre of the site. At some of the sites we surveyed we found traces of where ores had been prepared and where the charcoal had been stored. In the past small huts were built outside the tuyere inlets, one for iron ore and one for charcoal. No signs or traces of roofs or poles were observed, but more intensive archaeological examinations of the sites might yield interesting results of constructions or buildings related to the site activities. An abundance of potsherds and charcoal was located in association with possible hearthstones. Normally sleeping huts (including cooking and resting) were observed approximately 100 to 200 metres from the actual smelting sites, something that fits Stirnimann’s statements (Stirnimann 1976: 273).

We observed several pots and potsherds in situ within the furnace pits at some of these sites. But at other sites there were no signs of the furnace pot, ing’anjo, beneath the furnace shaft. When we observed furnaces with empty pits the aged smelters said that some masters used to take the pot home after each smelting season. There was a contradiction in statements on these matters. The smelters of the Mtweve clan from Masimbwe/Mlangali claimed the

Measurements of the remains of the furnace heights, circumferences and wall thickness were recorded. Tuyere ports were localised and measured, both length and

UPANGWA, ING´ANGITOLI SMELTING SITE

DENSE FOREST N

ARRINING AT THE SITE FROM THIS DIRECTION CLAY SOURCE

1m

FIREPLACE PREPARING OIL, MONO

JERRY CAN OF WATER ORE UMDAPO

RESTING PLACE

KILULEPO MAKING LIDOPE

RE

EK

KWIBEPO

SM AL LC

SLAG

0 TUYERE INLET KILULEPO MAKING TUYERES AND DRUMBELLOWS

UMLOMO

DENSE FOREST Tegn.E.Hoff

Figure 8.3 Sketch plan of a Pangwa smelting site as observed in a re-enactment at Ing’angitoli

119

Randi Barndon 1995, another style emerged than the one observed in Ufipa. Although the technical processes or the transformations were similar in resulting in bloomery iron and an overall similar Bantu philosophy of thermodynamics was recognised in both cases, an altogether different spatial and stylistic picture was seen. Most important in the Pangwa case was the location of smelting sites close to a river with the furnace main tuyere inlet or ‘mother opening’, umlomo, and slag removal facing towards the river. In the Fipa case the ‘mother opening’, palinyina, faced the setting sun and west but nor rivers or creeks.

pot was never removed once a furnace was ready built. On the other hand the Sanga master (from Magasiwalale) said his father had always taken the pot with him home at the end of each smelting season. If a smelt failed one of the reasons could be that the sacrifices had not pleased the ancestors and the pot would be removed. Perhaps the forefathers were not pleased with the food, such as beans and flour as well as beer that was presented to them and the location was abandoned. Hence, if a furnace could not produce forgeable iron, the furnace was destroyed, the pot removed and the smelting site abandoned. Iron smelting as an approach that values operation, decoration and location as a combined pattern or style is best seen in the furnaces themselves. Of greater archaeological relevance is however the fact that this style and the smelters activities are structured spatially in accordance with their way of conceptualising the smelting process. The workspaces as documented in Ufipa and Upangwa have few similarities. Notwithstanding, since I had some control over ‘the dynamics’ (Binford 1988) in the field it was easy to identify how the work and the sites in Ufipa were spatially structured in accordance with both practical and symbolic dimensions of the iron smelting. This ‘expressive space’ (Huffman 1984) structured not only thought but also action.

In both the Pangwa and in the Fipa case the furnaces operated like humans. The furnaces were decorated to become women and were located in secrecy as if they were real women in delivery. But the actual workspace was different. In Fipa culture the concepts ‘head’ and ‘loins’ as opposed to each other guided and structured the workspace. In Pangwa this set of codic oppositions were not seen as an elaborated part of their conceptions of the smelting process. In Pangwa thought the purity and coolness in bodily imagination was seen and may be related to their location of furnaces close to rivers and the main tuyere inlet facing the water. This may be seen as connected to the Pangwa’s emphasis on ritual washing in kiwalasha before all sorts of activities (cf. figure 5.1.).

When I surveyed smelting sites in Upangwa as well as during the re-enactment of smelts conducted in 1993 and

UPANGWA, ING´ANGITOLI SMELTING SITE

DENSE FOREST N

ARRIVING AT THE SITE FROM THIS DIRECTION

0

SLAG

SM AL LC

RE

EK

1m

DENSE FOREST Tegn.E.Hoff

Figure 8.4 Sketch plan of the Ing’angitoli smelting site imagined as an ‘archaeological’ site

120

From ethnoarchaeology to archaeology, from recent history to prehistory reflect higher-level systems of meaning within the society (Lemonnier 1992; Speth 1992). This was, preliminarily explored above. Lecthman and Steinberg were early in their criticism of archaeologists not paying enough attention to technological systems and style (Lecthman and Steinberg 1979). They also criticised archaeologists for not paying enough attentions to cultural aspects of technology during excavations. Lemonnier has criticised the ‘Excesses of the Cambridge school’, especially Hodder and students, for limiting artefacts to symbolic and structural studies in their data and theories within archaeology (Lemonnier 1992:96-97). My point here is that we must reject an approach or attitude towards technology that ascribes to culture much the same dominant role that ecological determinists once ascribed to environment (Trigger 1995:450). Both sides are equally important. Lemonnier proposed an ‘anthropology of technology’ in which he argues that ‘the study of a living tradition, of a living material culture provides a unique opportunity to investigate many aspects of technologies that lie behind archaeological remains’ (Lemonnier 1992:98). H stated that ‘the more technological relations are dealt with in the observations and analysis of a living material culture, the better our chances of constructing an accurate interpretation of archaeological remains’ (ibid.).

8.3 Summary As argued for in my introduction, comparisons lie at the very base of anthropological research simply because we have to compare in order to study the differences on which we comment. If our concern is the past or cultural processes that might have happened in the past we have to compare relics from the past with the present. This is the most important aspect in Lemonnier’s urge for more elaborate ethnographic recordings of chaînes opératoires in technological behaviour. Lemonnier and Sigaut have stated that technologies express systemic order and that these systems can be studied (Lemonnier 1992; Sigaut 1990, 1994). At a time when actual knowledge of pre-industrial iron smelting is about to disappear both metallurgical and nonmetallurgical aspects should if possible be taken into account. What has been lacking in most of the suggested paths and attitudes towards ‘technology’ is magical knowledge and treatments of the furnaces and ores as integrated parts of the chaînes opératoires and hence, integrated in the use of space on smelting sites. What makes the study of African iron making technologies specifically relevant to archaeology is that there is a clearcut mark between physical constraints and higher levels of meaning, yet always a dialogue between them. This dialogue, between physical constraints, social opportunities and metaphorical construction is essential for more holistic understanding of technology. In ethnographic studies acknowledging this bond allows the researcher to move between functionalist and symbolist sets of aspects and regard them both as equally relevant.

As argued above, and narrowing down my arguments to the actual practice of doing ethnoarchaeology and making it relevant, I believe the space between the dimensions, operation, decoration and location is where archaeologists concerned with East African iron working should try to document, describe and explore the material manifestations of iron working. In this chapter I have tried to imagine pre-industrial iron working with a conceptual approach that values technological agency from the point of view of the subject. I have used a proposed Bantu philosophy of thermodynamics and extracted three dimensions that informed and structured the smelting of iron. Operation, decoration and location in iron smelting were intrinsically linked. Finally, I have discussed how Fipa and Pangwa smelting sites were part and parcel of this philosophy of thermodynamics and I have discussed traits from the workspace observed on Fipa and Pangwa smelting sites.

Heidegger has suggested five aspects to be the characteristics of structuring in technologies; techniques (tools, implements, apparatus and machines) products (consumer and non consumer), nature (material and power), theory (the role of science) and inter subjectivity (the social organisation of labour). For Heidegger any account of technology must include a discussion of all these five essential features (Hood 1983:354). Research on technology has illustrated that technologies involve many arbitrary decisions and choices that are not dictated by immediate physical constraints. Instead many of them

121

Randi Barndon

122

Masters of metallurgy, masters of metaphors

9. Masters of metallurgy, masters of metaphors The perception of African iron technology as backward compared to that of Europe is an artefact of Western thought (Schmidt 1996:xv). symbols in the smelting process. Although old ethnographic accounts are valuable to modern anthropology and especially ethnoarchaeology they must be considered as all science; a product of their time.

9.1 Unity within diversity Livingstone was impressed by African iron production and commented on its abundance in Ufipa by stating that he could not walk without tuyeres, slag and abandoned kilns at his feet (Livingstone 1874). The Fipa master smelters been famous ironworkers ever since, but from a local viewpoint, the masters in Ukinga, and although to a lesser extent, the Lungu, the Pangwa, the Nyiha and the Tabwa master smelters, were also known. All of them were lived within this restricted area of the south circumTanganyika region and they practiced iron working at the same time. For an Africanist archaeologists the Fipa are among the best known of the Tanzanian peoples of the interior due precisely to their iron working tradition which has been paid a great deal of attention from the time of Livingstone’s early comments. Ethnographically better known are their neighbours to the north, the Nyamwezi (Robert 1968), to the south the Nyakyusa (Ngonde) (Wilson 1958) and to the west, the Bemba (Richards 1939, 1982 [1956]). Although the Fipa may have been famous ironworkers in pre-colonial and colonial times in the interior of Africa, they were certainly not the only skilled ironworkers in this region. Brock and Brock mention how the Nyiha masters were famous for their iron smelting skills (Brock and Brock 1965), and Kootz-Kretschmer confirms how the Nyiha were traders in iron with the Safwa (Kootz-Kretschmer 1926:180, see also Cline 1937:25). Chaplin stated the fame of the Lungu (Chaplin 1961), while Wilson focused on the Kinga as the famous ironworkers of the region (Wilson 1958). Stirnimann confirms Wilson’s observations of the Kinga and claims that they were the supreme smelters in the Southern Highlands (Stirnimann 1979). While Stirnimann illustrates that the Pangwa practised an identical technology to the Kinga he does not question the Pangwa’s lack of fame or relation to the Kinga master smelters (ibid.).

Ethnographic and historical information is scarce on the dynamics and interrelationship between the various iron working communities in the southern parts of the interior and in particular the master smelters. This is interesting to note, especially since the region has been one of the most frequently visited areas, in terms of both explorers and ethnographers. During the years before smelting was abandoned only the Fipa, Nyakyusa and Pangwa were studied at length by ethnographers or social anthropologists. This has resulted in somewhat biased views of the relations between the peoples of the interior. The ethnographers with the best insight into the cultures and languages of the interior have not given much attention to their material culture and been almost ignorant of their regional relations. It is difficult to establish the relation between master smelters of one group to those of another group. None observed the entire smelting processes among their hosts (the Fipa, Pangwa or Nyakyusa), or their neighbours. However, they did reconstruct the smelting processes based on informant’s statements and visited old abandoned smelting sites. In my introduction I quoted Wilson who described the area which the Pangwa and Fipa today inhabit as the very heart of Africa (Wilson 1958:1). From an archaeological perspective the areas between Lake Rukwa and Lake Malawi are important in regard to the early migrations of Bantu-speaking peoples (Clark 1974; Willis 1981:3).162 Clark stated that the south Lake Tanganyika region must have played a significant role almost as a highway for people passing from the north to the south and from the interior to the coast (Clark 1974:3). Within the area numerous archaeologists and ethnographers have conducted studies of the people of the southern lake region and their pre-industrial iron working practices. However, evidence for any Early Iron Age populations in the region is scarce. Kalambo Falls on the Zambia/Tanzanian border in-between Ulungu and Ufipa excavated by Clark still stands almost isolated within the region (ibid. and cf. also Derricourt 1985).163 This may

Social anthropology has been occupied with social and cultural aspects of societies. The detailed ethnographic accounts that included descriptions of the manufacture and use of material culture were only included in monographs in the last decades of the eighteenth and first decades of the nineteenth centuries (cf. Eriksen 1995; Keesing 1987; Pfaffenberger 1988, 1992; Trigger 1989). Since then material culture studies have been considered peripheral (Pfaffenberger 1992). Related to the general theoretical perspectives and philosophical considerations of the time, cultural and social aspects of iron smelting were treated as ‘ethnographica exotica’ (e.g. Lanning 1954; Wyckaert 1914). In most of these early accounts we find vivid and detailed descriptions of the rituals and

162 Clark has argued that iron working cultivators lived in the Kalambo Falls area from about A.D. 300 (Clark 1974:57). He has suggested, although vaguely, a cultural continuity between the A.D 300 people who inhabited Kalambo Falls and the present Lungu and Fipa people (ibid.). This view has continuously been referred to in ethnographic work from the region (e.g. Stirnimann 1976, 1979; Tew 1954; Willis 1964, 1981; Wilson 1958). 163 A distinct technology from the one I have described among the Fipa of the plateau has been observed by Mapunda (Mapunda 1995).

123

Randi Barndon partly be explained by the lack of archaeological surveys in this part of the interior.164 Apart from Mapunda’s (1995) excavations along Lake Tanganyika which has established that there were iron using people in the interior practising a different smelting technology prior to the Fipa, no systematic surveys of iron smelting furnaces on the plateau have ever been conducted. A broader understanding of the first iron using communities on the Ufipa plateau, as well as those of the Livingstone and Kipengere Mountains, must await further research. The Pangwa and Fipa who live in their respective areas today cannot be directly associated with material remains of past inhabitants of these regions, nor should they be. Based on historical and ethnographic studies (written sources and genealogical studies) they have been recorded as inhabitants of the area from the 15-17th centuries (e.g. Kjekshus 1977; Koponen 1988; Stirnimann 1976, 1979; Willis 1981). If our concern is the history of the late colonial peoples, the Rukwa region, with the historical village Milansi and several royal courts localised on the plateau seems especially promising for these purposes. In Upangwa the area surrounding the Mdapo mines and Manga village would be another interesting area to investigate. In the first half of this century African Iron Age research was coloured by discussions of the origins of iron technology and of the possible role of Egypt in the advent of iron in the Old World (Kense 1985; Merwe 1980; Miller and Merwe 1994). However, it eventually became clear that Egypt was rather late in its development of an Iron Age compared with the rest of the Old World and not the initiator. Egyptologists continued to claim that Ancient Egypt had al least served as the primary channel through which the knowledge of iron metallurgy had been transmitted to southern Africa and this Egyptian preeminence over sub-Saharan Africa was accepted without questions (Kense 1985; Wainwright 1942). Egypt is central because discussions of the African Iron Age divide the continent into two distinct regions with possibly independent metallurgical histories. The northern areas consisting of the countries of the Mediterranean basin, the Nile Valley and the Red Sea coast contain areas which were participants in the rise of the Bronze Age civilisations, most notably Egypt (Merwe 1980:463). Unlike Europe and most other parts of the Old World, Africa never went through a true Bronze Age. Thus, except in Egypt and parts of northern Africa, neither a distinct Bronze Age nor a true Copper Age occurred (Childs and Killick 1993:320; Merwe 1980:463; Phillipson 1993:158). According to Schmidt and Childs this might have led to different metallurgical histories with experimental modes that are distinctive to Africa (Schmidt and Childs 1995:524), and perhaps equally Mapunda suggests that the Lake technology with the kataruka furnaces are of an earlier date than the plateau technology (ibid.). 164 Cf. Phillipson 1988:191 for distributions of early finds.

124

interesting; might have given the iron objects and their producers a different symbolic and social value. This makes the history of iron using and iron producing communities within this region different from those that had been included within ‘the Old World’ (ibid). In spite of the variety in African furnaces (both prehistoric and ethnographic) it is not possible to postulate with certainty any of the routes of diffusion for metallurgical knowledge into Africa south of Sahara (Herbert 1993:9, and cf. Appendix D, figure 2). It becomes even more problematic if one attempts to establish the different routes of diffusion within the interior. Early dates from the northern Interlacustrine area indicate that this region was one of the areas from where the iron working knowledge spread southwards, along the coast and into the Tanganyika-Nyasa Corridor. Through time the mosaic of the interior also influenced the interior itself by the movements of people and in particular of master smelters. It is likely that people from the interior, from Central Africa, migrated into the Rukwa and Ludewa areas bringing with them a developed iron technology that encountered the movements of people from the south during the various interior turmoils such as the Ngoni invasions. This led to the smelting practices that we can partly grasp in re-enactments today were developed. Along with the early research on origins of metallurgy came the various and detailed ethnographic descriptions of smelting practices and among the most famous and still often cited volume is that of Walter Cline (Cline 1937).165 Apart from the detailed regional descriptions of early travellers and missionaries such as for instance Fülleborn (1914) or Wyckaert (1914) more recent research which includes ethnographic descriptions still has some references to Cline’s ethnographic overview (Haaland and Shinnie 1985; Herbert 1993; Kense 1983; Miller and Merwe 1994; Shinnie 1971). However, like Willis and Stirnimann, few of the numerous travellers and explorers that went through the interior ever actually observed the entire smelting process although they all state that furnaces were abundant (Bornhardt 1900; Burton 1961 [1860]; Fülleborn 1906; Kerr-Cross 1890; Livingstone 1874). Even fewer made any comments about the forge or whether the iron smelters were trading their iron to any of the other groups in the region.166 The picture we have of the internal relation between iron smelters and blacksmiths among the different groups in the region is therefore at the moment lacking in detail. Keeping my description of our re-enactments in Upangwa and Ufipa in mind and the above discussion of 165

Pole correctly notes that Cline’s review from the 1930’s provided very little in the way of a framework for a discussion of smelting practices because of his classification of furnaces merely into large and small ones without references to choice of criterion (Pole 1985:142). 166 But cf. Kootz-Kretschmer who gives an elaborate desription of Safwa blacksmithing and trade with the Nyiha (Kootz-Kretschmer 1926).

Masters of metallurgy, masters of metaphors In work mainly concentrated towards the west of Ufipa and Upangwa, Derricourt documents enormous variations in furnace design and furnace operation as well as variations within the symbolic repertoire used during smelting (Derricourt 1985). In this work Derricourt suggests that the Lungu belong within the Mwika complex (ibid.: 103). On the other hand, the secondary refining furnace and concepts (linguistic terms and metaphors) connected to this furnace are part of the Southern Highland smelting traditions. Hence, the Pangwa and their neighbours combine terms for the smelting practices with the northern tradition of the Lungu and Fipa peoples (cf. Appendix D, table 10). Within the northern Rukuru basin in Northern Malawi great variations in furnace operation have also been documented (Davidson and Mosley 1988; Childs 1991b). Similar variations are confined within smaller regions such as between the Chewa, Chulu and Phoka in Malawi (Killick 1987, 1990; Phillipson 1968; van der Merwe and Avery 1987).

embodiment in smelting symbolism I will discuss some of the reasons for the unity and diversity as it has been observed expressed between the different groups. I will extract aspects of the philosophy of thermodynamics as seen in iron smelting among the Pangwa and Fipa and connect this to the various neighbouring groups in the interior. I will discuss variations and similarities within the broader region to which both the Fipa and Pangwa belong. As I see it the master smelters used ‘a core vocabulary’ or ‘a symbolic repertoire’ that was based on a deeper ‘symbolic reservoir’ of bodily metaphorical constructions that developed through time and by the interaction among peoples and their ways of doing, learning and experiencing technologies. In the following discussion of the interior mosaic, the area between the lakes Tanganyika, Rukwa and Nyasa constitutes my geographical baseline. I will take as a starting point Wilson’s analyses of the people that from her point of view surrounded the Nyakyusa and Ngonde peoples towards the end of the 1950’s (Wilson 1958). The same clusters of peoples that she then recognised live in the region today, while the villages have moved and towns have all expanded and become more ethnically complex and international.167 I will concentrate on those peoples within the region that are specifically relevant in view of iron smelting and iron smelting symbolism (cf. Appendix D, table1). Hence, my conceptual frame or attitude towards technology is the platform for a discussion of the iron smelting practices as seen among a selected group of peoples within the broader, interior corridor region.

I will argue that if iron working is the focus, the Fipa are among the Pangwa’s distant neighbours in the Tanganyika-Nyasa Corridor region. They share among many of the peoples in this broader region lexical terms, rituals and practices, social significance of the smelting process as well as the social status of the master smelters. The Lungu and the Bemba The immediate neighbours southwest of the Fipa are the Lungu people whose closest neighbours are the Tabwa, the Mambwe and the Bemba. The derivation of the name lungu is obscure, the term iluungu refers to iron smelting furnaces in Lungu language as well as in the Fipa, Mambwe and Nyiha languages (Willis 1966:40).168 Clark observed old Lungu furnaces in Mkamba village on the Tanzania side (Clark 1974:5). The furnaces Clark surveyed are according to figures in his publication identical to the Fipa’s tall natural furnaces and secondary refining furnaces that I have surveyed observed in use. Chaplin observed Lungu iron smelting in the early 1960’s when a group of smelters under the surveillance of a former master smelter demonstrated their skills outside the Chibote Mission Station. Chaplin noted that the rituals in iron smelting that he observed among the Lungu were less significant than those recorded in Ufipa and among the Bemba, who had practised iron smelting in similar tall natural draft furnaces (Chaplin 1961:54). Chaplin suggested that the reason behind the reduces use of rituals and medicines in Lungu smelting could be explained by the fact that the re-enactment had been carried out in a public setting at the mission station (ibid.: 53-54).

From about at least 1890 the peoples immediately surrounding the Fipa are the Lungu, the Nyamwanga, the Lambya, the Nyiha and the Bungu and to the north the Pimbwe (Koponen 1988:16-17). Surrounding the Pangwa were and are the Kisi, the Matengo, the Kinga and the Bena and further south the Ngoni (Koponen 1988; Stannus 1914; Tew 1950). If iron working is the focus the Bena, the Kinga and the Matengo relate to the Pangwa (Culwick 1935; Culwick and Culwick 1967; Koponen 1988:261; Stirnimann 1979; Sutton 1985), while the Nyiha (Brock and Brock 1965), the Mambwe (Clark 1974; Koponen 1988; Willis 1964) and the Lungu (Clark 1974:5; Chaplin 1961; Willis 1964) relate to the Fipa. The Tabwa in eastern Zaire/northern Zambia with their connections and towards Uluba and the Luba kingdoms (Kriger 1999; Reefe 1981; Mapunda 1995; Willis 1981) are also important. The Chewa and the Phoka smelting practices towards the south, in Malawi, adds to the complexity since these smelting practices contain aspects of both the natural draft furnace process and the smaller anthropomorphic furnaces of the Pangwa and Kinga (Phillipson 1968; van der Merwe and Avery 1987 and see figure 1.4), and the Fipa. 167

Chaplin describes how the Kaonde furnaces always faced east, as did the Lungu furnace for his observed demonstration smelt (ibid.: 54-55). He is however not certain if this applies to the ruins of furnaces that he 168 Traditions recorded among the Lungu suggest that some of the Lungu are of Bemba and ultimately of Luba origin (Willis 1966:40).

Cf. Koponen 1988:16.

125

Randi Barndon observed in the regions. The Fipa furnaces, according to my field observations, always faced west (Barndon 1992; Mapunda 1995). Another difference between the two practices was seen in the way in which the furnaces were charged. In both the Bemba and Fipa methods, alternate layers of charcoal and iron ore were packed in the furnace (Barndon 1992; Wembah-Rashid 1972; Wise 1958a; Wyckaert 1913). The Ushi following the Lungu had only a small nest of ore at the top of the charcoal filled furnace (Chaplin 1961), while the Nyiha added charcoal at intervals during the entire smelt. As noted by Chaplin a Bemba furnace was driven by natural draft but the tuyeres were inserted separately and directly into the furnace walls (Clark 1961:56). Further up the chimney were ventilation holes placed to make the draft (ibid.; Willis 1966). The natural draft furnaces of the Bemba were similar to the Lungu and Fipa’s tall furnaces. Richards notes that for the Bemba ‘the dogma relating sex and fire is the idée mâitresse behind most ritual behaviour’ (Richards 1982[1956]: 30; Herbert 1993:118). Herbert explains that the Bemba say sexual activity makes a couple ‘hot’; a state in which it would be perilous to approach the ancestral spirits in any prayer or sacrifice (ibid.). However, as for the Chewa, Herbert continues, this has permutations throughout Bemba social life (ibid.). Brelsford described how the Chisinga, a Bemba subgroup in Zambia, explained the importance of sexual abstinence during smelting (Brelsford 1949). Again the notion of adultery is recognised. The furnace was seen as the smelter’s wife and to sleep with his human wife meant adultery. To commit adultery when the wife is pregnant means that the child will die. Using this analogy the furnace would not produce good iron (ibid.: 48). The Chisinga iron smelting practice, gender symbols and sexual abstinence because of a fear of adultery causing misfortune resembles the other cases described above. The linguistic terms used on the furnace and the process are similar. The Chisinga smelters were so afraid of adulterous behaviour that they would not let women cook for them and had young boys to do the cooking (ibid.: 28; Herbert 1993:79). Other Bemba furnaces differed from the Lungu and Fipa tall smelting furnace and had several small openings evenly distributed around the top or mouth of the furnace in order to assist the air draft (Barndon 1992:155; Chaplin 1961:55). The Nyiha The Nyiha are by linguists not associated with the M10 group of which the Fipa and Lungu belong, but the M20 group constituting the Wanda, Mwanga, Malila, Safwa, Iwa and Tembo (Felberg 1999). However, the concepts used in the smelting technology are the same as in the Fipa vocabulary (cf. Barndon 1992 and cf. Appendix B). The Nyiha in the Mbeya district of Tanzania have because of similarities in smelting technology, been closely linked to the Fipa (Clark 1974; Barndon 1992;

126

Brock 1968; Mapunda 1995). Clark suggested that the Nyiha and the Kinga were the initial Bantu immigrants into the Corridor prior to the fifteenth century, and were followed by several subsequent groups (Clark 1974:3). Kerr-Cross who was a missionary for the Free Church of Scotland in the area between the three lakes, Nyasa, Tanganyika and Rukwa in the 1880’s, commented upon Nyiha iron working. He passed through Nyiha country and wrote ‘iron abounds and is smelted and extensively wrought into spears, hoes and axes’ (Kerr-Cross 1890:289). And further, ‘the industry and skill of the people of Unyika greatly impressed me, traces of ironstone are found everywhere, and, in places old workings are seen. On one hillside I counted five smelting kilns, standing in the bush not many hundred yards from each other. These stand full nine feet high, five feet in diameter at the base, and three at the top and are built of reeds and clay. They will contain half a ton of iron’ (ibid.). According to Brock and Brock the Nyiha stopped iron production by the beginning of this century primarily because of competition with mass produced goods, and also possibly because of governmental repression (Brock and Brock 1965). In another of Brock’s studies it is indicated that there have been at least three different furnaces in use in Unyiha (Brock 1963). This has also been commented upon by Mapunda who has documented two different furnaces in the Lake Tanganyika area of Ufipa (Mapunda 1995:101). However, in the Nyiha case their temporal sequence is unknown. The first was described by an unknown writer in 1938 and recorded in Mbeya District Book. This furnace was 3 metres in height, 1.5 metres in diameter at the base and 70 cm in diameter at the top. The furnace was built of red clay and the wall thickness was uniform, about 12.5 cm (Brock and Brock 1965:98). This fits well with the dimensions of a ‘standard’ Fipa natural draft furnace (cf. Barndon 1992:83 and cf. Appendix D) as do the description and position of the tuyeres in the same publication. In spite of some discrepancies and uncertainties in the documentation by Brock and Brock there is a close relation with the Fipa practice. As Brock and Brock note: ‘The furnace was lined with short pieces of wood and charcoal up to about chest height, and the centre filled with iron ore. Charcoal was added to fill the kiln to the top. Fire was lit on one side; the four holes at the bottom were closed and the top left open. Charcoal was added at the top for two days. When the fire was going well, the holes were opened to increase ventilation. The whole was allowed to burn for four days, and the iron taken out on the fifth. The larger openings were filled with pipes and the slag allowed flowing out. Then the pipes were removed and the iron taken out’ (Brock and Brock 1965:98).

Masters of metallurgy, masters of metaphors Brock and Brock and more recently Collett suggest that the second refining furnace was only used when iron of especially good quality was desired (ibid.; Collett 1983). However, the Brocks admit that the account of the second refining was not very clear (Brock and Brock 1965). Nevertheless, the tall natural draft furnace was called ilungu while the secondary refining furnace was called ishitengwi (ibid., Brock 1963). Unfortunately no explanations for the terms are suggested.

During the smelting season the master and his assistant lived out in the bush and they were not allowed to wash or cut their hair before they returned to the village. Moreover, they were not allowed sexual intercourse. Women were allowed to bring food, but could not stay at the smelting site. If women participated iron would not come out of the furnace and because of this women were not allowed to come near the medicines contained in a medical basket, called intangara (ibid.: 99).

The second furnace, which was of a different type, was according to Brock observed in Unyiha by Dantz in 1899 who reports smelting furnaces called ilungu. One furnace had been discovered northeast of Vwawa town. This furnace was similar to the tall natural draft furnace but had only eight tuyere ports at the base.169 In addition to ore it was charged with charcoal and grass (Fülleborn 1906:168). Fülleborn notes that a smelt took from one to one and a half days and that this furnace could be in use for up to fifteen smelts before it was abandoned (ibid.: 168).170 Another furnace called ulungu by the Nyiha had only four tuyere ports at the base (ibid.; Brock and Brock 1965:98), rather than eight or ten as in the Fipa case. Information about this furnace is based solely on informants’ statements and there are several discrepancies in the technological descriptions in the text provided by the Brocks (ibid.). The informants said they used to collect iron ore, called inyimbo from bog ores. The Nyiha, as described by Brock and Brock conducted a secondary refining of the ore in a smaller forced draft furnace called ishitengwi (Brock and Brock 1965:98). Again this furnace is similar to the one used in the Fipa process.

In the Nyiha smelting practice the most important content of the medicine basket were the meat of a snake called inyuvwila, said to be phosphorescent and water-repellent. One of Brock’s informants said it was a snake that flew. The meat of the snake was so strong it had to be kept away from women and was covered with beeswax wax. Only the master could see it and when he used it in the furnace trench only the wax was actually used, not the meat itself. The snake is no longer found in Unyiha but Brocks’ informants said that it was still living in the lakes in Iwa in Northern Zambia (Brock and Brock 1965:100). The master prayed for success of the smelt to his ancestors and Mulungu, the Nyiha divinity. When he prayed, chickens were killed and the blood spread on the large and small furnaces. Sexual intercourse was forbidden, and as for the Fipa, women were only allowed to bring food to the smelters. Unfortunately Brock and Brock do not provide any descriptions or explanations for the rituals or taboos nor the use of magic in Nyiha iron smelting. The Tabwa Of interest to the Mwika group as defined by Lechaptois, is Derricourt’s study in the Northern Province of Zambia (Derricourt 1985). The furnaces in the Northern Province are of the natural draft type similar to the Fipa, Lungu and Nyiha furnaces. He sees a possible connection between the Lungu and the Mwika complex based on pottery and smelting technologies. Derricourt believes this complex in which also the Tabwa are discussed, includes the entire Tanganyika-Nyasa corridor (ibid.: 103). The people who today call themselves Tabwa live in northern Zambia, in the area between Lakes Mweru and Tanganyika and in south-eastern Zaire in the corridor running between the Luvua River to the west and Lake Tanganyika to the east (Cancel 1989; Davis 2000).

Brock discusses how smelting and forging were separate skills in Unyiha but, as in Ufipa, men could be both master smelters and master smiths. Brock’s main informant, Pima Nkota stated that iron smelting was not confined to certain clans or lineage’s, while other informants said that the Nkota name itself referred to one of the clans and chiefdoms in former Unyiha (ibid.: 99). Informants also claimed that on the death of a master smelter the (Mwasenga) chief arranged for his representatives and the members of the Nkota clan to select a new master smelter among those eligible by heredity and training (ibid.). When a suitable strong man was chosen he was caught and shut up in a house where he was invested with the tools of his trade. After a day he was brought out to be shown to the people as the new master (ibid.). As Brock describes, it what follows resembles hunting rituals and especially those described in chapters 6 and 7, the Kalulu ritual in Fipa iron smelting and wedding ceremonies.

The Tabwa ironmaking process has been recorded by Belgian missionaries and more recently accounted for by Roberts (Roberts 1991, Roberts and Mauer 1986).171 Cline, as well as Herbert, associate Tabwa furnaces with the other gynecomorphic furnaces such as the Chockwe (Lunda), Luba, Luchasi and Lozi and the Bena, Kinga and Pangwa furnaces further south (Cline 1937:53; Herbert 1993:32). The Tabwa furnace was sometimes referred to as ‘the mother of twins’ and the bloom was extracted

169

In his reading of Dantz, Fülleborn noted that the furnace had five tuyere inlets (Fülleborn 1906:168). 170 In Ufipa the master smelters said a Fipa iluungu was used about twenty times before it was abandoned.

171 Cf. also Schmidt and Mapunda (1997) and Schmidt (1997) for fuller descriptions of Tabwa iron working.

127

Randi Barndon from the furnace at the new moon.172

the short shaft furnaces (Ngoni and Shona furnaces) that operate with forced draft (Avery 1988:267). Hence, they resemble the Pangwa, Kinga, Matengo, Bena furnaces in the Southern Highlands region.

Three different furnaces have been recorded in Utabwa in more recent surveys. However, it is uncertain whether they represent different technologies or if it is the three stages of the entire iron working process as in the Fipa, Nyiha and Lungu smelting practices that have been observed (Schmidt 1997). The Tabwa masters, sisislungu, used magic in iron working. Similarly to the Fipa they placed magical and medical substances in the foundation trench of the furnace and in the middle of the furnace chamber. Robert describes how the Tabwa smelters bathed in smoke before the smelting. The bath contained herbs otherwise used by Tabwa women suffering from infertility and delivery problems and plants associated with keeping malicious spirits at a distance (Roberts 1993 in Schmidt 1997:233235).173 Similarly to the Pangwa they used medicines that were normally used by women and they believed heat and wind transformed herbs and plants , not too different from the Pangwa’s use of the umboocha poison.174 Features of the cosmology in Fipa iron smelting discussed in Chapter 6 and 7 are recognised in Tabwa iron smelting. Willis included the Tabwa in eastern Zaire/northern Zambia in his 1991 study on Fipa use of body metaphors and argued that the use of body metaphors is based a resemblance in cosmology between the two peoples (Willis 1991:275). Willis suggests that the Tabwa like the Fipa conceptualise the body as polarised between two forces or principles called ‘head’ and ‘loins’ (ibid.: 277).175 Thus both the Fipa and Tabwa seem to in distinguish between head and loins as concepts related to the human body that are transferred to concepts of gender and cosmology (Willis 1967, 1991), and furhter transferred to their iron smelting. Herbert suggests a link between Fipa and Tabwa based on Roberts’ notions of Fipa twin symbolism (Herbert 1993:143, and cf. Robert 1949:244) and the Tabwa’s celebration of new chief and twin births (Roberts and Maurer 1985:183). Mapunda further elaborates the link between Fipa and Tabwa in his discussion of pottery and investigations of iron technology along the shoreline of Lake Tanganyika (Mapunda 1995:55,104), while Schmidt elaborates on the similarities in iron working practices (Schmidt 1997). Avery has suggested that the tall natural draft furnaces confined to the lakes Tanganyika, Rukwa and Malawi should not be compared with the tall natural draft furnaces of western Africa; rather they are comparable to 172

In Tabwa culture the rising of the new moon is a central theme and an image often occurring in rituals, art and thought (Herbert 1993:89). 173 See Schmidt and Mapunda 1997:78. 174 Illustrating the complexity in linguistic terms and loanwords of the region, the Tabwa have a clan called the Bena Lungu, ‘the calabash clan’ (Willis 1966:40 and fn.199). 175 Cf. Chapter 7.

128

Kinga A village called Ku Ndapa, said to mean ‘near the iron’ and more specifically the iron ore, mdapo, was singled out by Bornhardt as the main smelting centre in the Livingstone Mountains (Bornhardt 1900:80).176 The village was located at 2.500-m altitude in the Hugilo range, not far from the German mission of Bulongua which is today’s centre in Ukinga (Kjekshus 1976). Kjekshus basing his information on Bornhardt describes how ‘the ore was dug from the mountainsides in pits driven up to five metres into the ground, and smelted in cone-shaped ovens that seldom rose more than one metre over the terrain. Blooms of approximately five kilos were produced from each smelt, which lasted two full days’ (Bornhardt 1900:167; Kjekshus 1976:88). Elton had earlier written about ‘the iron mountain’ of the Kinga and claimed that these ‘tribesmen’ supplied most of the market for iron and tools among the peoples living north of Lake Nyasa (Elton 1879:321 in Kjekshus 1977:88). Kerr-Cross told a similar story from his visit to Rungwe in 1893 when he found that the area was abundantly supplied with iron (Kerr-Cross 1895:118). Bornhardt (ibid.) also commented on the from 2 to 5 metres deep mines in the Kinga region. Fülleborn provides a full description of Kinga smelting furnaces. The furnace was about 1 meter tall and made out of clay. The furnace chamber was 70 cm high and 40 cm in diameter. Three sets of clay pipes of about 5 centimetres in diameter were placed approximately 6 cm within the furnace chamber (Fülleborn 1906:167). The tuyere and bellow system is identical to the Pangwa system and a drawing of a Kinga furnace with its pot of medicines confirms the relation between the two traditions (ibid.: 169). In addition to the Pangwa ing’anjo pot with its perforated lid, the Kinga (Ngoni?) in Fülleborn’s description made openings from the pot and through to the furnace chamber (ibid.: 169). This was not observed during my own fieldwork. Fülleborn noted that Kinga blacksmiths also used medicines in forging (Fülleborn 1906:170). Apart from noting the actual use of a medicine pot beneath the furnace chamber Fülleborn did not describe or comment on symbols of the obvious female attributes on the Kinga and Ngoni furnaces that he observed. In Fülleborn’s drawings these traits are missing (ibid.: 168-169). According to Sutton the Kinga furnaces were about one metre high and 8 cm wide at the base. The furnace operated with three pairs of drum bellows (ibid.: 177). Sutton states that the actual smelting would take up to 9 hours (Sutton 1971b), not two days as Fülleborn noted (Fülleborn 1906:167). Sutton notes that the process resulted in a forgeable bloom that did not need further 176

Cf. also Fülleborn 1906:166-167 and Kjekshus 1976:88.

Masters of metallurgy, masters of metaphors refining. Sutton concludes that even if some of the furnaces were located close to resources such as charcoal and ore, this association was not at all a rule (ibid.). While Sutton and his collaborators did not actually observe iron smelting in Ubena, Ukinga, Upangwa or Umatengo his analysis is important and his drawings of a Kinga furnace depict the similarities in furnace operation with the Pangwa furnace that I have observed.

The Chewa, Phoka and Chulu Phillipson reports from his archaeological and ethnographic investigations in Zambia that the Chewa used two furnaces in their smelting process. The first was a domed shape furnace called mucelo that was operated by natural draft. A smaller furnace, kantengwa was used for a secondary refining of the iron (Phillipson 1968:102). At the furnace base of the tall natural draft furnace Phillipson observed small holes that his informants could tell him had been used for medicines. Medicines were according the aged informants only used once, not before each smelt as recorded in Ufipa (Barndon 1992).

Kinga iron smelting has unfortunately not been studied in more recent times.177 Further south in the highlands Sutton has documented Bena and Matengo furnaces that appear to be identical to the Kinga and Pangwa furnaces (Sutton 1971b, 1985), but again no recent studies of their smelting practice exist, and detailed observations of the use of magic is lacking.

Merwe and Avery’s study of the Chewa and the Phoka smelting practices on the Malawian side of Lake Nyasa illustrates in detail the similarities within the region (van der Merwe and Avery 1982). The Phoka smelting process was carried out in two furnaces. The large and primary furnace operated by natural draft and was called ng’anjo in both Chewa and Phoka languages (Tumbuku).179 The product from these furnaces was refined in a secondary smaller furnace called chiramba in Chewa and katengu in Tumbuku (ibid.). Merwe and Avery describe the use of medicines and rituals in the Chewa process. The first ceremony in Chulu smelting starts when the foundation trench has been dug out. Pegs marked the circumference of the furnace and doors. Medicines and maize flour were pounded and grounded ‘by old women without a husband’ i.e. a woman both infertile and celibate. The smelters asked their gods and ancestors to bless the work with success and counter evil influences. The aged smelters sacrificed a white rooster and its blood was spilled outside and inside the ng’anjo furnace. The head of the chicken, its liver and one wing were placed under a musolo tree as an offering along with a small pot of beer. At this stage villagers who had overseen the ritual were told to leave the site. Lowering a boom over the path to the village sealed the smelting site. A font consisting of a bowl of tree bark in a forked upright stood next to this gate. Pieces of bark, wood and water were added to the font. As the boom was raised for each participant to enter they washed their hands. According to Avery this act served both as a purification of the smelters and a distancing of them from ordinary people( ibid.).

Wilson claims that the Kinga provided chiefs for all the peoples in the region (Wilson 1958:12). They were also responsible for the hoes, the fire, the food and the stock (ibid.). When we were in Upangwa we discussed the Pangwa relation to the Bena and the Kinga. Our informants first claimed they were as Bena as the Bena themselves, referring to their similar dialects.178 But on several occasions during our visits the smelters claimed their origins went back to Ukinga. They said they had only arrived in Upangwa after a decline in the demand for iron in Ukinga, and one of our own masters was himself of Kinga origin. Like everybody else that was interviewed, he claimed the first Pangwa smelters were of Kinga origin. They had, like his father, immigrated to Upangwa because there were too many smelters already in Ukinga. The Kinga master smelters are believed to have provided both the Nyakyusa and the Pangwa with their master smelters (Stirnimann 1976; Wilson 1958). In the Livingstone mountains north of Lake Nyasa peoples such as the Bena, Kinga, Pangwa and Matengo smelted iron in similar, approximately one meter high furnaces (Bornhardt 1900:80; Fülleborn 1906:168; Koponen 1988:261, fn.155). Similar, but smaller furnaces were used among the Ngoni (Fülleborn 1906:169; Koponen 1988:261). The furnaces of this region varied in height but were of Kense’s (1983) and Childs’ (1991b) Type B; a low forced draft furnaces. In 1968 Sutton conducted a survey of the northeastern Southern Highland region (Sutton 1969, 1971b, 1985). This survey of Bena, Kinga and Matengo furnaces has confirmed that iron working was common among most of the Southern Highland peoples until the beginning of this century, not only the Kinga.

Smelting started when the participants were told of taboos such as total isolation from outsiders and abstinence from washing their bodies during the smelting days (ibid.: 158). Merwe and Avery reports of no rituals associated with the secondary furnace, chiramba (ibid.: 159). The Chulu smelters were not able to produce forgeable iron and unfortunately Merwe and Avery do not report local explanations for the failure. The answer had to do with the kinds of ore they had used (ibid.: 163). When the smelting was completed the villager were allowed to come and look. In Merwe and Avery’s re-enactments two of the Phoka master smelters, murungu, together with a

177 The only more recent study of the Kinga that I have come across stems from Koizumi who does not include aspects of iron smelting in her study of religious changes among the Kinga of Bulogwa (Koizumi 1995). 178 Cf. Culvick and Culwick 1935.

179 The Phoka belong within the Tumbuku language group (Merwe and Avery 1982:147).

129

Randi Barndon medicine man, msofi, led the process (ibid.: 159).180 Merwe and Avery notes that the Phoka smelting was conceptualised as human production. While building the furnace it was called a pot, chimpani but when the furnace was ready built it was called mwali, a young woman ready to marry. Traditional fire drill sticks made the furnace fire. The smelters used an extensive amount of medicines that constituted the most of Phoka pharmacopoeia for dealing with witchcraft, epidemics and venereal diseases that reduced fertility. Other items were symbols of fertility. A small hole was dug out in the centre of the furnace ground and medicines, sepo, were placed in the hole after the appropriate medicine had been determined by divining. The medicines were of two major kinds, ‘chases evils’ and ‘chases away ancestor spirits’ (ibid.: 160). When the foundation trench was completed a brown rooster was killed and its jerking body left in the middle. Where it finally flopped would decide the doorway of the furnace. Accompanying the killing of the rooster the smelters said ‘allow this place, allow that this pot produces a wealth of iron’ (ibid.: 161).181 When the furnace was ready it was celebrated. Women, children and the smelters ate maize porridge and the rooster and sang and danced around the furnace. Millet and water were sprinkled where the doors were to be opened. The master closed the smelting camp by placing medicines, sepo at the paths. From now on nobody could leave the camp or wash themselves. Again each participant was told to observe the vows of celibacy, peaceful conduct and isolation. The secondary refining furnace, kathengu, operated by two bellows, which were made of goatskins. Before starting the work three pieces of sepo medicine were placed in a small hole in the furnace middle. The description seems to combine smelting concepts and perceptions of the technology from both the Fipa and the Pangwa. 9.2 About the interior mosaic, several peoples but one tradition? The technological practices described above are certainly connected. Nevertheless ethnicity or how observed data have been described in ‘tribal’ terms are problematic. While there were most likely pre-colonial ethnic differences and social boundaries between peoples, these differences did not correspond to the colonial and ethnographers idea of a ‘tribe’ (Koponen 1988:43; MacEachern 1998:111). Koponen further makes it clear that when we speak of pre-colonial peoples such as the ’Hehe’ or ’Chagga’ this does not necessarily refer to the modern populations with these names, not even always to their ancestors but to the people who happened to live in the areas 180

In 1982 the murungu and msofi were two different persons while in 1983 one man assumed both offices (Merwe and Avery 1987:159). 181 cf. also Cline for a slightly different description among the Chewa of Dowa district (Cline 1937:119).

130

which the modern peoples currently inhabit (Koponen 1988:43). Following Koponen, the important thing is not whether ethnic terms refer to integrated self-conscious ethnic groups, something they often did not, but whether they can function as categories and help to depict similarities and dissimilarities in the characteristics of different peoples. Most people within the area claim that their chiefs and kings originated from a different region than the rest of the peoples, such as the Fipa whose kings have been said to originate from the Northern Lacustrine region. However, Wilson claims that the Nyakyusa kings or chiefs originated from the Kinga (Wilson 1958:12), and Stirnimann (1976) claims a similar Kinga influence for the Pangwa. The interesting point here is not the actual origin, which might be impossible to establish anyway, but that in all the instances the kings and chiefs were associated with a knowledge of iron smelting that they brought to the people (cf. Maret 1985). Chief Kalula Ilungu, the mythical founder of the Luba state, is connected with iron and iron making (Reefe 1981:81-82; Maret 1985:74). Reefe describes how Kalula Ilungu in Luba, Zaire, was thought to smelt iron in a more sophisticated way by building the furnace modelled on the human body (Reefe 1981:82, cf. also Herbert 1993:32; Maret 1985). This touches perhaps upon the very initial metaphoric construction in iron smelting and may even be an explanation for the widespread East African use of body metaphors in iron smelting. Pottery making may serve as another (Gosselain 1999), although any original source for the embodied use of metaphors has not been suggested. The Fipa, Lungu, Nyiha (and Pimbwe) technologies on the south-eastern side of Lake Tanganyika are not based on the same materialised human model since their furnaces (amaluungu) do not visually connote a woman. Nor do the Tabwa furnaces in northern Zambia, even if the Tabwa are geographically closer to the Luba. The metaphorical link is imaginative more than material over vast areas of the region. Nevertheless, when a new furnace was ready to be charged with fuel and ore it was decorated as a young bride (Barndon 1992; van der Merwe and Avery 1987; Wembah-Rashid 1972; Willis 1981; Wise 1958b; Wyckaert 1914). Additionally, the secondary refining furnaces, ichinteengwe, kitieengwe etc. used in all these technologies had indeed a design that brings them within ‘the group of overtly gynecomorphic furnaces’ (Herbert 1993:32). According to Lechaptois all the peoples in the Mwika group were famous ironworkers (Lechaptois 1914:254; Wilson 1958:25). The name given the natural draft furnace, iluungu (pl. amaluungu) is said to mean iron smelting furnace in Fipa, Mambwe, Lungu, Nyiha and Pimbwe languages (Willis 1964 and see Appendix D, table 10). Clark claims that the Lungu arrived in the valley after the Fipa. But Clark believes the Fipa to have arrived rather late in the region because their pottery differs from the LIA pottery from Kalambo Falls. As

Masters of metallurgy, masters of metaphors furnace liteende, but the medicines in the furnace pot, ing’anjo, they called ing’anzo. The Ngoni furnace was called nganzo like the medicines in the Pangwa furnace pots. Fülleborn described how the Ngoni used to place a medicine pot with a lid with four perforated holes in it (Fülleborn 1906:169). The Phoka (Tumboka) called their natural draft furnace n’ganjo (van der Merwe and Avery 1987). Among the Pangwa the tuyeres were called n’gelwa, similar to the Chewa, ncelwa. Phillipson describes a furnace called muculu (Phillipson 1968), a term that resembles the Fipa, Lungu, Nyiha and Mambwe term for tuyeres, incehlu. The Phoka described by van der Merwe and Avery, called the furnace a pot instead of a womb, while their wives were called inganjo instead of wife etc. It is recorded that the Phoka called their furnace a pot, chimpani, while it was being built, but when the tuyere inlets were cut out from the furnace wall it became a young woman ready to get married, mwali (Merwe and Avery 1987:159). This exemplifies both the gendering of the furnace as a woman, its boundedness and openness. The Fipa called their secondary refining furnace by many names, such as ‘the one who reveals secrets’, a ‘tunnel’ and a ‘cooking pot’. This use of material and linguistic metaphors points at how societies use some basic embodied metaphors as memories that transcend the actual technological differences. The importance is the performance of rituals that make the smelters aware of the most essential aspects of the smelting; its prohibitions and possibilities that again are basic moral laws in the societies (see Connerton 1989).

Sutton has later remarked this approach to linking cultural change with peoples is questionable (Sutton 1985:139). Sutton points at another disturbing element in Clark’s work; the connected notion that tribes exist, and existed and arrived as tribes (ibid.). Most likely knowledge of smelting in natural draft furnaces has its origins to the west. The Fipa smelters I encountered on this issue associated the name iluungu directly with the Lungu people. They claimed that the furnace main tuyere port and the smelters’ entrance into the furnace chamber and finally the opening for slag tapping during the smelt should always face towards Ulungu, the land of the Lungu (Barndon 1992).182 The linguistic term iluungu may also be connected to eating salt as indicated in the Tabwa’s use of the verb ilunga which itself is derived from a verb that refers to ‘putting salt in one’s food’, ‘to join things together’ or euphemistically to uniting sexually (Roberts 1991:250 with references). Roberts add to the story that one of his Tabwa informants told him that to assert that this verb meant ‘to add salt to food’ was ‘to hide [reference to] the male ‘seed’ or sperm, because the word really referred to incest (ibid.). Over the entire region the procreative process and male semen is conceptualised as a ‘seed’. Thus, seeds of life mixing in the mother’s womb with female substances such as the Pangwa breast milk, is a common way to explain procreation (Barley 1994; Herbert 1993; Heusch 1980; Stirnimann 1969, 1979; Willis 1991). Herbert has pointed at how the Bemba say ‘a woman in menstruation is ‘hot’ and therefore avoiding seasoning food with salt, which is sure to ‘cut’ (i.e. cause illness) to anyone eating it’ (Herbert 1993:118). The Pangwa add salt (from Ubena) to the offerings made for the ancestors, dead masters and spirits of the smelting place and the same use of salt in iron smelting has been observed among the Phoka and Chewa (Avery et. al. 1988:273).

Brock believes that the Nyiha iron working tradition originates in Luba but has passed through Unyamwanga and Uiwa. In Unyamwanga north of Unyiha many informants claimed a Luba origin. His master smelter said they were from Nyamwanga. Hence, the Nyamwanga-Iwa iron working tradition is linked, although vaguely, to the great developments of iron working in the Luba-Lunda kingdoms of Zaire. This does not contradict the fact that the Fipa believe in a close relation with the Lungu people. The Lungu people were themselves skilled ironworkers (Chaplin 1961) and their furnaces were very similar to those of the Fipa (cf. Barndon 1992). Willis believes that their name both for the furnace and the people themselves is connected with the smelting furnaces in Ufipa (1966:40).183 Discussing these matters with my informants they told me that an ordinary medicine man learnt about smelting once a piece of a star, intaanda, fell into his medicine basket (Barndon 1992, 1996b). Others said the first smelters were not Fipa but Lungu that once lived in Zambia but came to Kisuumba, southwest of Sumbawanga town to settle in Ufipa. In Kisuumba they passed over their knowledge to those people who were already living in Ufipa. Then both the Lungu and Fipa chiefs left Kisuumba and settled in Milansi. Once the people of Milansi learnt about iron working it became well known within all of Ufipa. Because of this all

Phillipson reports that his Chewa informants mentioned a furnace, called n’ganjo that he was not able to locate or find traces of (Phillipson 1968:102). Van der Merwe and Avery refer to the term n’ganjo as a generic term for smelting furnaces among the Chewa (Van der Merwe and Avery 1987:148). Davidson and Mosley suggests that the term should be related to the more recent Ngoni invasions in the 1840’s and that it should be seen as a loanword (Davidson and Mosley 1988:87). The Ngoni called their furnaces nganzo (Fülleborn 1906:168), the Ila called their furnace ing’anzo (Smith and Dale 1968 [1920]:206). The Pangwa called their 182

Mapunda (1995) correctly notes that Ulungu country is not directly west of Ufipa, nor does the main tuyere port on the tall Fipa furnace face west, but he has misinterpreted my focus on the local meaning of the placement of the main entrance. West means towards west, towards the setting sun and towards Uluungu. Mapunda confronts this with the use of the ‘western baseline’ of a magnetic west which is more or less irrelevant.

183

131

As noted in Chapter 7 the Fipa call a substitute salt for iciluunga.

Randi Barndon smelters in Ufipa were called asiluungu and their furnaces were called iluungu. When discussing ‘his’ master smelter, Chaplin notes that he was a Lungu who had lived among the Chishinga people, a subgroup among the Bemba, all his life (Chaplin 1961.53). Although he had lived outside Ulungu his entire life he stated that he was a Lungu. He had learned iron working from his father who, in turn, had learned it from his father. Whether the former master smelter’s father or even his grandfather had moved into this area or not, is not made clear in the report by Chaplin. Wilson claimed that the Kinga provided chiefs for all the peoples in the region (Wilson 1958:12). They were also responsible for the hoes, the fire, the food and the stock (ibid.: 12). When we were in Upangwa we discussed the Pangwa relation to the Bena and the Kinga. Our informants first claimed they were Bena referring to their similar dialects.184 However, on several occasions during my fieldwork the smelters traced their origins back to Ukinga. They said they had only arrived in Upangwa after a decline in the demand for iron in Ukinga, and one of our own masters was himself of Kinga origin. Like everybody else that was interviewed, he claimed the first Pangwa smelters were of Kinga origin. In the smelting traditions described above, as well as what I was told in Ufipa and Upangwa, most of the master smelters claimed their origin outside the area where they were living and had once practised iron smelting. The Nyiha’s Nkota clan came from Iwa, the Lungu masters came from Mambwe, the Pangwa masters from Ukinga and the Fipa masters from Mambwe, Lungu and Nyamwanga. In both my case studies many of the masters that I interviewed told me they had moved several times during their careers as smelters. This was not only because of Ujamaa and villagization politics that forced villagers to abandon their homes (e.g. Johnson 1991) but because of the decline in demand for locally made hoes. The decline occurred decades before villagization politics. The first Nkota (master smelter) is believed to have come from Iwa in Zambia.185 The Iwa were ‘offshoot’ of the Nyamwanga people, the western neighbours of the Nyiha. One of my Fipa informants, an elderly master smelter, claimed his origins were Nyamwanga. At least during the last days of active smelting this indicates that master smelters were mobile within their own regions, such as the fathers of our Fipa and Pangwa master smelters, and that some of them migrated into new areas and managed to assimilate their tradition with the technology practised in their new areas. 9.3 From symbolic repertoires to symbolic reservoirs It is almost needless to say that we know all cultures to be dynamic and we also know that people change, to 184

Cf. Culvick and Culwick 1935. The Iwa belong, together with the Nyiha, within the M20 language group (Feldberg 1999). 185

132

impose and that they imposed and are imposed upon by other ways of seeing and doing things. I believe that the Pangwa and Fipa master smelters, among many other iron workers, contributed to both the diversity and unity we see in symbolic reservoirs or beliefs and in particular perceptions of the body in Bantu-speaking East Africa. Studying in detail two different iron working technologies in Tanzania has confronted me with various comparative levels that have led me into a broader understanding of iron smelting behaviour. I have attempted in my comparisons to move beyond classical anthropological comparisons of ‘they do this, they do that, they use bellows, they don’t, they smelt here, they there etc.’ (cf. Sutton 1971a, 1985). Gosselain suggests that an element or object in the ritual as a whole may seem illogical and must therefore be seen as derived from a distant context in which it was logical (Gosselain 1999). If seen in a broader African perspective the arbitrariness in iron smelting lies mainly in the way practical technological levels of the process have been organised. The cosmology, beliefs and customs have a common foundation as a social practice, from one region to the other (Cline 1937; Herbert 1993; Kense 1983; Schmidt 1997:264). Why this is so probably leads to more questions than answers in my comparison of the smelting traditions of some of the people in the interior. Still, I hope that my approach can contribute to a change in focus, something that I have argued that we need in order to figure out where to look in the ethnographic record for the cultural processes of change and continuity in the histories of iron making. Sterner has refined McIntosh’s term ‘symbolic reservoir’ in an inter-ethnic and cultural historical study of Mandara pottery (McIntosh 1989 in Sterner 1992:72). By definition ‘a symbolic reservoir contains the core of symbols, beliefs, values and ideas upon which cultures are founded (ibid.: 72). Based on archaeological and ethnoarchaeological research on pottery Sterner argues that the peoples of the Mandara Highlands in Cameroon have been drawing from the same symbolic reservoir for almost two millennia. While a symbolic reservoir may change over time because it is combined, recombined or reproduced, Sterner argues that it is still based on a limited range of themes (ibid., see also Schmidt 1997:264)). Sterner promotes an approach not too different from Barth’s account for variation in Inner New Guinea (Barth 1987) in which a multitude of peoples draw from the same vocabulary, material cultures and behaviours (Sterner 1992:74). Sterner argues that ‘although there are variations between the different groups in specific decoration forms and styles in pottery this should not prevent us from seeing that they derive from a complex, shared culture history’ (ibid.: 74). The similarities in the various smelting practices may be seen as part of the symbolic reservoir and the broader Bantu

Masters of metallurgy, masters of metaphors resignation “Some spirit does not want this. Let us appease him with a sacrifice”. Or else he will say with resignation “My medicines are not worth anything anymore, let us find others!”. And he will really work very hard at coming up with rarer bones and more extraordinary feathers and the skins of even nastier snakes. But most often he will cry angrily: “Again our women are behaving badly in the village. They are spoiling our work” and he will send spies to spy on them. Heaven help the poor accused or suspected one! She will have to prove her innocence by the poison test’ (Wyckaert 1914:375).187

philosophy of technology. The master smelters were able to expand beyond ethnic or social boundaries because of their skills in using the unity in symbolic thought (or, as Sterner calls it, a symbolic reservoir) with its long and deep tradition, and using this as part of an active repertoire (cf. also Kriger 1999:14). A similar symbolic or semantic model was part and parcel of Bantu pottery making (e.g. Barley 1994; Collett 1985, 1993; David et al. 1988; Sterner 1992; Herbert 1993; Heusch 1980; Gosselain 1999), cooking (e.g. Collett 1985, 1993; Herbert 1993; MacLean and Insoll 1999) and salt making (Collett 1993; Davison 1993). Bodily based metaphors were also used in rain making rituals and during the installation ceremonies of new chiefs or kings (Sanders 1998). One may certainly argue that as such it was based on themes from a shared ‘symbolic reservoir’. Using the concept ‘symbolic repertoire’ Bekaert has demonstrated that such a reservoir of symbols may have multiple layers of meaning that are contextually played out. Looking at Sakata forging Bekaert argued that the smiths played out different parts of their symbolic repertoire, a repertoire of meaning that draws both from phenomenology and structuralism. Both symbolic reservoirs and symbolic repertoires are valid terms when East African iron smelting symbolism is discussed. I believe iron smelting was based on a reservoir of symbols both within and beyond regional boundaries but certainly it was a symbolic repertoire that the smelters experienced and explicitly used.

Following this, Wyckaert describes the various ordeals and poison tests the accused smelter or his wife was asked to go through.188 When I confronted my Fipa informants with what would be the causes for a failure they said, similar to what Wyckaert reports, that most commonly it was one of the wives of a smelter that had committed adultery, not any of the smelters. If a smelter was accused of adultery he was called, ‘one of those that commit adultery’, wachililungu, meaning that he was like a woman. My informants said women were the most adulterous, therefore if a man was accused of adultery it was a great shame, insoni (Barndon 1992:110-111, see also Johnson 1991).189 If a man was forced to admit adultery he had to offer the master a white rooster. If the master accepted his excuses the master would re-smelt the iron. If the master did not accept the offer he excluded the adulterous smelter from his crew (Barndon 1992:111).

It was important to perform rituals, apply medicines and magic and mandate taboos in order for the thermodynamics of the furnace to function well. Kriger argues that mastery of the technology was not the masters’ only fortée, continuing and modifying the knowledge of furnace construction and operation was significant also in order to produce a host of social relations. As the furnace owner a master claimed the authority to divide and mobilise the labour of women, children and clients in preparing a smelt and in dividing the bloom afterwards. Kriger also includes the importance of how the master smelter manages skilled labour during the smelt and embellished it with ceremony (Kriger 1999). Kriger concludes that this was how he maintained his craft tradition while at the same time creating a highly prized and useful metal (ibid.). Hence, not only iron but prestige was produced during a smelt.

Stirnimann reports in a similar fashion that among the Pangwa a diviner was called in to find out reasons for a smelting failure. He would see whether a witch had visited the smelting site and disturbed the medicines or if any of the smelters had violated the taboo against sexual intercourse.190 If the diviner believed a witch had caused the failure the master had to find a new smelting site. If one of the smelters had been adulterous he had to give his portion of charcoal and ores as well as iron blooms to the mkoyo and his crew and could risk exclusion from the group (Stirnimann 1976:286-287). When we actually faced a failure in our re-enactment of a smelt the aged smelters accused the young men in our crew of adulterous behaviour but they also discussed the possibilities of the furnace not being dry enough and the possible low quality

Wyckaert’s discussion of a Fipa master’s reaction to a smelting failure illuminates what was at stake if a master did not make forgeable iron, and in his own words:

187 The quote above is translated by the author but adjusted to Merve and Avery’s translation of Wyckaert (Merwe and Avery 1987). 188 Wyckaert reports that occasionally people were killed but concludes that in his days of observation these ordeals were put upon the accused’s dog instead of the person who had broken taboos (Wyckaert 1914). 189 Today the the Fipa and Pangwa marry one wife but may have more ’wives’ within separate housholds (cf. Johnson, H. 1991). 190 Stirnimann reports that ’if one of the workmen during the night got the nightly flow of seed in the camp then he told his umkoyo in the morning. The master smelter sendt him home to wash himself. Then after sleeping in the boys house in the village the man returned to work the next morning’ (Stirnimann 1969:9).

‘What does the chief do then? Does he check his ore to see if it was good or bad quality? Does he check his flux to see if it was suitable to the quality of the ore? Will he try to find out if some natural cause has upset the operation? Occasionally yes, he will ask himself these questions. Ordinarily no! He will simply say with

133

Randi Barndon of the ores they had used. In both cases moral explanations for the failures were made explicit. Rather than providing practical or technological explanations alone the improper behaviour of people was the reason given. If our observations of failures had been within an authentic setting rather than a re-enactment the reputation of the master smelter would have been seriously threatened. If a master was not able to produce iron he would no longer manage to mobilise a labour group and followers. If his reputation as a skilled master was weakened he would have to move to a new area and re establish himself among a new group of smelters. The rituals performed in iron smelting were important at multiple social and individual levels beyond the smelting context. They were performed not only to secure a successful outcome or prevent a smelting failure; they were performed in order to stage the importance and skills of the masters. Technology may serve as an arena in which different kinds of interests can be defined, expressed and negotiated (Dobres 1995:27). Through rituals the master smelters were able to communicate that their technology was indeed complicated and that it was not ‘a given’ that everybody could master and manage. Ritualisation of the process insured that training and experience of the craft was transmitted only to those the masters wanted to be their heirs. In sum these major and minor events took place within a context where ‘a core vocabulary of dominant symbols’ or ‘a symbolic reservoir’ (Sterner 1992) was used. As illustrated in my comparison of Fipa and Pangwa smelting symbolism in the previous chapters with those of their neighbours, the symbolic reservoir that the participants in the technology, and thereby participants in a social event, seemed to refer to, were gender combinations as in sexuality, fertility and procreation. This conceptualisation was part of a ‘Bantu philosophy based on a thermodynamic image of the body. Hence, it was an experience that went far beyond metallurgy. 9.4 Iron as a thing of value Perhaps the relatively revolutionary impact of iron in subSaharan Africa can be explained precisely by the lack of any previous metal producing societies (Schmidt and Childs 1995). Throughout pre-colonial and early postcolonial times iron was considered first and foremost ‘a thing of value’ (Maret and Nsuka 1977; Phillipson 1988). It had its functional side because as tools it became the means, which facilitated work such as forest clearance, and tilling the fields. It became a source of wealth because those who possessed the knowledge of iron making became rich in fame and in fortune, in cattle and wives. Iron commonly forged into hoes became a medium of exchange for cattle and as tribute to the kings. Eventually it became a general money medium.

134

The master smelters were often given a special ritual or magical status; they were men of power and prestige. A Fipa master smelter was often called king, mweene, or mweene iluungu and when his assistants entered the smelting camp two wooden sticks were beaten together as a royal greeting (Barndon 1992, also reported in Greig 1937). How did the Fipa and Pangwa master smelters become men of prestige and power and what were they given in return for their iron blooms and iron hoes? The fact that there existed a local and a long distance trade network in the interior even before the Europeans penetration of the land was mentioned in my introduction. Another, and I would suggest, more essential reason behind the local power and prestige assigned iron master smelters was that they were the primary suppliers of ‘wedding hoes’. The Pangwa for instance and, probably also the Kinga, produced ‘wedding hoes’ for the Bena (Stirnimann 1976; Culwick and Culwick 1967). In pre-colonial Ufipa bridewealth hoes were also the most important, not cattle as in the northern areas. Willis comments that apart from a goat given at the wedding feast, animals were not essential (Willis 1981).191 This makes the southern lacustrine societies different from those in the north, mainly because iron and not cattle was the matter of prestige, of wealth and essential for living. The iron produced by the master and his labour force was of value to everybody. What was primarily at stake for the smelters was to produce iron, a material of value because it could be forged into tools and weapons and a currency that made people able to marry and expand their wealth. Nonetheless, after my study among the Fipa it became evident that many other issues of concern to personal and collective morality and social control were apparently also important to stage or play out during times of iron smelting. And this formed part of the smelters symbolic repertoires and reservoirs. In Upangwa and Ufipa, like in most of the neighbouring regions, blacksmiths made hoes for working the fields and gardens. The hoes were used in market exchange for other commodities such as cloth, salt, cattle or sugar; thus, they were the most central currency. Hoes were the most important symbol and gift exchange medium in marriage arrangements and were given as tribute to chiefs or kings. In Upangwa in the colonial days locally made iron objects were beginning to be disliked and feared. The increasing number of Christians believed traditionally made iron objects were dangerous and could harm them. This belief was grounded in a newly established fear of pagan rituals and the use of magic associated with local iron working. Mgina interestingly suggests this fear to be one of the reasons behind the decline in Upangwa local iron production (Mgina nd.).192

191

Cf. also Johnson for a more recent study of Fipa bridewealth customs (Johnson, H 1991). 192 Cf. Wyckaert’s photos of the master smelter wearing catholic necklaces (Wyckaert 1914).

Masters of metallurgy, masters of metaphors The symbolism associated with smelting has indeed been interpreted as attempted male control of female attributes and powers, of female fertility (MacLean 1996:236). However, if we try to understand the symbols as meaningful it becomes clear that the technology was experienced at multiple levels of which only some were of a clear social significance. Many were part of a longer and deeper symbolic reservoir, inter tribal and inter ethnic while the genderisation could be used for social purposes to limit the distribution of the technological and ritual knowledge. It did indirectly control female powers and assisted in a nuanced understanding of the technology, the event that transformed the ore to iron. Access to knowledge was controlled by not letting unmarried women participate or observe. Denying wives participation prevented in-laws neighbours from gaining access to the mysteries and magic of making iron. Finally, they were in use to control the entire group of smelters. Only masters knew the rituals and the importance and differences between female exclusions, female absence and presence. Only masters knew the real importance of the sexual taboos and magical applications that were vital for a successful outcome. They feared a spread of their valuable knowledge to in-laws and strangers. The most important knowledge was the ritual and the magical, not the practical or technological. The magic and symbols masked and mystified but they also simplified and explained. Did the iron masters in the past fear a similar uncontrolled spread of their knowledge, as I believe the Fipa or Pangwa did? Were they guarding the rank and respect with which the iron provided them?

People in Upangwa were ready to abandon old values attached to the locally produced hoes and other iron objects because they were against the magician and had been convinced to follow the missionaries and colonial administrators. This came along with wage labour and an increased use of colonial currency. I believe this insight may be relevant within a broader area than the Livingstone Mountains and Upangwa. The context the objects had been produced within was no longer admired or wanted and this was probably one of the forces behind the change from local hoes to imported ones being objects of desire. This was followed by the general decline in iron working because of economic and political interests. Kriger has recently discussed the decline in iron working, especially iron smithing in the Zaire basin (Kriger 1999). The same abrupt decline as she has recognised in Central Africa is seen in Eastern Africa. It is, therefore, as Kriger suggests, probably a response to the loss of regional ‘markets’ of which the currency market had been among the most important (ibid.: 236). Further west in central Africa metal wires were commonly the unit, but the ironworkers involvement in making and converting units of currency diminished along with the decline in value attached to iron (ibid.). In Ufipa and Upangwa iron hoes were the currency. Payment of bridewealth was gradually being more commonly paid in the colonial cash currency. Insignia and iron weapons were not produced as before. Moreover, the local chiefs and kings were not in office and the work in the forge diminished to minor repairs and the forging of special tools that were not available among the imported goods. It was not only an efficiency argument, that it was both cheaper and easier to buy imported hoes at the local market, that was the cause of the decline.

I believe it was important to perform rituals and express taboos in order to exclude other men. Knowledge, skill and innovation turned a smelter and a smith into a master. The secrecy and prestige that surrounded the smelters’ occupation made it difficult for opportunists to enter the profession. Only a small segment of a community or society was really allowed entry among the group around a master smelter. By using ideology and morality master smelters managed to keep the occupational group rather small and exclusive. Strong sanctions were often the response towards those who did not perform their duties and some apprentices were even forced to leave the group (cf. Barndon 1992; Wyckaert 1914). Iron working locations either outside living areas or within, as with the village forge, were socially defined settings. These settings were moral spaces as Beidelman calls them (Beidelman 1993). A setting where one could demonstrate mastery skills and play out questions about ideology and culture, about values and morality. This both constrained the spread of iron working techniques, the master skills, knowledge and know-how and made it possible to move beyond ethnic boundaries. However, this worked by magic, and without magic, no iron. This has been recorded in all the above mentioned documentation, it was the same in Ulungu, in Unyiha, Utabwa, and Ukinga, something that indicates that the same magical precautions and actions were taken in order

Kriger argues that one of the reasons behind the decline was that important aspects of the smelting and blacksmithing work and its source of prestige eventually disappeared (ibid.). She suggests the colonial currency as a major cause. I believe that cultural and specifically religious changes played a significant role in the decline and loss of prestige of local ironworkers and their products. The iron made in the smelting furnaces and iron objects forged were never only utilitarian and mundane because they were made in social contexts wherein medicines, magic and morals played a central role. Both arguments, prestige and religion, focus not on the efficiency of the produced objects, in agriculture or war, but rather the social significance of the iron objects and the pre eminence of the social status of the master smelters who had forged them. Iron had been ‘a thing of value’ and the makers of value had been socially highly regarded until the cultural value of iron declined. The purpose behind the notions of morality in the symbolic repertoire employed during a smelt was that in these societies gender symbols were used precisely to control power over ritual and technology and more specifically over an otherwise uncontrolled distribution of metallurgical (magical and technological) knowledge.

135

Randi Barndon to control the size of the occupational group within the entire region. Accepting the master smelters social role as magicians and specialists enable us to see how they were through their occupational group the transformers of traditions, beliefs and practices from one village to another (Kriger 1999:14). The masters were masters in metallurgy but also masters of making and using metaphors. This enabled them to use a repertoire of metaphors and symbols that was part of a shared interethnic or inter-tribal symbolic reservoir. By this they managed to move from one region to another. Possibly one of the most important factors and contributors to the transmittance and conservation of iron working practices was the master smelter. He was the one who tried to confine the knowledge of ore transformation within a limited range of people. The master smelter was also the one who because of his skills and talent through metaphorical entrepeneurship could cross ethnic and geographical boundaries, thus transmitting his knowledge beyond ethnic boundaries. When crossing ethnic boundaries he transmitted parts of his ideology and culture and supressed other parts, resulting in diversity within unity. Among the Fipa there are no direct associations between descent and territory. Cultivation rights are associated with village residence (Willis 1967:352,fn.5), something, which is common throughout the region. This made it possible for a master to settle into a new area and be first provided cultivation land and thereafter rights to use the land for hunting and iron making. What the master smelters managed was to isolate the profession from other domains of life. Therefore, one can say that they used a general level of embodied thinking and experience in order to play out and control the power and prestige in their work. By the use of general knowledge of morality in line with the shared symbolic repertoire of concepts and experiences they managed to prevent other men from establishing their own iron smelting groups. This they managed by taking their own folk model of morality into the smelting context and developed it in line with thr rules of sexual abstinence and fertility magic. In line with this their position as patriarchs provided the foundation on which their craft tradition developed (Kriger 1999:80). Many choices in iron smelting as a practice, as well as the metaphors in the symbolic repertoire, were part of a symbolic reservoir or ‘entrenched in traditions’ and based on tacit knowledge (i.e. Bekaert’s multiple levels). Nevertheless, metaphors and embodiments were, as Schmidt suggests, created by iron smelters and then worked to control an ideology surrounding the smelting (Schmidt 1983). The African master smelters have often been labelled ‘the masters of fire’ (Heusch 1956:59; Maret 1980, 1985). They were indeed masters of fire and as Kriger suggests they were the masters of the furnaces (Kriger 1999). But, as my approach has illustrated, their use of magic and ritual was even more important. The medicines and herbs the head smelters used were the

136

same as ordinary magicians would use in curing humans. What made smelters into master smelters was the combination of magical and practical knowledge and skills, hence mastering both was equally essential in order to build up a reputation as a master. This illustrates that the smelters were more than anything masters of metaphors because they used metaphorical thinking to conceptualise and operate the furnace technology. This conceptual framework was based on experience and used embodied metaphors that moved from the phenomenological type towards oppositional pairs and categories. These were staged in order to allowed people an entry or not in smelting. What the smelters used as symbols and metaphors was already there (as part of their symbolic reservoir) and already experienced by themselves and their onlookers or even intruders. They used embodied and experienced thinking and collective morality, and because of this they were able to move into new areas, assimilate within new ethnic sub-groups and mobilise new labour forces. 9.5 Summary What made iron working and iron workers special as opposed to other crafts such as pottery making or weaving was the value attached to the iron itself and the objects made from this valuable material. Using metaphors and their indigenous ‘folk model of morality’ the masters managed to uphold their reputation as masters. They created physical and embodied boundaries and ‘settings’ that were moral as much as metallurgical settings. This limited entry to the craft but was simultaneously one of the forces and possibilities that were at the centre of the process of transmittal of the technology. This repertoire was used by the masters in order to mobilise a labour force, control access to the craft and was taken into use when smelters crossed regional boundaries. The technology since it was based on a symbolic reservoir of deeply experienced metaphors made the masters of metallurgy into masters of metaphors. Their mastery of metaphors was played out when they were protecting their occupational group (e.g. by exclusion of those who had committed adultery) or when they crossed social boundaries (i.e. adjusted their own symbolic repertoire with that of their hosts). The master smelters used a local folk model of morality based on embodied metaphors in order to mobilise a crew and in order to exclude unwanted people from participation. Notably women were excluded but so were men whom the master did not want among his apprentices. When using metaphors and moral folk models he not only set the social rules for the actual smelting but he was able to control his own prestige. When comparing the different smelting practices I have argued that the master smelters used the same folk model

Masters of metallurgy, masters of metaphors of morality and embodied metaphors when crossing ‘ethnic’ boundaries. Because of this I have argued that the masters of metallurgy were even more the masters of metaphors founding their skills and know-how and working as an occupational group of specialists within and by a more general Bantu philosophy of thermodynamics.

137

Randi Barndon

138

Towards Conclusions

10. Towards conclusions Cultural meanings are only unimportant for those who choose to make them so (Tilley 1994:2). between traditional and industrial technologies and between magic and science. As distinct categories these could only be studied in different parts of the world or in different times or periods of the past and were therefore not open for comparisons. Tambiah believes, as I do, that ‘virtually everywhere life manifests totalities in which technico-causal and expressive-performative features are linked’ (ibid.: 136, cf. also Latour 1996, 1999). A translation or transcription of these totalities as belonging exclusively to one or the other domain echoes the earlier belief that magic or other animistic beliefs and acts were some kind of fake ‘science’ (Tambiah 1993).

10.1 A second discussion of technology; its individual bodily and collective basis. Heidegger stated that only by overcoming a purely instrumental conception of technology can we ever hope to understand it (Hood 1983:353). As he argued ‘man does not stand in some external relation to technology, it is not as once perceived, something apart from man’s being’ (Heidegger 1937 in Hood 1983:352). Mauss emphasised technology in all his studies, demonstrated the importance of the body in technological agency and noted that technological practice express the social (Schlanger 1998). This was taken as the starting point for the early French school of technology developed by Leroi-Gourhan and Haudricourt, a view upon technology that has been Lemonnier and Lecthman’s points of departure, both as an approach and a method in describing technologies. Latour argues that the divide between craftsmanship (the poeisis) and industrialised technologies should be rejected because it has never really existed (Latour 1999:195). Contrary to the opinion of Heidegger, Latour argues that there is a great continuity from handicrafts to industrialised technologies because the link between actors (humans) and objects or non-humans (computer chips, machines etc.) is still there. However, the chains in these collectives (the social) such as the number of subjects and object has increased (ibid.: 196).

In my introduction I stated in line with post-modern philosophical streams of thought that objective science does not bring us closer and closer to a final truth. Nor will we ever reach a final truth through comparison. Still, through comparison one might better understand the social and cultural content in transmission of technological knowledge and processes that through history lead either to regional changes or to unity within diversity in technological practices. The ‘Standard view’ certainly denied us a broader approach towards transmission of knowledge between occupational groups of smelters. I have tried to demonstrate that in Pangwa and Fipa iron smelting technological agency was a personal and collectively experienced embodiment. On the one side embodiment was used in making iron under the socially defined correct moral standards. On the other hand embodiment was crucial in order to provide a correct combustion or balance when applying ores and charcoal, medicines and magical substances or while working the bellows. Embodiment moved from individually ‘lived experience’ of everyday life towards seasonal technological activities, thus from one domain to another. The social setting in which this embodiment was used metaphorically was a technical and a moral space and this was the setting the masters controlled, conserved and changed through their technical skills, knowledge and innovation. This was the setting for moving beyond cultural boundaries. Because of the masters sophisticated use of metaphors they were able to adjust or change cultural norms in accordance with new social settings. As my cases have indicated master smelters commonly moved and established their way of making iron within other traditions. Accepting the master smelters and blacksmiths social role as magicians and specialists enable us to see how they were through their occupational group the transformers of traditions, beliefs and practices from one village to another (Kriger 1999:14).

In my discussion of a conceptual shift in the approach towards technology I criticised the ‘Standard view’ because of its misplaced influence on archaeological as well as anthropological studies of technologies. In its extreme form ‘the Standard view’ has denied us an interesting pathway towards how people create and recreate meaning in practices guided towards technological goals. It has ignored an empirical foundation for a broader understanding of the significance of rituals and magic in bridging domains from everyday life and the supernatural with technological activities (cf. Barndon 1992). The long lasting ‘Standard view’ encouraged anthropologists to exclude and ignore material culture in their studies and as a consequence of this neglect technology was given a peripheral status (Pfaffenberger 1992:492, Olsen 2000). Within ethnoarchaeology observations of magic, as something separated from technology but connected to traditional beliefs, have been applied to legitimise the use of indigenous peoples and their cultures as ‘cautionary tales’ rather than as leading on to new dimensions of technological agency. It has neglected the relevance of studying technology, both in terms of concepts and agency and for better approaching the relations between ‘science’ and ‘magic’ (cf. Tambiah 1993:24). The ‘Standard view’ has uncritically made a distinction

In Chapters 3 and 4 I discussed the historical background for, and Lemonnier’s chaîne opératoire approach and I

139

Randi Barndon experience. Nor do we fully understand how it was practised and transmitted within groups and between groups of peoples. More combined work is needed in order to establish possible links between the two fields of cultural and metallurgical approaches towards technology.

argued that this attitude provides a useful approach for transcriptions of technological agency in iron smelting. In Chapters 5 and 6 I described the Pangwa and Fipa smelting processes in terms of a partial implementation of a chaîne opératoire approach. In Chapter 7 I elaborated on how technological operations were perceived and seen as identical to transformations in the human body and I discussed how the technology, in particular the furnaces, were decorated. The furnaces were given female breasts (i.e. Pangwa), female thighs (i.e. Fipa) and washed in ritual cleansing water (i.e. Pangwa), or wedding decorations (i.e. Fipa). Thus, the furnaces became humans. As in all relations to humans, people live by moral rules and therefore sexual taboos were mandated and rules of avoidance became important. However, by reference to Bekaert’s discussion of levels of symbolism moving from a phenomenological type to a structural codic opposition (Bekaert 1998:23), I argued for a multiple way of experiencing and expressing embodied metaphors. Lemonnier’s ‘anthropology of technology’ is concerned mostly with my first dimension, operation, and the overall human component in technologies. I have argued that the material dimension, that is not lacking in Lemonnier’s approach but not made explicit, may point at how the material is imposing on the human component (after Latour 1999).

Connerton claims that ‘ritual is not only an alternative way of expressing certain beliefs, but certain things can be expressed only in ritual’ (Connerton 1989:54). Preliminary rituals and specifically magic were not separated but integrated parts of the overall conception of the chaînes opératoires of the smelting process. Douglas has suggested that ritual (and the use of the physical body in the ritual) effect social operations and were there in order to make statements about social and political structure (1996 [1970]: 70; Strathern 1996:16).193 Turner was however also concerned with the psychophysiological level in rituals and expanded his approach outwards from Douglas’ more narrow social explanation (Turner 1969; Strathern 1996:16). Strathern suggests that a combination of these approaches, a combination that both values the significance of embodiment and society, is more in line with the phenomenological thinking that brings these levels of society and the individual more closely together (ibid.).

I have discussed how the use of magic, prescriptions and prohibitions (e.g. sexual abstinence) was central. It partly structured participation in activities and the conditions under which raw materials were collected, prepared and transformed in iron smelting (see figures 5.1. and 6.1). These restrictions were linked to local models of human physiology and a philosophy of thermodynamics. Morality as such, the correct balances of the correct substances, therefore also partly formed the ‘paths’ or chaînes opératoires in the technology.

I think East African iron smelting worked on both a phenomenological and a structural level. As I discussed in Chapter 9, metaphorical constructions were memories and experiences of personal and communal character but used by the masters smelters in order to demonstrate, maintain and expand their own prestigious status. The iron smelters were not only managing metallurgy they were also particularly skilled in making and using medicines. From the beginning of the mobilisation of the work and during the various stages and steps of the smelting process until the final product, the bloom, had been achieved, they created and recreated meaning through their use of metallurgical, magical and metaphorical skills. Combined, this was their (own) chaîne opératoire approach based on a symbolic repertoire that moved from the phenomenological type to structural codic oppositions.

The sequences of a technology, or a society technologies, when once recorded are according to Lemonnier’s dream, one step in the direction of entering into the heads of the actors (Lemonnier 1992). I have tried to argue that the symbolic repertoires in the Fipa and Pangwa peoples iron smelting practices were based on embodied experiences and metaphorical imagination. The technology and transformation that took place in the furnace was based upon the use of magic and a metaphorical bridging of domains of human experiences of ‘being-in-the-world’. Because of this the heat transformation in the furnace chamber became equivalent to the ‘heat’ between man and woman during sexual intercourse. The iron in the furnace chamber had to be nourished with various homeopathic substances as the foetus in a woman’s womb. Within this line of thought sexual taboos, rules of avoidance and magic, incorporating gender combinations and oppositions became central. I have attempted to highlight the body used in technological activities, and the body as a model for the process. But few questions in terms of the actual embodiment and experiences as such have been raised. We still do not fully understand how embodiment is part and parcel of technology as an

David argues for a holistic approach in studies of African iron working technologies. He suggests that we should look for how magical and technical actions are imposing upon each other (David 2001). In our modern industrialised society we perceive technology and technological behaviour as something mundane or as a phenomenon existing outside our own control. In African societies where most technologies were, and still are in some places, seasonally adjusted activities, such as

193 I am here referring to Douglas’ statement that ‘the human body is always treated as an image of society, and there can be no way of considering the body that does not involve at the same time a social dimension’ (Douglas 1970:70). Douglas was greatly influenced by Mauss and his emphasis of the social dimension in body beliefs and techniques (e.g. 1935, cf. also Herbert 1993:235; Strathern 1996:13).

140

Towards Conclusions moved the setting from individual and bodily experienced realities towards collective sharing notions of morality guided by the supernatural. Taboos and secrecy surrounding the participants in the process did not prevent other members of the local community to be aware of the importance of correct behaviour during times of smelting. The season of smelting effected most people. It was a time for performances of social status, magical skills and the establishment of prestige. Prohibition concerned gender and age, related to sexual maturity and fertility, hotness or coldness. In Ufipa female relatives of the ironworkers would fear accusations of adultery if something went wrong in a smelt. And in the past they were indeed often those indirectly blamed. In Upangwa women were warned against the danger of looking at the smelting furnace and thus also blamed if a failure was the result. Poison or sorcery warned off men who had not been allowed entry into a master’s crew of smelters while sacrifices and magic assisted the masters.

hunting- and gathering activities, and the production of artefacts such as pots and iron, these activities were social events. Specific body decorations, specific food, taboos and preliminary magical rituals marked the activities and connected them metaphorically to life and morality. In these contexts metaphors were evoked and used. However, the symbols in iron working were never used isolated from other contexts of production or social life, they were part of the people’s modes of thought, thus parts of their folk models of morality. But iron smelting was an arena where social and moral values were staged and aspects of morality were part of verbal and nonverbal, material metaphors used in the process. Herbert has suggested that ‘a gendered universe’ in which gender oppositions run through the entire repertoire of symbols and work roles that created the specific iron smelting technological setting (Herbert 1993:224). Keeping Herbert’s insight in mind I attempted to add deeper levels of meaning to the picture. Thus, I argued that the power of a gender combination was staged in iron smelting because of an overall perception of the smelting technology as corresponding with concepts of the human body (male or female) in terms of thermodynamics or balances of substances.

Although concepts of the smelting process were equally for the private well-being of persons and for the more public moral folk model, making iron was not an everyday event. A ritual ‘cosmos’ was created around the smelting furnace and the smelting location because the smelting site was materially and magically ‘sealed’. Medicines were placed in the ground at the edges of the camp or placed on sticks at the entrance or ‘doorway’ to the site. The path to the smelting location was closed or footsteps in the path were taken away (van der Merwe and Avery 1987). Intruders were not allowed entry and those who participated were not allowed to leave the site before the smelting was completed. The participants had to wash themselves or were not allowed to wash themselves, had to cut their hair or were not allowed to. Sacrifices were conducted to please the ancestors and spirits of the place. The local conception of smelting was linked to cosmology (i.e. the Fipa ‘head’ and ‘loins’ concepts) but especially embodied experiences. Most smelters sang and decorated their furnaces or celebrated the event of a new built furnace as a wedding ceremony. The rituals and decorations gave the anthropomorphic furnaces its gender (Herbert 1993). Moral concepts guided the actions of the participants and onlookers. I have tried to demonstrate that in order to fulfil and maintain their position as masters the smelters used what was already at hand, individual and shared embodied experiences. Therefore, the masters of metallurgy were above all masters of metaphors.

Rather than only gender categories working in opposition to each other I have argued that the gender model worked by a combination of female and male powers that were equally essential for the well-being of persons and furnaces. The use of magic and medicines became important because thinking technology (or transformations) was thinking body, hence, curing the furnace body was like curing the human body. Because of this there were no clear-cut distinction between what was ritual or magic and what was not while the pre-eminence of the master smelter was constantly evoked. Many Pangwa and Fipa aged smelters that I talked with claimed that they had participated in iron working since they were boys. Yet, their work was not easy or merely based on unconscious routine because even if the smelters followed rules and prescriptions they could not predict the outcome and failures occurred. Many choices, actions and step by step operations were negotiable. Often new actions (magical or practical performances) were developed within the old paths or sequences. When using natural draft furnaces, as the Fipa did, there was little the smelters could do to control the outcome once the reduction process had started (Killick 1990; Rheder 1987). Hence, whether they managed to produce a bloom or not was often uncertain and beyond human control. Making the outcome a success the pre-smelt tasks such as selection of good ores and high quality charcoal were fundamental, but so were the preliminary rites, sacrifices and magic. These activities linked the masters to the spiritual world (cf. Herbert 1993:117). The decorations of the furnace and of the smelters themselves were performances grounded in metaphorical imagination. In order to achieve a success the use of magic became essential and an embodiment of the smelting context

In the previous chapters I have discussed how various levels of a symbolic repertoire worked and were used by the smelters, in particular the master smelters, in order to render meaningful technological agency and technological transformations, both the successful ones and failures. I have tried to move beyond the mere fact that the smelting process was modelled on the female body, fertility, sexuality and procreation. If fertility symbolism is provided as an explanation in itself it is merely an oversimplification of what was being conveyed 141

Randi Barndon in the smelting context. Of major importance were conceptions of the body, the power of gender combinations and moral standards for social behaviour. My description and discussion of embodiment in iron smelting has demonstrated how the Pangwa and the Fipa conceptualised technology and how people experienced the transformation from iron ore to bloomery iron, not as different from human transformative powers of procreation, but equal to it.

10.2 What did we observe in Ufipa and in Upangwa? The technology involved in iron smelting is partly a physical experience. Much of the knowledge is tacit; this means it cannot be verbally expressed or explained, only experienced. In a re-enactment or an ethnographic reconstruction of iron smelting we only see parts of the surface of what was once a living practice. We see only the first uncertain attempts to make something that is partly forgotten because most smelters do not possess the complete embodied experience and knowledge any more. This provides parts of an explanation to why the Pangwa smelters were not able to recreate a totally successful smelt and produce bloomery iron. Years had passed since they were building furnaces, making tuyeres, preparing ores, blowing the bellows and examining the temperatures within the furnace. The three aged masters in Upangwa had not practised their ‘know-how’ and skills in iron making after they stopped regular smelting. Because of this their bodily based experience in making tuyeres or blowing bellows was partly forgotten. What we managed to observe in Upangwa was their first steps in the direction of a recapturing of the embodiment of the technological experience. This kind of knowledge is tacit and cannot be recaptured in a moment. Like trying to ride a bicycle or knit a jumper one fails several times before one gets the balance and rhythm back in the body.

I have attempted to provide a multi-levelled explanation for why and how the embodied philosophy was related to iron smelting. I have argued that, since the technology was based on a thermodynamic perception of the human biological body, similar moral standards and a wish for the well-being of the furnace was demanded. This resulted in both dangerous oppositions between, and powerful combinations of, gender and associated cosmological phenomena. Therefore, one of the most fundamental issues to bear in mind is that Pangwa and Fipa technicians did not divide objects and subjects (furnaces and people) into separate categories (i.e. Latour 1996, 1999:147). My work indicates that for the Pangwa and Fipa smelters there was no ontological distinction between humans and objects, between women and furnaces. Finally, this results in a view of material culture, and technology in particular, in which objects or technologies (such as the smelting furnaces) are not representations of something else. Humans and furnaces (objects) were a totality or a meeting place that was needed in order to reach technological goals.

One commonly neglected part of iron smelting such, as the rhythm of the work is forgotten (Espelund 1997 and pers. comm. spring 1998). Espelund has pointed at how most experimental smelts conducted in Norway have resulted in failures. He suggests that this must be seen in relation to the fact that although one has knowledge of the furnace structures and operations in large, those conducting the experiments have not been familiar with the rhythms and procedures of the process, the chaînes opératoires. They have not possessed the lived experience of making iron. I believe this was witnessed in Upangwa. Iron smelting is a tacit, non-verbal knowledge. Not only were the ‘secrets’ or magic and techniques hidden and stored within the occupational group of smelters which was a small segment of society it was also tacit knowledge learned only through many years of apprenticeship. Most smelters I interviewed had been with their fathers observing smelts from ten or twelve years of age and gradually learned the work. As an experienced adult smelter one cannot fully explain the process verbally and the chaînes opératoires can only be demonstrated through actual behaviour. The technical operations carried out by individuals and groups of ironworkers as the rituals, songs and dances were part of a registry arranged in order to get the correct rhythm and balance of steps or operational sequences in the process. This is fundamental in the bellow working and in the charging of the furnace. Incorrect rhythm and balance will not provide the correct temperatures and combustion and will not provide the correct air draft in the furnace. Since my informants had abandoned iron smelting several decades ago many of the unconscious embodied action had escaped their memory.

From my point of view, based on a study of conceptions of the human body and furnaces, the master smelters in Upangwa and Ufipa were indeed related to each other. The relation between them is central, not because they were both iron working communities, but because they both modelled the smelting technology on conceptions of morality, containment and embodied metaphors. The functions and balance of the body and the operations in the smelting furnace were guided by the correct use of behaviour by the participants and others and correct magical and medical ingredients consistent with the masters symbolic repertoire. Hence, their concepts of technology and technological agency was in line with their folk model of morality and a Bantu philosophy of thermodynamics that included people and furnaces. This formed parts of the chaînes opératoires, and as I have argued, were more than social representations. Through an approach that has focused on ‘magic’ rather than ‘science’ the knowledge and skills of the master smelters as magicians has demonstrated why the furnaces were treated very much like humans. Their embodied thermodynamic perception of the technology effected spatial arrangements and layouts of the smelting sites (Barndon 1992, 1996a). Thus, embodied imagination was made into material metaphors.

142

Towards Conclusions metaphors and they had magical skills that they advanced as important in the process of smelting.

The master and his crew who demonstrated iron working to us in Ufipa were able to produce forgeable blooms. The Fipa smelters are famous for their iron working and the master smelter who personally contributed to this fame has been an ‘active’ smelter through the numerous re-enactments of smelting. He was a skilled and experienced master smelter and had the opportunity to keep tacit, embodied and bodily experiences, parts of the smelting intact decades after he had stopped regular smelting.

10.3 What messages were conveyed and to whom? What I have described above was part of a symbolic repertoire that was staged in our re-enactments of smelting practises. Nevertheless, our setting was a constructed setting for the purpose of research and therefore also a new setting that opened for a new moral space to be created. Some messages were more important to convey to us as foreigners than they would have been to an ordinary audience of co-smelters and co-villagers in the late 1940’s or early 1950’s. Co-villagers and cosmelters were after all familiar with their own philosophy and their symbolic reservoirs and repertoires. Maybe we were more like the audience of a smelter who tried to establish his reputation as a master in a new area. More than anything our aged master smelters had to convince us and the young blacksmiths about the importance of their skills and knowledge, their use of magic and sacrifices and how their own success would rely on the goodwill of their ancestors.

What both Pangwa and Fipa smelters had profound memories of was the symbolic repertoire consisting of preliminary rituals, prohibitions or taboos and sacrifices. As in the past this embodiment (personal and shared) was part of everyday life, part of marriage, the delivery of a child, pottery making, wood working, basket making, gardening and rainmaking rituals or ancestor worshipping. Within this symbolic reservoir sacrifices and prohibitions were easy to recapture as important in iron working because they are still practised. This being the case, illustrates not only why they were easy to recapture in iron smelting, but that parts of the symbolic repertoire were at hand and easy to adjust with existing symbolic reservoirs. In short, both my case studies illustrate that the non-metallurgical aspects (as we would frame them within a Western base line) have been kept more intact.

When I returned to Ufipa for the second field season in 1991 and while we were in the midst of preparations for the third smelt a peculiar event occurred. At first I did not connect it with the iron working or status of the ironworkers. I did not understand the messages the smelters tried to convey. It was in the early afternoon of 20th August 1991 and since we had left the smelting site, Tupa, earlier than we used to, I decided to do some laundry. Among the laundry was a pair of jeans that I hung out in a tree outside the house to drip off the water and dry in the warm afternoon sun. After dinner we sat in front of the house talking with some of the neighbours. Just before sunset I went around to the back and looked for my pair of jeans that were not hanging in the tree anymore. I asked the women in our compound since I thought that perhaps one of them had taken them inside the house for me. We looked around but it was getting dark and I decided to look for them the next morning when two of the ironworkers, who were active blacksmiths in the village forge and part of our crew, came to our house. The women immediately told them about my missing trousers. The men believed they had been stolen. We went inside the house and the men said that if they were able to get into trance they would find them for me and perhaps even find the person who had stolen them. One of the smiths painted his forehead with a white powder, not the red unnkolo powder he would use in iron smelting, and sat down in a corner of the room. We were sitting around the fire, waiting. He gradually moved into a trance, whispering words while his body was shivering. When he started to talk with a singing voice, the other smith could tell us where the trousers were and the name of the boy who had taken them. A third person went to collect them. I was happy to have my trousers back and impressed by the smiths. What really happened, possibly did behind my back, I will never know. Nonetheless, the message was clear. The smiths

A Fipa master smelter as his apprentices were known to have magical skills and were known to move into trance and trace thieves. This has not been paid enough attention in studies of iron smelting due to an overemphasis on sexuality and fertility symbolism. The focus has been on the furnace as a woman and the iron as a foetus while the ambivalent and admired role of the masters has been missing. Of primary importance throughout this work has been to discuss the way the smelting was seen in terms of thermodynamics similar to that in the human body. The masters were masters in metallurgy but also masters of making and using metaphors. This enabled them to use a symbolic reservoir that was partly inter-ethnic (a Bantu thermodynamic philosophy). They managed to use this deeply experienced symbolic reservoir as a repertoire in which magic and medicines were central in order to move from one region to another. In Ufipa no iron could be produced without the masters basket of magical substances. In Upangwa no iron was produced without assistance from the medicine pot in the furnace pit. The same emphasis on magical skills and knowledge was stated in reports from Ulungu, Unyiha, Utabwa and Ukinga and numerous other cases such as the Phoka and Chewa smelters. This indicates that medical and magical knowledge, being based on a conception of bodily thermodynamics, the correct balance of hot and cold substances (blood, semen and breast milk), were taken into use in order to control the size of the occupational group and render the process meaningful. The masters used the local and inter-tribal repertoire of symbols and 143

Randi Barndon ambivalent and ambiguous. The masters were both admired and feared. While we were waiting for the Pangwa furnace to dry, we saw heavy dark clouds arriving and the crew started to worry. One of the aged smelters announced that we did not have to worry he could blow the clouds away from our site. It was said in a joking and laughing atmosphere, and the younger smiths were amused, as were we. Then, a few days before our first smelting was about to begin in Kuking’ande an old man in Mkiu village suddenly became sick and died. We lived in a compound close to this man, who was a distant relative of the younger blacksmiths. The young blacksmiths believed one of our elders, one of our umkololo’s had used sorcery in order to state his powers as a master of smelting and a magician. Again, I had been reminded of the master smelter’s power to transform and control essences and materials. Moreover, we were lucky; it never rained!

had demonstrated that they were able to find thieves by getting into trance. I believe this event illustrates that they had confidence in me and were willing to inform me about this more secret role a smith can have. In Ufipa the village forge used to be a place where one could discuss juridical and domestic problems with the master blacksmith. A thief or a person accused of adultery would also stay in the forge until his case was taken to the village elders and village headman. When I arrived in Upangwa two years later and was told that the elders in our group there were magicians and medicine men, I gradually realised how the masters in Ufipa were indeed also medicine men and magicians during their work processes. I realised the importance of medicines and magic in combination with the supernatural. I saw more clearly how the status of the master smelter and master blacksmith was both

144

Appendices

Appendices Appendix A Pangwa terms used in the text mangulowo: gods maoka: spirits mawele: female beasts mene: goat mkololo: master smelter mooto: hot mulindi; the top opening on the furnace, also called ixinyalulepulwa (Stirnimann 1969) mwinja: bride pamokoka: making the smelting site panyi: down pori: forest sonda: goatskin bellows, skins used around shoulders when it rains twitiekhela : sacrifice ulweko : making charcoal umbihi: drum sticks used in iron smelting umboocha: ‘the furnace guardian’ umdala: woman umdala panyi: woman who sleeps on the ground (and not in her husbands bed) because she is in menstruation umdapo: iron ore umdebedebe: wives or husbands, vadebedebede, having sexual intercourse are considered hot and dangerous, vapyio, to many tasks. umkoyo: master smelter, leader of lutanana, descent group and work group umponji: generic term for smelter and blacksmith umtwangilo: stick used by women when preparing clay for pottery making and used by the smelters when they make tuyeres umkovi: the leader of community work or leader of a working and beer party umlindi: furnace pit and grave umlomo: ‘mother opening’ or ‘mother mouth’ (vagina), main tuyere inlet on the furnace umunyalisitu: ‘a woman in the bush’, a woman in delivery upechelo: fire drill sticks vakololo vajango: a group of smelters vapyio: hot after sexual intercourse

amajango: tuyere inlets, doors amatongo: testicles, the skin bags of the drum bellows amukanga: medicine men awalele: cold, pure hutwanga: kneading termite soil ifiloso: holes in furnace iksideksu/ixidexu: the vomit (xudexa:to vomit), slag ilala: ‘plants taken home’, herbal medicines ilidope: clay for making pots and furnaces ilivuka: pit for charcoal making imono: casteroil from beans, medicines in iron smelting ing’anjo: medicine pot ing’anzo: substances in the medicines pot ing’elu: tuyeres ing’elwa: tuyeres ing’oma: drum bellows, generic term for drums iseke: disease caused by having sexual intercourse with a woman in menstruation ixibolo: penis ivutali: iron iwefia: tubes, hollow sticks connecting the tuyeres to the drum bellows kalilo: small basket for flour kihivi: basket used for carrying iron ore kilulepo: tuyere inlet kiwalasha: purifying ritual water, a three and root of a plant used for medical and ritual purposes kualopila: ‘making something hot’, sexual intercourse kufukuta: to work the bellows to make iron or in the forge kulitekelo: ‘the place for prayers’ kwibepo: ‘the wind-making hole’, peephole at the back of the furnace libihi: sticks used when making tuyeres liteende: smelting furnace, ‘something beneath’ or ‘the one who smelts’ lutanana : descent group, work group luuto: furnace foundation trench, grave machocolondo: ‘the vomit from the furnace’, slag makara: charcoal mamlungu: gods mangala: iron bells

145

Randi Barndon Appendix B Fipa terms used in the text kalesa: Fipa god kalulu: hare, dry yellow flowers used when making fire with drill sticks kanawiinga: young marriageable woman kateembwa: firewood kasula: master blacksmith kasunga wachelu: tuyere makers kulumbilisha: ritual at furnace foundation, sacrifice. liindi: grave miaandi: wooden poles used for scaffolding the natural draft furnace mikoa: creepers used in scaffolding the furnace misuungu: long thin sticks used when making tuyeres mizanti: scaffoldings mweela: ‘something good’, a term given the tall natural draft furnace nawiinga: bride nkoko: cockerel ntaanda: the peephole on the natural draft furnace, a star palinyina: ‘the mother opening’, the entrance and exit to the furnace chamber and major tuyere inlet palisi: ‘the father opening’, next largest tuyere inlet at the eastern side of the furnace pamweela paliynina: ‘the loincloth for the maternal opening’ sawiinga: bridegroom ukusula: forging ulufula: foundation trench for the natural draft furnace ululu: pure and forgeable iron uluko: descent group ulusuumba: a termite mound umbululu: half reduced iron ore umpakasi: ‘the furnace caretaker’, a cockerels wing umuuwa: goatskin bellows umwaami: master smelter, leader of an elephant hunt undumenda: a boy unnsana: ‘loins’ unnsole: master smelter’s closest assistant unndosi: a witch, sorcerer unnkolo: red powder used in wedding ceremonies and during smelting unnsa: ritual earth used in marriage ceremonies unsakwe: a smelting site unsoosi: a sing’aanga changing into a sorcerer unsuungu: a girl unntwe: head siiko: a fireplace or cooking place, open fire in forge yaeemba: body decoration, clay and beads used in wedding rituals wachililuungu: smelters who have committed adultery

akio: bowl used in wedding rituals amaasi ileembo: magical water amaasi isi wiinga: ritual wedding water amalombwe: slag, midwives amapiso: hot amasisiwe: iron ore aluumba: master smelters assistants asing’aanga: medicine men asiluungu: iron smelters entoa: measuring stick, ceremonial wedding stick ifingila: ‘things which enter’, magical substances ifisiimba: ‘latent things, magical substances ifo: termite soil used for building a furnace ichinteengwe: ‘the one who reveals secrets’, secondary refining furnace icitifiti: ‘black iron ore’ iciluunga: substitute salt ikala: charcoal ikata: women’s carrying circlet ilawa: wedding spear ileembo: medical barks, generic term iluungu: natural draft furnace imbwe: hammerstone impaasa: ‘axe iron’, strong iron impaamba: ‘inside of female thighs’ inchelu: tuyeres inchila: sickness, pains in the back caused by adultery, a disease attacking men coming close to a fire where a woman in menstruation is cooking. incaango: a slag, ore and charcoal admixture that forms a ‘cake’ at the bottom of the natural draft furnace, literary it means ‘the iron stool’. It is also called untare: impure iron. ing’aanga: knowledge possessed by a sing’aanga, medicine man ingailo: herbal medicines used in iron smelting, barks from the eastern and western side of trunk of trees ing’aanga: sorcerer’s knowledge and medicines, refers to power incuunda: ‘ores for making knives’ insuka: making charcoal insoni: shame, low quality iron intakaaso: mixture of water and clay used in wedding rituals intaangala: basket of magical substances, generic term inteendo: cooking pot, term given secondary refining furnace inyiimbo: iron ore, generic term isiluungu: a professional smelter isuuli: goatskin worn by smelters around the waist

146

Appendices Appendix C Historical events in the Upangwa (Ludewa) and Ufipa (Rukwa) regions 1100 AD 1235-110 1410- 80 1550-1750 1800 c.1820 1832 183518401842184518501857-59

1897 1886 1888 1890 1897 1899 1902 1905-1906 c. 1912 1914-18 1916-61 192519301936 1939-45 19391953 19601961

Pottery (one C-14 date) excavated by Mapunda at Kirando in Rukwa region of the Kalambo Falls type. Ivuna salt plains are located south-east of Ufipa and north of Ukinga/Upangwa. Cattle bones found at Ivuna, near Lake Rukwa (Fagan, B. and E.J. Yellen 1968). Pottery like Kisi ware on Mbande hill in Ngonde Introduction of the katuruka iron working technology and related pottery to the Lake Tanganyika shoreline area (Mapunda 1995a and b).First written document about the Fipa by the Brazilian explorer Francisco Jose Maria de Lacerda writing about the Mussucumas (after Willis 1981:6). Change in direction of ivory trade: first crossed lake in time of Kyungu Mwangonde, whose son Mpeta fought Ngoni (Wilson 1977). Trade route from Katanga to the coast pioneered by Nyamwezi traders from central Tanzania. Along the route came new crops: maize, cassava and later firearms (Willis 1970:248). Portuguese explorer A. C. P. Gamitto reported on the Sukuuma, their location in today’s Bemba country, and how they (Musukumas or Fipa) had invaded the country (after Willis 1981:6). Ngoni crossed Zambezi on day of solar eclipse (Iliffe 1979). First Ngoni raids in Ngonde, Arabs arrived at Ujiji (Burton 1860). First Ngoni arrived in Ufipa. The Ngoni leader died in Ufipa. One section of Ngoni established them selves in the Songea area where they incorporated Pangwa and Ndendelu into a Ngoni state (Wilson 1958:4). From 1855 until 1900 the Pangwa were constantly invaded by Ngoni hoards (Stirnimann 1979). Richard Burton travels with John H. Speke through the Great Lakes region and comments on the considerable amounts of iron produced in Ubena and Urori as well as the abundance of slag in Ufipa (Burton 1961 [1861]:273). Slave trade (such as among the Bisa, Bemba and Lunda), Tanzanian slave trade grew from the early decades of the 19th century to some 25000 in the late 1860s and 1870s (Koponen 1988:85). Definitive German occupation of Ufipa (Willis 1981:231).Joseph Thomson visits the Fipa king of Nkansi (ibid.) The first European missionaries of the White Fathers (French Roman Catholic missionary society) settled in Ufipa (Willis 1981:199). Kerr-Cross travelled through Rungwe valley and over the Ndali hills. Mission station opened at Kararamuka (Kerr-Cross 1890, Wilson 1977). Agreement that brought Ufipa formally under the rule of Imperial Germany (Willis 1981). German administrative control was effectively established over Fipa territory (ibid.). Taxes collected by German administration in the corridor region. The first German missionary station Milo (Milow) in Upangwa was established (Stirnimann 1979). The Pangwa participate with the Ngoni in the Maji-Maji rebellion (Strinimann 1979:14), Milo mission station in Upangwa was almost ruined. Wyckaert observed iron smelting on the Ufipa plateau (Wyckaert 1914) World War I Fighting between British and Germans through Rungwe valley (Wilson 1977; Iliffe 1979). British rule in Tanganyika. Establishment of Indirect Rule under British Mandated Territory of Tanganyika. Ukinga, Ubena and Upangwa formed as a separate district and called Njombe. Numerous mission stations are established in Upangwa (Stirnimann 1976:263), Mumford established a school in Mlangali (1930). Greig reports approximately 20 high shaft furnaces still in use on the Fipa plateau (Greig 1937). World War II All German missionaries required to leave Rungwe District (Wilson 1977). Père Robert, a White Father missionary lives in Ufipa. Mzee Stefano Malimbo stopped iron smelting (pers. comm. July 1990). A small number of the Fipa were still smelting and forging iron during Roy Willis’ first stay on the Ufipa plateau (Willis 1981:278ff3). Independence of Tanganyika.

147

Randi Barndon 1962

1963 1964 1967

1969 1971

Roy G. Willis arrives in Ufipa the 8th of December in order to conduct anthropological fieldwork among the Fipa. Kalambo Falls excavations initiated by Clark, simultaneously he carries out ethnographic surveys among the Lungu and Fipa (Clark 1974). Chiefs and village headmen dismissed in Tanganyika. Migrant labour to South Africa from Tanganyika prohibited. Tanzania created as the new state (Tanganyika united with Zanzibar). Stirnimann arrives in Upangwa to conduct ethnographic and linguistic studies among the Pangwa. Arusha Declaration. Village Development Committees abolished, Functions taken over by TANU leaders. Two practising Fipa smiths demonstrate their craft in Dar es Salaam and the process is film recorded (cf. Wembah-Rashid, 1969 and 1973). Government’s proposals on Uniform Law of Marriage. The law of Marriage Act, Villagization acts, Ujamaa and Uhuru.

148

Appendices Appendix D Dimensional aspects of Pangwa and Fipa smelting furnaces List of figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5

African types of smelting furnaces Distribution of prehistoric and ethnographic smelting furnaces Sketch of Pangwa furnace at Ing’angotoli II Sketch of furnace base at old smelting site, Kumaanga Sketch of base on Fipa high shaft furnace used in re-enactment in 1990

List of tables Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 Table 8 Table 9 Table 10.

Reconstructions and re-enactments of iron working Pangwa furnace dimensions in re-enactment of smelts in 1993 and 1995 Samples of Pangwa tuyere dimensions in re-enactments and fragments from old smelting sites Pangwa furnace dimensions on old smelting sites Fipa furnace dimensions in re-enactments in 1990 and 1991 Fipa secondary refining furnaces used in re-enactments in 1990 and 1991 Tuyeres in Fipa re-enactments in 1990 and 1991 Old Fipa natural draft furnaces dimensions Fipa tuyere inlets and dimensions on old furnaces Local linguistic terms on furnaces in the region

149

Randi Barndon

A: BOWL

B: LOW SHAFT

C: HIGH SHAFT

Figure 1 African types of smelting furnaces (modified after Childs 1991b)

10

20

0

10

20

30

40

50

30

30

20

20

10

10

0

0

N 10

10

20

20

200 4

6

8 10 12 14 16km

30

30

10

Key:

0

Types

10

20

A

30

B

40

C

50

Unknown

Archaeological Ethnographical Distribution of furnace types

Figure 2 Distribution of prehistoric and ethnographic smelting furnaces of Types A, B and C (modified after Kense 1985:20)

150

Appendices Table 1 Reconstructions and re-enactments of iron working

Group

Furnace

Country

Reference

Year

Fipa

Type C and B

Tanzania

Wyckaert 1914

ca. 1909- 1913

Fipa

Type C and B

Tanzania

Grieg 1937

1936

Fipa

Type C and B

Tanzania

Wise 1958

1956

Nyika

Type C

Tanzania

Brock and Brock 1965

Lungu

Type C

Tanzania

Chaplin 1961

1960

Kaonde

Type C

Zambia

Chaplin 1961

1960

Fipa

Type C and B

Tanzania

Wembah-Rashid

1967

1966, 1971

(Dar es Salaam)

Fipa

Type C and B

Tanzania

Barndon 1992

1990 and 1991

Pangwa

Type B

Tanzania

Barndon 1996b

1993 and 1995

Phoka

Type C

Malawi

van der Merwe and

1983, 1983

Avery 1987 Phoka/Chewa

Type C and B

Malawi

Killick 1987, 1990

151

Late 1980’s

Randi Barndon Table 2 Pangwa furnace dimensions in re-enactment of smelts in 1993 and 1995, in cm

Furnace, liteende

1993

1995a and 1995b

Total height from pit to top

130

134

Pit

40

43

Height from ground surface to top

75

91

Bottom external diameter

88

70

Bottom interior diameter

48

50

Top external diameter

44

40

Top internal diameter

20

28

Number of tuyere inlets

3

3

Main tuyere port, umlomo

40 – 30

Width from ground surface 28

Height from ground surface Other tuyere inlets Width from ground surface

12

Height from ground surface

25

approximately 150 cm. During smelting the total height of the furnace was from base to top 91 cm. Three tuyere inlets were wide enough to fit in one tuyere into each tuyere port. The tuyeres were partly inserted into the furnace chamber. At the back of the furnace above one of the tuyeres a small hole was made, kwibepo that was used as ventilation and to control the combustion during the smelting. These were observed old abandoned furnaces. After smelting the clay around the tuyeres ports was vitrified as were the tuyeres observed at old smelting sites. Height from furnace base to kwibepo was on the first furnace built 40 cm and the internal diameter was 8 cm.

The furnaces in our Pangwa re-enactments and surveys varied in height from 75-100 cm and had an average inner diameter at the furnace base of 85 cm. For the first re-enactment of smelting the furnace’s inner height was during smelting 91 cm from base to top. A pit was made below the furnace chamber where the smiths placed the ing’anjo pot medicines immediately before the smelting started. The pit and pot were completely sealed with medicines and a layer of charcoal before the smelters started the charging of the furnace with charcoal and ores. As an example the second furnace built had a total height of 133.5 cm inside the furnace chamber including the pit for the medicines. The outer circumference at the base was 280 cm while the outer top circumference was

152

Appendices Table 3 Samples of Pangwa tuyere dimensions in re-enactments and fragments from old smelting sites, in cm

Tuyeres

1993

1995a

1995b

Ing’angotoli

Ing’angotoli

I

II

Kumaanga

Length

36

49

38

41

35

43

Width

9

10

11

9

11.5

9

Inner diameter

4

5

4.5

4.5

5.5

4

In our re-enactments the tuyeres were inserted 4.5 to 5 centimetres inside the furnace chamber. Table 4 Pangwa furnace dimensions on old smelting sites, in cm

Furnace, liteende Height from ground surface to top

Kumaanga

Ing’angitoli I

Ing’angitoli II1

60

80

93

Bottom external diameter Bottom interior diameter

55

41

Top external diameter Top internal diameter

36

Number of tuyere inlets

3

3

3

50

43

Main tuyere port, umlomo Width from ground surface Height from ground surface

33

Other tuyere ports Width from ground surface

12

15

12

Height from ground surface

25

20

24

Height from base to kwibepo

31

Kwibepo internal diameter

8

6

1

At Ing’angotoli II the master smelter of the furnace was buried 3 metres west of the furnace. Close to the site was a rock shelter where the smelter and his assistants used to stay while they were smelting. We observed a fireplace numerous potsherds and chicken bones.

153

Randi Barndon observed in the pit (or below) nor outside the furnace. At all sites observed in Upangwa tuyeres and tuyere fragments were observed close to the furnaces. At many sites complete tuyeres were observed inside furnaces but not in situ

At Kumaanga potsherds were found in situ in the furnace pit. Some of these sherds had perforated holes and must have been part of the lid. At Ing’angotoli I potsherds were found outside the furnace while at Ing’angotoli II no potsherds from the ing’anjo pot were

Table 5 Fipa furnace dimensions in re-enactments in 1990 and 1991, in cm

Furnace, iluungu

1990

1991a

1991b

Height from base

280

275

285

Height from base to peeping-hole

127

130

136

Bottom external diameter

207

207

182

Bottom interior diameter

165

159

142

Top external diameter

102

110

88

Top internal diameter

81

81

68

Number of tuyere inlets

10

10

10

Height from ground surface

56

54

45

Width from ground surface

49

48

40

Height from ground surface

50

48

38

Width from ground surface

19

18

15

Height from ground surface

41

41

38

Width from ground surface

18

18

15

Main western door, palinyina

Main eastern door, palisi

Smaller doors

154

Appendices Table 6 Fipa secondary refining furnaces used in re-enactments in 1990 and 1991, in cm

Furnace, ichinteengwe

1990

1991a

1991b

Height from ground surface

40

43

45

Height from furnace interior

48

63

53

Bottom external diameter

37

?

35

Bottom internal diameter

30

25

15

Top external diameter

48

50

52

Top internal diameter

30

35

35

1990

1991a

1991b

Length

24

25

25

External diameter

6

6

6

Internal diameter

4

4

4

Length between exterior end

5

4

4

Height from ground surface

15

?

19

Width (diameter)

25

23

25

Height from ground surface

17

14

18

Width from top

25

?

12

Table 7 Tuyeres in Fipa re-enactments in 1990 and 1991, in cm

Furnace, iluungu

and iron pipes Openings

Front opening

155

Randi Barndon Table 8 Old Fipa natural draft furnaces dimensions, in cm

Furnace, iluungu

A

B

C

D

E

F

Height from ground surface

3.20

3.25

2.84

2.93

2.92

2.70

Height from base to ntaanda

1.20

1.15

1.43

1.25

1.40

1.04

Bottom internal diameter

2.05

1.85

1.75

1.62

1.74

1.96

External diameter at ntaanda

1.75

2.04

?

1.02

1.17

1.38

Top internal diameter

1.20

1.25

0.63

0.55

0.69

?

Table 9 Fipa tuyere inlets, dimensions on old furnaces, in cm

Furnace, iluungu

A

B

C

D

E

F

Number of tuyere ports

10

10

10

10

10

10

Height from ground surface

0.70

?

?

?

0.47

0.50

Width from ground surface

0.40

?

?

?

0.46

0.44

Main tuyere port, palinyina

N

KWIBEPO

MILYANGO

LITEENDE

MILYANGO tuyere port

0

0,6m

Site: INGANGITOLI II Tegn.E.Hoff

LIBEPO/ UMLOMO

Figure 3 Sketch of Pangwa furnace at Ing’angotoli II 156

Appendices

N

UMPAKASI

PALINYINA tuyere port

PALISI or ordinary tuyere port

ILUUNGU

0

200cm

Tegn.E.Hoff

Figure 4 Sketch of furnace base at old smelting site, Kumaanga

N

MILYANGO tuyere port KINYA LULEPU

LITEENDE

MILYANGO

LIBEPU

0 Tegn.E.Hoff

Site: KUMAANGA II

Figure 5 Sketch of base on Fipa high shaft furnace used in re-enactment in 1990

157

0,6m

Randi Barndon never observed furnaces with more than ten tuyere ports. The old furnaces, including the one used for the third re-enactment, were placed sloping slightly downwards at the base with the main door, palinyina, at the lowest part. This resulted in the iron, incaango, being concentrated towards the main tuyere inlet.

The dimensions found at the furnace base and the size on tuyere inlets on the old registered Fipa furnaces were equivalent to the furnace used for our third reenactment observed in 1991. Even on furnaces with a comparatively wide bottom or base, the tuyere arrangements had similar narrow dimensions. We Table 10. Local linguistic terms on furnaces in the region

Group

Country

Furnace Type C

Furnace Type B

Tuyeres

Fipa

Tanzania

iluungu

Ichinteengw

Inchelu

Mambwe

Tanzania

iluungu

Ichinteengwe

Inchelu

Nyiha

Tanzania

Ilungu

Ishintengwi

Lungu

Tz/Zambia

Iluungu

Mucelo

Aushi

Zambia

Chishinga

Zambia

Bemba

Zambia

Pangwa

Tanzania

Liteende

Ing’elu

Kinga

Tanzania

liteende

Ing’elwa

Ngoni

Tanzania

nganzo

Chewa

Malawa

Ng’anjo

Chiramba

Chewa

Malawi

Ng’anjo or

Katengwa/

mucelo

kathengu

(Phillipson

Ilungu

1968) Phoka

Malawi

N’ganjo

Zambia

Ing’anzo

Tabwa Ila

158

Kintengu

Nchelo

Icintemgwa

Inchelo

Appendices Appendix E Participants in re-enactments of iron smelting in Upangwa and Ufipa Upangwa 1993 Randi Barndon George Gardias (field assistant and interpreter) Smelters and smiths: L. Mtweve, A. Mtweve, A. Mgimba, Mr. S. Mgimba. Upangwa 1995a and 1995b Randi Barndon Emanuel T. Kessy (field assistant and interpreter) Amin Msuya (field assistant) Ali Hamza (driver from History dept. University of Dar es Salaam) Smelters and smiths: L. Mtweve, A. Mtweve, A. Mgimba, S. Mgimba, P. B. Sanga Damakira, M. Mtitu, J. Mtitu, Mrs. P. B. Sanga Damakira. P. Mhagama. Members of our crew with the same names, such as Mtwewe are brothers. Mr. Mhagama is Mr. Damakira’s nephew. The Sanga name derives from Ukinga where Sanga was the leading iron working clan while Mtwewe was the leading clan in Lugarawa, Mlangali in northern Upangwa (cf. Stirnimann 1976:35.fn.2). Ufipa 1990 Randi Barndon Billham Kimathi (field assistant and interpreter) Smelters and smiths: D. Kakwaya, H.D.Kakwaya, J. D. Kakwaya, P. D. Kakwaya , J. Kaliela, N. J. Kanyama, F. Makumi, S. Malimbo, L. Masilanga, A. Mwanisawa, E. Manisawa, G. Manisawa, H. Manisawa, J. Manisawa, N. D. Manisawa, P. Manisawa, W. Manisawa, I. Upina Ufipa 1991a Randi Barndon, Bilham Kimathi, (field assistant and interpreter), Randi Haaland (supervisor) Smelters and smiths: D. Kakwaya, D. Kakwaya, H.D. Kakwaya, J. Kaliela, N.J. Kanyama, F. Makumi, S. Malimbo, L. Masilanga, A. Mwanisawa, E. Manisawa, G. Manisawa, H. Manisawa, M. Manisawa, P. Manisawa, W. Manisawa, I. Upina, P. Wangao Ufipa 1991b Randi Barndon Billham Kimathi (field assistant and interpreter) Smelters and smiths: D. Kakwaya, D. Kakwaya, H.D. Kakwaya, J.D. Kakwaya, J. Kaliela, N.J. Kanyama, F. Makumi, S. Malimbo, L. Masilanga, L Mtalimbo, A. Mtukwa, C. Mtukwa, A. Mwanisawa, E. Manisawa, G. Manisawa, H. Manisawa, M. Manisawa, P. Manisawa, W. Manisawa, I. Upina, P. Wangao

159

Randi Barndon Appendix F Pangwa medicines used in our 1995 re-enactment

your way. Ntaanda is a light and we need light when we work’; unnkulo: a piece of a red powder prepared from a tree root. According to my informants it is not found in Ufipa and was part of their inter-tribal trade with the Nyamwanga people. The powder was in the past used as facial decorations symbolising their occupation and they coloured their foreheads with this powder while they worked with the furnace and the day they lighted the fire on the natural draft furnace. The master smelter and his closest associates painted their foreheads also when sorting the content of the basket. A small piece of the powder was laid down in the furnace foundation trench; ulupeembe: a small black horn from unidentified animal; impoombo: a gazelle hoof; kantalala: a piece of an elephant ear. I was explained that when the elephant sleeps the ears are moving. Like the elephants ears the work inside the high shaft furnace will never stop even if the smelting is at night. Elephant bones are also part of the substances; twiga: a kiswahili word: hair from the neck of a giraffe. The master told me: ‘When the giraffe is out on the grassland and the wind is blowing you can see the vibrations in his hair, we the asiluungu must work in the same manner, with a light hand on our work’; kalulu: roots from plants with yellow flowers called kalulu. These flowers or grasses are used when making fire with wooden drill sticks. Kalulu also means hare in iciFipa and the plant has been given the same name since it is in between the dried yellow flowers that the hare prefers to stay. The hare is one of the human-like or a personified creature in Fipa stories (Willis 1978). It is thought of as an intelligent animal. I was told: ‘A hare will never return empty handed’. Informants claimed the flowers are red but according to our colour classifications they are yellow; impala: a piece of skin from a gazelle. I was told: ‘The gazelles are always together, when you see them out on the savannah they are never alone; when we add a bit of this skin we will have a lot of iron, and the iron will come out together in one piece’; dik-dik: a piece of porcupine skin or bone. A Fipa legend tells; ‘One day some intruders arrived in Ufipa to steal cattle from the farmers. But the Fipa were good hunters and when the thieves saw how the Fipa hunters were able to catch a dik-dik who was standing on the other side of a termite mound, they fled Ufipa. Hunters who could hit an animal standing behind a termite mound they had never even heard of in their own country. The Fipa could from that day keep their cattle unprotected’; inntankala: unidentified bones; insweela: antelope skin; inntaamba: barks from various medicinal trees; innconguusa: creepers with white juice, which according to medicine men in general prevent epidemics and cure sickness. Some of these creepers are also collected before each smelt and specifically placed in front of the western door on the natural draft furnace and in a small

Contents in the ing’anjo pot: kiwalasha:- leaves from the cleansing plant and a whiteish juice; imono:- beans imono nyamwaha, oil from beans; mukoda:- leaves from a small scrub; ing’apulya:- yellow flowers used for making fire (and similar to the Fipa kalulu plant used for the same purpose); lumbange:- leaves from a female fertility medicine; mukurumuko: long soft creepers used as a male fertility medicine; ilikoniasuli:plant with red fruits, a symbol of female fertility’ red earth from crossroad: symbol of male and female fertility Pangwa smelting medicines recorded by Stirnimann (1976) In the ing’anjo pot the mixture, kivika, consisted of: ingolomoxo: creepers with white juice, 30 cm long sticks are used as phallic symbols ilitupa: plant with white jiuces and a symbol of semen leaves from a rare plant, symbol of male fertility ilikonyasude: a small plant with redfruits , a symbol of menstruationblood ilixuganguluvi: a fertilitymedicine alsoi used on the fields ans garden imono nyamvaha: beans as a symbol of ferlitity and perhaps more like the Fipa magical substances (see below): beeswax: a symbol of male and female fertility, earth from a crossroad red earth as symbol of blood, earth from the opening of a mouldwarps nest: a symbol of female and male fertility Fipa magical substances in the intaangala basket in reenactment 1990. The content listed below is based on my initial study of Fipa iron working (Barndon 1992:91-95).2 ntaanda: a piece of rock crystal representing a piece of a star. The Umwaami who possessed the first intaangala immediately knew that ntaanda was to help him produce hoes and should be used on the furnace, iluungu, to give it light. It was a part of the peephole that allowed the smelters to control the smelting inside the furnace chamber. A piece of ntaanda was always placed inside the furnace wall. The smelters said: ‘when you walk out in the night under the stars they light up 2

Cf. lists of contents in Roberts 1949; Wembah-Rashid 1973, and see Willis 1978b for list of contents in a Fipa doctors magical basket.

160

Appendices shells as containers for medicines to treat eye diseases. In the smelting context the shells connotes the importance of secrecy; unnckoko: feathers and bones from a cockerel, used against threats of witchcraft; ikala: a piece of old charcoal. The Fipa said: ‘The present burning in the furnace will be strong and powerful as the one when this charcoal was used’. A piece of charcoal was traditionally also placed under the master smelters bed, preventing his wife from being adulterous; inteendo: a piece of an old cooking pot; iciloowa: a piece of old hearth stones; unnembo: birds feathers, also used during wedding ceremonies; ulusaambo: small threads from a cloth made in Rukwa; ifuupa: various unidentified bones; ilepolasi: a red porous powder, sprinkled on the magical substances.

pit in the smaller refining furnace. I was told the true meaning and function of the inncunguusa is only known to the master smelter and his closest assistants. The smelters are generally restricted from sexual relations during the smelting process but by applications of innconguusa the master smelter and his assistant are freed from these taboos; ululu: a piece of pure iron from a previous smelt. The master told me they have to put a piece of iron in the ifingila preventing the intaangala basket from bewitching somebody; unnsuku: special pieces of bark which were always kept in the intaangala, like bark from the Unnsuku tree which has red edible fruits; incalaango: pieces of bones from the neck of a lion. The lion is seen as a fierce animal. Pieces of bone are laid both at the furnace base, in the mouth of a sacrificed cockerel and in the smaller refining furnace; ulwiili: fresh water shells. Medicine men used these

161

Randi Barndon

162

Bibliography Barndon, R. 1992. Traditional iron working among the Fipa - An ethnoarchaeological study from South-western Tanzania. Unpublished Cand. philol. thesis, University of Bergen, Norway.

Agrosah, E. K. 1990. Ethnoarchaeology: the search for a self-corrective approach to the study of past human behaviour. African Archaeological Review, 8:189-202. Andah, B. W. 1995. European encumbrance to the development of relevant theory in African archaeology, in Ucko, P. (ed.). Theory in Archaeology, A World Perspective. 96-108, London and New York: Routledge.

Barndon, R. 1995. Report on Pangwa iron technology. Report submitted to Tanzania Commission for Science and Technology, Report on file, Dar es Salaam, Tanzania. Barndon, R. 1996a. Fipa iron working and it’s technological style, in Schmidt, P. (ed.). The Culture of African Iron Working. 58-74, Gainsville: University Press of Florida.

Andrén, A. 1997. Mellan ting och text. Brutus Östlings Bokforlag Symposium, Stockholm: Stehag. Ani, M. 1996 (1994). Yurugu. An African-centred critique of European culture, thought and behaviour. Eritrea: Africa World Press.

Barndon, R. 1996b. Mental and Material Aspects of Iron Working - A Cultural Comparative Perspective. Proceedings 10th. Pan African Associations of Prehistoric and Related Studies. Harare, Zimbabwe.

Appadurai, A. 1986. The Social Life of Things, Commodities in Cultural Perspective. Cambridge: Cambridge University Press.

Barndon, R. 1999a. Iron Working and social control: the use of anthropomorphic symbols in recent and past East African contexts. Kvinner i arkeologi i Norge, 22-23:5977.

Appadurai, A.1988. Putting Hierarchy in Its Place. Cultural Anthropology, 3:36-49. Aremu, D. A. 1999. Preservation of Ampara and Delimiri Ironworking sites: towards the promotion of Nigeria’s cultural heritage. African Archaeological Review, 16(4):199-210.

Barndon, R. 1999b Etnoarkeologien og historier fra virkeligheten rett før nåtiden. Primitive Tider, 2:70-82. Barth, F. 1974. On Responsibility and Humanity: Calling a Colleague to account. Current Anthropology, 15:100103 .

Atherton, J. H. 1983. Ethnoarchaeology in Africa. African Archaeological Review, 1:75-105.

Barth, F. 1987. Cosmologies in the Making. A generative Approach to Cultural Variation in Inner New Guinea. Cambridge Studies in Social Anthropology, Cambridge: Cambridge University Press.

Augé, M. 1998. A sense for the other. The timeliness and relevance of anthropology. Mestizo Spaces, Stanford: Stanford University Press. Avery, D. H., N. J.van der Merwe and S. Saitowitz. 1988. The metallurgy of the iron bloomery process in Africa, in Maddin, R. (ed). The beginning of the use of metals and alloys. Cambridge: Mit Press.

Beidelman, T. O. 1968. Fipa Symbolism. Man, 3:308310. Beidelman, T. O. 1980. Women and men in two East African Societies, in Karp, I. and C. Bird (eds.). Explorations in African Systems of Thought. Washington and London: Smithsonian University Press.

Avery, D. and P. Schmidt. 1996a. Preheating, practice or illusion? in Schmidt, P. (ed.). The Culture and Technology in African Iron Production. 267-77, Gainsville: University of Florida Press.

Beidelman, T. O. 1993. Moral Imagination in Kaguru Modes of Thought. Washington: Smithsonian University Press.

Avery, D. and P. Schmidt. 1996b. Use of preheated air in ancient and recent African iron smelting furnaces; a reply to Rheder, in Schmidt, P. (ed.). The Culture and Technology in African Iron Production. 240-47, Gainsville: University of Florida Press.

Bekaert, S. 1998. Multiple levels of meaning and the tension of consciousness. Archaeological Dialogues, 5(1):6-29.

Barley, N. 1994. Smashing Pots, Feats of Clay from Africa. London: British Museums Press.

Bell, C. 1992. Ritual theory, Ritual practice. Oxford. Bell, D. H. A. 1945. Three tales: translated from Chifipa by D. H. A. Bell. Tanganyika Notes and Records, 20:6164.

Barndon, R. 1991. Fieldwork among the Fipa. Report submitted to Tanzania Commission for Science and Technology, Report on file, Dar es Salaam. 163

Berglund, A. I. 1991. Fertility as a Mode of Thought, in Jacobsen-Widding, A. and W. Van Beek (eds.). The Creative Communion. Stockholm, Uppsala: Studies in Cultural Anthropology, Vol. 15.

Budd, P. and Taylor, T. 1995. The faerie smith meets the bronze industry: magic versus science in the interpretation of prehistoric metal-making. World Archaeology, 27(1):133-43.

Berlin, B. and P. Kay. 1969. Basic Colour Terms, their universality and evolution. Berkley: University of California Press.

Buchli, V. A. 1995. Interpreting material culture, in Hodder, I. (et als.). Interpreting Archaeology, Finding meaning in the past. 181-193, London and New York: Routledge.

Bernard, F. O. 1962. Two types of smelting furnaces on Ziwa Farm (Inyanga). South African Archaeological Bulletin, 17:235-236.

Bugliarello, G. and Doner, D. B. 1979. Preface, in Bugliarello, G. and D. B. Doner (eds.). The History and Philosophy of Technology. Illinois: University of Illinois Press.

Bernstein, R. J. 1983. Beyond Objectivism and Relativism, Science, Hermeneutics and Praxis. 28:21725, Philadelphia: University of Pennsylvania Press.

Burton, R. F. 1889 (1961). The Lake Regions of Central Africa. New York: Reprinted, 2 vols.

Binford, L. 1962. Archaeology as Anthropology. American Antiquity, 28:217-225.

Cancel, R. 1989. Allegorical Speculation in an Oral Society. The Tabwa narrative tradition. California: University of California Press.

Binford, L. 1988. (1983). In Pursuit of the Past. Decoding the Archaeological Record. London: Thames and Hudson.

Caplan, P. 1995 (1987). Introduction, in Caplan, P. (ed.). The Cultural Construction of Sexuality. 1-31, London: Routledge.

Bloch, M. 1992. Prey into hunter. Lewis Henry Morgan lectures, Cambridge: Cambridge University Press.

Caplan, P. 1995 (1987). Celibacy as a soulution? Mahatma Ghandi and Brahachraya, in Caplan, P. (ed.). The Cultural Construction of Sexuality. 271-295, London: Routledge.

Boast, R. 1997. A small company of actors, a critique of style. in Material Culture, 2(2):173-198. Bourdieu, P. 1973. The Berber house, in Douglas, M. (ed.). Rules and Meaning. London: Penguin Modern Sociology.

Carsten, J. and S. Hugh-Jones. 1996 (1995). Introduction, in Carsten, J. and S. Hugh-Jones (eds.). About the House. Cambridge: Cambridge University Press.

Bourdieu, P. 1977. Outline of a theory of practice. Cambridge: Cambridge University Press.

Celis, G. R. 1989. La métallurgie traditionelle au Burundi, au Rwanda et au Buha. Anthropos, (84):25-46.

Bourdieu, P. 1995 (1979). Distinksjonen. Oslo: Pax. Chaplin, J. H. 1961. Notes on traditional smelting in Northern Rhodesia. South African Archaeological Bulletin, 16:53-60.

Bornhardt, W. 1900. Zur Oberflächengestaltung und Geologie Deutch-Ostafrikas. Berlin.

Brock, B. P. 1963. The Nyiha of Mbozi, Tanzania Notes and Records, 65:1-30.

Childs, S. T. 1991a. Iron as Utility or Expression: Reforging Function in Africa, in Ehrenreich, R. (ed.). Metals in Society, Theory Beyond Analysis. Philadelphia: MASCA Research Papers in Science and Archaeology, 8(2): 57-67.

Brock, B. P. 1968. The Nyiha (of Mbozi), in Roberts, A. D. (ed.). Tanzania before 1900. Nairobi: East African Publishing House.

Childs, S. T. 1991b. Technology, Styles and Iron Smelting Furnaces in Bantu Speaking Africa. Journal of Anthropological Archaeology, 10: 332-359.

Brock, B. P. and W. Brock. 1965. Iron Working among the Nyiha of South-western Tanzania. South African Archaeological Bulletin, 20:97-100.

Childs, S. T. and W.J. Dewey, 1996. Forging Symbolic Meaning in Zaire and Zimbabwe, in Schmidt, P. R. (ed.). The Culture and Technology of Iron Production. 145171, Gainesville: University Press Florida.

Brelsford, V. W. 1949. Rituals and Medicines of Chisinga Iron Workers. Man, 49:27-29.

Brown, J. 1995. Traditional metalworking in Kenya. Oxbow Monographs, 44. Cambridge:Cambridge Monographs in African Archaeology.

Childs, S. T. and Killick, D. 1993. Indigenous African Metallurgy: Nature and Culture. Annual Review of Anthropology, 22:317-37. 164

Childs, S. T. and Schmidt, P. R. 1985. Experimental iron smelting: the genesis of a hypothesis with implications for African prehistory and history, in Haaland, R. and P. Shinnie (eds.). African Iron Working, Ancient and Traditional.121-141, Oslo: Norwegian University Press.

Culwick. G. M. 1935. Pottery among the Wabena of Ulanga, Tanganyika Territory. Man, 35: 165-169.

Clark, J. D. 1974. Kalambo Falls Prehistoric Site. Vol. II, Cambridge: Cambridge University Press.

Damm, C. 1998. Forhistoriske ritualer: En diskussion omkring mening og handling, Religion og materiell kultur. Christensen, L. and S. Sveen.(eds.). Aarhus: Aarhus Universitetsforlag.

Culwick, A. T. and G. M. Culwick. 1967. The functions of bridewealth in Ubena of the Rivers. Africa, 2:140-159.

Clifford, J. 1986. On ethnographic allegory, in Clifford, J. and G. E. Marcus (eds.). Writing Culture, The poetics and the politics of ethnography. Berkley: University of California Press.

Dantz, D. 1900. Ergebnisse der Geologischen Expedition des Bergassessors, Mittheilungen von Forschungsreisenden und Gelehrten aus dem deutchen Schutzgebieten, xiii:39-44.

Clifford, J. and G. E. Marcus (eds.). 1986. Writing Culture, The Poetics and Politics of Ethnography. Berkley: University of California Press.

David, N. 1982. Prehistory and Historical Linguistics in Central Africa: Points of Contact, in Ehret, C. and M. Posnansky (eds.). Archaeological and Linguistic Reconstructions of African History. Berkley: California University Press.

Cline, W. 1937. Mining and Metallurgy in Negro Africa. Menasha, Wisconsin: George Banta Publishing Co. Cohen, A. P. 1995(1994). Self Consciousness, An alternative anthropology of identity. London: Routledge. Collett. D. P. 1985. The Spread of E. I. A. Communities in Eastern and Southern Africa. PhD. thesis, unpublished, Cambridge: University of Cambridge.

David, N. 1992. Integrating ethnoarchaeology; A Subtle Realist Perspective. Journal of Anthropological Archaeology, 11:330-359. David, N. 1994. African Archaeology: Twix cancer and Capricorn. Journal of African History, 35:125-132.

Collett. D. P. 1993. Metaphors and Representations associated with pre-colonial iron smelting in eastern and southern Africa, in The Archaeology of Africa, Food, Metals and Towns. Shaw, T. et. als., One World Archaeology, London: Routledge.

David, N. 2001 (in press). Lost in the third Hermeneutic? Theory and methodology, objects and representations in the ethnoarchaeology of African metallurgy. Paper presented for Table Ronde Internationale ‘L’Afrique et le bassin Méditerrané en aux origines de la métallurgie de fer’, Geneva 4-7-June 1999, Mediterranean Archaeology.

Comaroff, J. and Comaroff, J. 1992. Ethnography and the historical imagination. Oxford: Westview Press.

David, N., Sterner, J. and K. Gavua. 1988. Why Pots are Decorated. Current Anthropology, 23(3):365-389.

Conkey, M. W. 1990. Experimenting with style in archaeology: some historical and theoretical issues, in Conkey, M. W. and C. A. Hastorf (eds.). The uses of style in archaeology. Cambridge: Cambridge University Press. Connah, G. 1989 (1987). African Cambridge: Cambridge University Press. Connerton, P. 1989. How societies Cambridge: Cambridge University Press.

David, N., et.al. 1989. Between Bloomery and Blast furnace: Mafa Iron-smelting ethnology in North Cameroon. African Archaeological Review, 7:183-208.

Civilizations.

David, N. and I. Robertson, 1996. Competition and change in two traditional African iron industries, in Schmidt, P. (ed.). The Culture and Technology of African Iron Technology. 128-145, Gainsville: University Press of Florida.

remember.

Cooke, C. K. 1966. An Iron-Smelting Site at Matapo Hills, Southern Rhodesia. South African Archaeological Bulletin, 14:118-120.

Davis, C. O. 2000. Death in abeyance. Illness and therapy among the Tabwa of Central Africa. Edinburgh: Edinburgh University Press.

Cory, H. 1949. The ingredients of magic medicines, Africa, 1:13-32.

Davis, K. 1997. Embodied Practices, Feminist Perspectives on the Body. London: Sage Publications.

Crapanzano, V. 1986. Hermes’ Dilemma. The Masking of Subversion in Ethnographic Description, in Clifford, J. and G. E. Marcus (eds.). Writing Culture, The Poetics and Politics of Writing. Berkley: University of California Press. Culwick, A. T. and M. Culwick. 1935. Ubena of the Rivers. London: George Allen and Unwin LTD.

Davison, S. and P. N. Mosley. 1988. Iron-Smelting in the upper North Rukuru Basin of Northern Malawi. Azania, XXIII:57-101.

165

Fernandez, J. 1980. Edification by Puzzlement, in Karp, I. and C. S. Bird (eds.). Explorations in African Systems of Thought. Washington and London: Smithsonian Institution Press.

Davison, S. 1993. Saltmaking in Early Malawi. Azania, XXVIII:7-44. Derricourt, R. 1985. People of the Lakes. Archaeological Studies in Northern Zambia. Zambian Papers, 13.

Fernandez, J. 1986. Persuasions and Performances. Bloomington: Indiana University Press.

Dietler, M. and I. Herbich. 1989. Tich Matek, The technology of Luo pottery production and the definition of ceramic style. World Archaeology, 21:148-164.

Fernandez, J. (ed.). 1991. Beyond Metaphor, The Theory of Tropes in Anthropology. Stanford: Stanford University Press.

Dietler, M. and I. Herbich. 1998. Habitus, Techniques, Style: An Integrated Approach to the social understanding of material culture and boundaries, in Stark, M. T. (ed.). The Archaeology of Social Boundaries. 232-263, Washington: Smithsonian Institute Press.

Fletcher, R. 1989. The messages of material behaviour: a preliminary discussion of non-verbal meaning, in Hodder, I. (ed.). The meaning of things. One World Archaeology, London: Harper Collins Academic.

Dobres, M. A. 1995. Gender and prehistoric technology: on the social agency of technical strategies. World Archaeology, 27(1):25-49.

Flikke, R. 1994. The Past in the Present, a semiotic exploration of urban Zulu Zionism in Durban, South Africa. Unpublished Cand. polit. thesis, Department and Museum of Anthropology, University of Oslo.

Dobres, M. A. 2000. Technology and social agency. Oxford: Blackwell Publishers Ltd.

Fowler, I. 1990. Babungo: A study of iron production, trade and power in a nineteenth century Ndop Plain Chiefdom (Cameroon). Ph.D. University College, London University (unpublished).

Douglas, M. 1966. Purity and Danger. London: Penguin Books. Douglas, M. 1973 (1970). Natural Symbols. London: Routledge.

Fülleborn, F. 1906. Das Deutche Njassa - und RuvumaGebiet, Land und Leute. Berlin: Dietrich Reimer.

Dreyfus, H. L. 1996. The current relevance of MerleauPonty’s phenomenology of embodiment. The Electronic Journal of Analytical Philosophy, 1-13.

Geertz, C. 1979. From the Native’s point of view: On the nature of anthropological understanding, in Rabinow, P. and W. M. Sullivan (eds.). Interpretative Social Science, A Reader. Berkley: University of California Press.

Eggert, M. 1977. Prehistory, archaeology and the problem of ethno-cognition. Anthrophos, 1:242-55.

Geertz, C. 1973. The interpretation of Cultures. New York: Basic Books.

Eggert, M. K. H. 1987. On the alleged complexity of Early and Recent iron smelting in Africa. Further Comments on the Preheating Hypothesis. Journal of Field Archaeology, 124:377-382.

Geertz, C. 1984. Distinguished Lecture: Anti Antirelativism. American Anthropologist, 86:263-278.

Eliade, M. 1978 (1958). The Forge and The Crucible, The Origins and Structures of Alchemy. Chicago and London: University of Chicago Press.

Geertz, C. 1995. After the Fact. Two Countries, Four Decades. One Anthropologist. Cambridge, Massachusetts: Harward University Press.

Eriksen, T. 1995 (1988). Small Places, Large Issues. London: Pluto Press.

Gell, A. 1988. Technology and Magic. Anthropology Today, 4:6-9.

Espelund, A. 1997. The «Evenstad» Process Description. Excavation, Experiment and Metallurgical Evaluation, in L. C. Nørbach (ed.). Early Iron Production - Archaeology, Technology and Experiments. Technical Report, 3: 47-58.

Gellner, E. 1995. Interpretative anthropology, in Hodder, I. Interpreting Archaeology. London and New York: Routledge. Gibbon, G. 1989. Explanation in Archaeology. Oxford: Oxford University Press.

Fagan, B. M. and E. J. Yellen. 1968. Ivuna: ancient saltmaking in southern Tanzania. Azania, 3:1-43.

Giddens, A. 1976. New rules of Sociological Method.

Feldberg, A.1995. Excerpt from Nyakyusa dictionary, Nyakyusa homepage (http://home. sol.no/feldberg/nyakyusa/nya_typology.htm).

Giddens, A. 1979. Central Problems in Social theory. London: MacMillian.

166

Traditional. 50-72, Oslo: Norwegian University Press.

Gilchrist, R. 1994. Gender and material Culture: The Archaeology of Religious women. London: Routledge.

Haaland, R. 1988. The role of Ethnoarchaeology and Experimental archaeology in the interpretation of prehistoric societies, Oslo: Arkeologiske Skrifter. Historisk Museum, 4.

Gilchrist, R. 1999. Gender and Archaeology. London: Routledge. Gilje, N. and Grimen, H. 1992 (1991). Sammfunnsvitenskapens forutsetninger. Bergen: Ariadne.

Hastrup, K. 1992. Det antroplogiske prosjekt - Om forbløffelse. Copenhagen: Nordisk Forlag.

Gilje, N. 1993. Anomalier i moderne vitenskapsteori. in Gilje, N. and H. Grimen (eds.), Kompendium i Almen Vitenskapsteori for Dr. Polit. og Dr. art.-graden, 1-87. Bergen.

Hastrup, K. and P. Hervik 1994. Social Experience and Anthropological Knowledge. London:Routledge. Hastrup, K. 1995. A Passage to Anthropology. Between experience and theory. London: Routledge.

Gosselain, O. P. 1992. Technology and Style, potters and pottery among Bafia of Cameroon. Man, 27:559-586.

Haudricourt, A. G. 1964. La Technologie, Science Humaine. La Pensée, 115: 28-35.

Gosselain, O. P. 1993. From clay to pottery, with style: 1990-1992, Fieldwork in Cameroon. Nyame Akuma, 39:2-7.

Heald, S. 1995. The power of sex: some reflections on the Caldwells’ ‘African sexuality’ thesis. Africa, 65(4):489505.

Gosselain, O. P. 1998. Social and technical identity in a crystal ball, in Stark, M. (ed.). The Archaeology of Social Boundaries. Washington: Smithsonian.

Hegmon, M. 1992. Archaeological research on style. Annual Review of Anthropology, 21:517-536.

Gosselain, O. P. 1999. In Pots we trust. Journal of Material Culture, 4(2): 205-230.

Hegmon, M. 1998. Technology, Style, and Social Practices: Archaeological Approaches, in Stark, M (ed.). The Archaeology of Social Boundaries. Washington: Smithsonian.

Gould, R. A. 1978. Explorations in ethnoarchaeology. Albuquerque: University of New Mexico Press.

Heidegger, M. 1977. The Questions Concerning Technology. Cambridge: Garland Publishing.

Gould, R. A. and Watson, P. J. 1982. A dialogue on the meaning and use of analogy in ethnoarchaeological reasoning. Journal of Anthropological Archaeology, 1:355-81.

Herbert, E. W. 1993. Iron, Gender and Power, Rituals of Transformations in African Societies. Bloomington: Indiana.

Gould, R. 1990. Recovering the Past. Albuquerque: University of New Mexico Press.

Heusch, L. de. 1956. Le Symbolisme du Forgeron en Afrique. Reflets du Monde, 57-70, Université Libre de Bruxelles.

Gray, J. 1957. Trading Expeditions from the Coast to Lakes Tanganyika and Victoria before 1857. Tanganyika Notes and Records, 49:226-246.

Heusch, L. de. 1975. What shall we do with the drunken king? Africa, 45(4):363-373.

Greig, R. C. H. 1937. Iron Smelting in Fipa. Tanganyika Notes and Records, 4: 77-81.

Heusch, L. de. 1980. Heat, Physiology and Cosmogony, Rites de Passage among the Thonga, in Karp, I. and C. S. Bird (eds.). Explorations in African Systems of Thought. 27-44, Washington, and London: Smithsonian Institution Press.

Goucher, C. 1981. Iron is iron till it rusts: trade and ecology in the decline of West African iron-smelting. Journal of African History, 22:179-189. Grosz, E. 1987. Notes towards a corporeal feminism. Australian Feminist studies 5 (summer):1-16.

Hodder, I. 1982. Symbols in Action. Cambridge: Cambridge University Press.

Guthrie, M. 1962. Some developments in the prehistory of the Bantu languages. Journal of African History, III(2):273.282.

Hodder, I. 1984. Ethnoarchaeology: Applications of the general theory, in Stjernquist, B. (ed.). Perspective on archaeological theory and methods. Report Series no. 20, London: University of London Press.

Haaland, R. 1985. Iron Production, it’s Socio-Cultural context and Ecological Implications, in Haaland, R. and P. Shinnie (eds.). African Iron Working, Ancient and 167

Hodder, I. 1989(1986). Reading the Past. Cambridge: Cambridge University Press.

Hastrup, K. and P. Hervik (eds.). Social experience and anthropological knowledge. (EASA), London: Routledge.

Hodder, I. 1992. The Present Past. New York: Pica Press.

Kense, F. J. 1983. Traditional African Iron Working. African Occasional Papers, 1. Canada: Department of Archaeology.

Hodder, I. 1999. The Archaeological Process, An Introduction. Oxford: Blackwell.

Kense, F. J. 1985. The initial diffusion of iron to Africa, in Haaland, R. and Shinnie, P. (eds.). African iron working, Ancient and Traditional. Oslo: Norwegian University Press.

Hood, W. F. 1983 (1972). The Aristotelian Versus the Heideggerian approach to the problem of technology, in C. Mitcham and R. Mackey (eds.). Philosophy and Technology, 347-363, London: The Free Press. Hore, E. C. 1892. Tanganyika: Eleven Years in Central Africa. London: Stanford.

Kerr-Cross, M. B. 1890. Geographical notes on the country between lakes Nyassa, Rukwa and Tanganyika. Scottish Geographical Magazine, 6:281-293.

Iliffe, J. 1979. A modern history of Tanganyika. Cambridge: Cambridge University Press.

Killick, D. 1987. Recent Iron Smelting in Central Malawi. Nyame Akuma, 28:27-29.

Ingold, T. 1990. Society, Nature and the Concept of Technology. Archaeological Review from Cambridge, 9(1):5-17.

Killick, D. 1990. Technology in its social setting: Bloomery iron-smelting at Kasungu, Malawi 1860-1940. PhD. dissertation, Ann Arbor: Yale University, U.M.I.

Jacobsen-Widding, A. 1990. Subjective body, objective space - An introduction, in Jacobsen-Widding (ed.). Body and Space. Uppsala: Acta Universitas.

Killick, D. 1996. On Claims For ‘Advanced’ Iron working Technology in Pre colonial Africa, in Schmidt, P. R. (ed.) The Culture and Technology of African Iron Production, 247-266, Gainesville: University Press of Florida.

Janzen, J. M 1992. Ngoma, Discourses of healing in Central and Southern Africa. Berkley: University of California Press.

Kjekshus, H. 1976. Ecology Control and Economic Development in East Africa. London: Heinemann.

Jeffreys, M. D. W. 1952. Some Notes on The Bikom Blacksmiths. Man, 75:51.

Koizumi, M. 1995. Meaning of Conversion: Cosmology, Politics and Moral discourse among the Kinga of South western Tanzania. unpublished PhD. dissertation, Ann Arbor: Yale University.

Johnson, H. 1991. Dilemmas on Development: Forms of peasant resistance among the Fipa in Rukwa Tanzania. Cand. Polit. thesis, University of Olso.

Koponen, J. 1988. People and Production in late Precolonial Tanzania. Monographs for the Finnish Society of Development Studies, 2, Finland: Gummerus Kirjapaino Oy.

Johnson, H. and B. Olsen, 1992. Hermeneutics and archaeology: On the philosophy of contextual archaeology. American Antiquity, 57(3):419-436.

Koponen, J. 1994. Development for Exploitation, German colonial politics in mainland Tanzania, 18841914. Finnish Historical Society, Studia Historica 49, Studien zur Afrikanischen Gescichte 10, Helsinki and Hamburg: Lit Verlag.

Johnson, M. 1987. The Body in the Mind. The Bodily Basis of Meaning, Reason and Imagination. Chigaco: University of Chicago Press. Johnson, M. 1993. Moral imagination, Implications of cognitive science for ethics. Chicago: University of Chicago Press.

Kootz-Kretschmer, E. 1926. Die Safwa. Berlin: Dietrich Reimer.

Johnson, M. 1999. Archaeological Theory, An Introduction. Great Britain: Blackwell Publisher Ltd.

Kriger, C. E. 1999. Pride of Men, Iron Working in the 19th Century West Africa. London: James Curry.

Jones, S. 1997. The Archaeology of Ethnicity. London: Routledge. Keesing, R. M. 1987. Anthropology as Interpretative Quest. Current Anthropology, 28(2):161-176.

Kuhn, T. 1962. The structure of scientific revolutions. Chicago: University of Chicago Press. Kuntz, M. 1935. The iron workers of the Kwanga tribe. Man, 202-204:186.

Kempny, M. and Burszta, W. J. 1994. On the relevance of common sense for anthropological knowledge, in 168

Lecthman, H. and Steinberg, A. 1979. The History of Technology: An Anthropological Point of View, in Bugliarello, G. and D. B. Doner (eds.). The History and Philosophy of Technology. Illinois: University of Illinois Press.

Lakoff, G. 1987. Women, Fire and Dangerous Things, What categories reveal about the mind. Chicago: Chicago University Press. Lakoff, G. and Johnson, M. 1980. Metaphors we live by. Chicago: University of Chicago Press.

Leeuw van der, S. 1993. Giving the potter a choice, in Lemonnier, P. (ed.). Technological Choices, Transformations in Material Culture since the Neolithic. Oxford: Routledge.

Lakoff, G. and Johnson. M. 1999. Philosophy in the Flesh. New York: Basic Books. Lane, P. 1986. Past practices in the ritual present: An example from the Welsh Bronze Age. Cambrigde: Archaeological Review from Cambridge.

Lemonnier, P. 1986. The study of material culture today: Towards an Anthropology of Technical Systems. Journal of Anthropological Archaeology, 5:147-86.

Lane, P. J. 1996a. The use and abuse of ethnography in Iron Age studies of Southern Africa. Azania, XXIXXXX, (1994–1995):15-27.

Lemonnier, P. 1989. Bark capes, Arrowheads and Concorde: On social representations of technology, in Hodder, I. (ed.). The Meaning of Things, Material Culture and Symbolic Expressions. One World Archaeology, 6, London: Harper Collins.

Lane, P. J. 1996b. Rethinking ethnoarchaeology, In Soper, R. and Pwiti, G. (eds.). Aspects of African Archaeology. 727-732, Harare: University of Zimbabwe Publications.

Lemonnier, P. 1990. Topsy, turvy techniques: Remarks on the Social representations of techniques, in Sinclair, A. and N. Schlanger (eds.). Technologies in the humanities. Archaeological Review from Cambridge, 9(1): 18-27.

Lane, P. J. 1998a. Engendered spaces and bodily practices in the Iron Age of Southern Africa, in S. Kent (ed.). Gender in African Prehistory. 179-203, London: Alta Mira.

Lemonnier, P. 1992. Elements for an Anthropology of Technology. Anthropological Papers, No.88, Museum of Anthropology, Ann Arbour, Michigan: University of Michigan.

Lane, P. J. 1998b. Ethnoarchaeological research - past, present and future directions, in Lane, P. Reid, A. and A. Segobye (eds.). Ditswa Mnung- The archaeology of Botswana. 177-205, Botswana: Pula Press.

Lemonnier, P. 1993. Introduction, in Lemonnier, P. (ed.). Technological Choices. London: Routledge.

Langer, M. M. 1989. Merleau-Ponty’s Phenomenology of Perception, A guide and Commentary. London: MacMillian Press.

Leroi-Gourhan, A. 1993 (1964). Gesture and Speech. Massachusetts: MIT Press.

Latour, B. 1996 (1991). Vi har aldri vært moderne. (Transl. by R. Myklebust), Oslo: Spartacus.

Lévi-Strauss, C. 1963 (1962). Totemism. Boston: Penguin Books.

Latour, B. 1999. Pandora’s Hope, Essays on the reality of science studies. Cambridge, Massachusetts: Harvard University Press.

Lévi-Strauss, C. 1994 (1966). The Savage Mind. Chicago and London: The University of Chicago Press. Lévi-Strauss, C. 1970. The Raw and The Cooked; Introduction to a Science of Mythology. Jonathan Cape.

Lawrence, D. L and S. M. Low. 1990. The built environment and spatial form. Annual Review of Anthropology, 19:453-505.

Lévi-Strauss. C. 1980 (1978). Myth and Meaning. London: Routledge & Kegan.

Leach, E. 1976. Culture and Communication. Cambridge: Cambridge University Press.

Livingstone, D. 1874. The Last Journals of David Livingstone in Central Africa. by H. Waller, 2 Vols. London: Murray.

Lechaptois, A. 1913. Aux Rives du Tanganyika. Algiers: Maison Carré.

Livingstone, D. 1872. (1990). Letters and Documents, 1841 - 1872. London: Holmes, T. James Currey. MacCormack, C and M. Strathern. 1980. Nature, Culture and Gender. Cambridge: Cambridge University Press.

Lecthman, H. 1977. Style in Technology - Some Early Thoughts, in Lecthman, H. and R. Merill (eds.). Material Culture; Styles, Organisations, and Dynamics of Technology. New York: West Publishing.

169

Merleau-Ponty, M. 1989 (1962). The Phenomenology of Perception. London: Routledge.

MacEachern, S. 1995. Scale and Stylistic Variability: The case of the Mandara Mountains. paper presented in the Symposium ‘Social boundaries, Technical Choices and Material Culture patterning’. 60th Annual Meeting of the Society for American Archaeology.

Merwe, N. J. van der. 1980. The advent of iron in Africa, in Wertime, T. and J.D. Muhly (eds.). The Coming of Iron. New Haven: Yale University Press.

MacEachern, S. 1996. Foreign Countries: The Development of Ethnoarchaeology in Sub-Saharan Africa. Journal of World Prehistory, 10(3):243-304.

Merwe, J. van der and D. H. Avery. 1982. Pathways to steel. American Scientist, 70:146-55. Merwe, N. J. van der and D. H. Avery. 1987. Science and Magic in African Technology: Traditional iron smelting in Malawi. Africa, 57(2):143-172.

MacEachern, S. 1998. Scale, style and cultural variation: technological traditions in the Northern Mandara Mountains, in Stark, M. T. (ed.). The Archaeology of Social Boundaries. Washington and London: Smithsonian Institution Press.

Mgina, A. N. J. not dated. Matumizi Ya Chuma Upangwa. Unpublished document.

MacLean, R. 1996. The social impact of the beginning of iron technology in the western Lake Victoria basin: A district case study. Unpublished Ph.D. dissertation, Department of Archaeology, Cambridge: University of Cambridge.

Miller, D. E. and N. J. van der Merwe. 1994. Early Metal working in Sub-Saharan Africa: A Review of recent Research. Journal of African History, 35:1-36. Mitcham, C. 1979. Philosophy and the history of technology, in Bugliarello, G. and D. B. Doner (eds.). The History and Philosophy of Technology. Chicago: University of Illinois Press.

MacLean, R. 1998. Gendered technologies and gendered Activities in the Interlacustrine Early Iron Age, in Kent, S. (ed.). Gender in African prehistory. London: Alta Mira.

Misango, K. 1996. Ceramics from the Upemba Depression, A diachronic Study, in Arnoldi, M. J. et al. African Material Culture. Bloomington and Indianapolis: Indiana University Press.

Malinowski, B. 1935. Coral Gardens and their magic, 2 Vols. London: George Allan and Unwin Ltd. Malinowksi, B. 1948. Magic, Science, and Religion and other Essays. Illinois: Glencloe Free Press.

Moore, H. L. 1990. Paul Riceur: Action, Meaning and Text, in Tilley, C. (ed.). Reading Material Culture. 85121, Oxford: Blackwell.

Mapunda, B. B. 1995. An Archaeological View of the History and Variation of Iron Working in South-western Tanzania. Unpublished. Ph.D. dissertation, Gainesville: University of Florida.

Moore, H. L. 1986. Space, Text and Gender. Cambridge: Cambridge University Press.

Maret, P. 1980. Ceux qui jouent avec la feu: La place du Forgeron en Afrique Centrale. Africa, 50(3):263-79.

Moore, H. L. 1994. A Passion for Difference. Oxford: Polity Press.

Maret, P. 1985. The Smith’s Myth and the Origin of Leadership in Central Africa, in Haaland, R. and P. Shinnie. (eds.). African Iron Working, Ancient and Traditional. Oslo: Universitetsforaget.

Moore, H. L. 1995. The problems of origins: PostStructuralism and Beyond, in Hodder, I. et. al. Interpreting Archaeology, Finding Meaning in the Past. London and New York: Routledge.

Maret, P. and F. Nsuka 1977. History of Bantu Metallurgy: Some linguistic aspects. History in Africa, 4:43-65.

Mumford, B. W. 1930. Mlangali School. Africa, III:1-3. Mumford, B. W. 1936. The Hehe-Bena-Sangu peoples of East Africa. American Anthropologists, 36:203-222.

Marwick, M. G. 1965. Sorcery in its social setting. Manchester: Manchester University Press.

Murdock, G. P. 1959. Africa: It’s Peoples and Their Culture History. New York: McGraw Hill.

Mauss, M. 1979 (1935). Sociology and Psychology. Trans. Brewster B, London: Routledge & Kegan.

Mutoro, H. W. 1998. Precolonial trading systems of the East African interior, in Connah, G. (ed.). Transformations in Africa. 186-202, London: Leichester University Press.

Mauss, M. 1972 (1950). A general theory of magic. London: Routledge & Keagan.

170

Perceul, R. W. and I. Hodder, 1996. Contemporary Archaeology in Theory. Great Britain: Blackwell.

Nedham, J. 1980. The evolution of iron and steel technology in East and Southeast Asia, in Wertime, T. A and J. D. Muhly (eds.). The Coming of The Iron Age. 507-542, New Haven, Yale University Press.

Pfaffenberger, B. 1988. Fetished Objects and Humanised Nature: Towards an Anthropology of Technology. Man, 23:236-52.

Newman, J. L. 1995. The Peopling of Africa, A geographic Interpretation. New Haven: Yale University Press.

Pfaffenberger, B. 1992. Social Anthropology of Technology. Annual Review of Anthropology, 21: 491516.

Noten, van. F. and J. Raymakers, 1988. Early Iron Smelting in Central Africa. Scientific American, 84-91.

Phillipson, D. W. 1968. Chewa, Leya and Lala Iron Smelting Furnaces. South African Archaeological Bulletin, 23:102-113.

Odner, K. 2000. Tradition and Transmission. Bergen: Norse Publications.

Phillipson, D. W. 1977. The Later Prehistory of Eastern and Southern Africa. London: Heinemann.

Olsen, B. 1987. Arkeologi, tekst, samfunn. Fragmenter til en post-prosesuell arkeologi. Stensilserie B, historie/arkeologi, 24, Tromsø: Universitetet i Tromsø.

Phillipson, D. W. 1988 (1993). African Archaeology. Cambridge: Cambridge University Press.

Olsen, B. 1991. Metropolis and satellites in archaeology: on power and asymmetry in global archaeological discourse. In Preucel, R.W. (ed.). Processual and Postprocessual Archaeologists; Multiple ways of Knowing the past. Occasional Paper 10:211-24, Carbondale: Southern Illinois University Press.

Plog, S. 1983. Change and conservatism in pottery production systems, in S. E. Van der Leeuw & A. C. Pritchard (eds.). The many dimensions of pottery. Amsterdam: University ver Amsterdam. Pole, L.M. 1985. Furnace design and the smelting operation: a survey of written reports of iron smelting in West Africa, in Haaland, R. and P. Shinnie (eds.). African iron working, Ancient and traditional. 142-164, Oslo: Norwegian University Press.

Olsen, B. 1994. Roland Barthes: From sign to text, in Tilley, C. (ed.). Reading Material Culture. London: Blackwell. Olsen, B., A. Hesjedal, and I. Storli. 1993. Camera Archaeologica. Tromsø: Tromsø Museum skrifter: XX III.

Popplewell, G. D. 1937. Notes on the Fipa. Tanganyika Notes and Records, 3:99-105.

Olsen, B. 1997. Fra ting til tekst, Oslo: Norwegian University Press. Olsen, B. 2000. Arkeologi og materiell unpublished manuscript, University of Tromsø

Potter, P. B. Jr. 1991. Self-Reflection in Archaeology, in Perceul, R. W. (ed.). Processual and Postprocessual Archaeology, Multiple Ways of Knowing the Past. Occasional Paper, 10:225-235. Southern Illinois: Centre for Archaeological Investigations, University at Carbondale.

kultur,

Parker Pearson, M. 1995. Tombs and territories, material culture and multiple interpretations, in Hodder, I. et als (eds). Interpreting Archaeology. London: Routledge.

Reefe, T. Q. 1981. The Rainbow and the Kings. A history of the Luba Empire to 1891. Berkley and Los Angeles: University of California Press.

Park, M. 1899. The life and Travels of Mungo Park. With supplementary details of the results of recent discovery in Africa, Edinburgh: W. P. Nimmo, Hay & Mitchell.

Reid, A. and R. MacLean, 1995. Symbolism and the social contexts of iron production in Karagwe. World Archaeology, 27(1):144-61.

Perceul, R. W. 1991a. Introduction, in Processual and Postprocessual Archaeologies, Multiple ways of Knowing the Past. Occasional Paper 10:1-14, Centre for Archaeological Investigations, Southern Illinois: University at Carbondale.

Rheder, J. E. 1987. Archaeomaterials, 2:47-58.

Natural

draft

furnaces.

Rheder, J. E. 1996. Use of preheated air in primitive furnaces. Comment on vies of Avery and Schmidt, in Schmidt, P. (ed.) The culture and technology of African iron production. Gainsville: University Press of Florida.

Perceul, R. W. 1991b. The Philosophy of Archaeology, in Perceul, R. W. (ed.).Processual and Postprocessual Archaeologies, Multiple Ways of Knowing the Past. Occasional Paper, 10:17-29. Centre for Archaeological Investigations, Southern Illinois University at Carbondale.

171

Rowlands, M. and Warnier, J. P. 1996. Magical Iron Technology in the Cameroon Grassfields, in Arnoldi, M. J. et. als.(eds). African Material Culture. Bloomington and Indianapolis: Indiana University Press.

Richards, A.1982 (1956). Chisungu: A girl’s Initiation Ceremony among the Bemba of Zambia. London: Tavistock Pub. Routledge. Ricoer, P. 1997 (1978). The Rule of Metaphor. London: Routledge

Sackett, J. R. 1977. The meaning of style in archaeology: a general model. American Antiquity, 42:369-80.

Rio, K. 1998. Sprialen: Det materielle produksjonsapparatet involvert i fremstillingen av menn og andre mannlige figurer i Vanuatu, Sentral-Melanesia. Primitive Tider, 1:33-45.

Sackett, J. R. 1990. Style and ethnicity in archaeology: the case for isochrestism, in (eds.), M. W. Conkey and C. A. Hastorf (eds.). The uses of style in archaeology. Cambridge: Cambridge University Press.

Rodman, M. C. 1992. Empowering place; Multilocality and Multivocality. American Anthropologist, 94(3):640656.

Sanders, T. 1997. Making Children, Making Chiefs: Gender, Power and Ritual Legitimacy. Africa, 68(2):238262.

Robert, R. P. J. M. 1949. Croyances et Cotumes MagicoReligeuses Des Wafipa. Kipalapala, Tabora Post Office: Tanganyika Mission Press.

Schiffer, M. and J. M. Skibo. 1987. Theory and Experiment in the Study of Technological Change. Current Anthropology, 28(5):595-621.

Roberts, A. D. 1968. Tanzania before 1900. Nairobi: East African Publishing House.

Schiffer, M. B. 1999. The Material Life of Things. London: Routledge.

Roberts, A. F. 1991. Where the king is coming from, in Jacobsen-Widding, A. (ed.). Body and Space. Acta Univesitatis Uppsalalensis, 16, Uppsala: Uppsala Studies in Cultural Anthropology.

Schmidt, J. 1985. Maurice Merleau-Ponty, Between phenomenology and structuralism. Houndmills: MacMillian. Schmidt, P. R. 1978. Historical Archaeology. Contributions in Intercultural and Comparative Studies, 3, Westport, Connecticut: Greenwood Press.

Roberts. A. F. and E. Mauer. 1986. (eds.). The Rising of a new moon, A Century of Tabwa art. Seattle. Robinson, K. R. 1961. Two iron smelting furnaces from Chisi Native reserve, Southern Rhodesia. South African Archaeological Bulletin, 16:20-22.

Schmidt, P. R. and D. H. Avery. 1978. Complex Iron Smelting and Prehistoric Culture in Tanzania. Science, 201:1085-1089.

Rorty, R 1979. Philosophy and the mirror of nature. Oxford: Blackwell.

Schmidt, P. R. 1981. The Origins of iron smelting in Africa: A complex technology in Tanzania. Research Papers in Anthropology 1, Brown University.

Rorty, R. 1989. Contingency, Irony and Solidarity. Cambridge: Cambridge University Press. .

Schmidt, P. R. 1983. More Evidence for An Advanced Prehistoric Iron Technology in Africa. Journal of Field Archaeology, 10(4):421-34.

Rosemond, C. C. 1943. Iron smelting in the Kahama District. Tanganyika Notes and Records, 16:79-84.

Schmidt, P. R. and T. Childs. 1995. Ancient African Iron Production. American Scientist, 83:524-533.

Rostoker, W. and Bronson, B. 1990. Pre-Industrial iron. Its technology and Ethnology. Pennsylvania: Archeomaterial Monographs. Rowlands, M. J. 1971. The Archaeological interpretation of prehistoric metal working. World Archaeology, 3(2):211-224.

Schmidt, P. R. 1996a. Cultural Representations of African Iron Production. in Schmidt, P. R. (ed.). The Culture and Technology of African Iron Production. Gainesville: University Press of Florida.

Rowlands, A. and Warnier, J. P. 1993. The Magical Production of Iron in the Cameroon Grassfields, in Shaw, T. et. als. (eds). The Archaeology of Africa, Food, Metals and Towns. One World Archaeology, New York/London: Routledge.

Schmidt P. R. 1996b. Reconfiguring the Barongo: Reproductive symbolism and reproduction among a work association of iron smelters, in Schmidt, P. R (ed.). The Culture and technology of African Iron Production. Gainesville: University Press of Florida.

172

Stiles, D. 1977. Ethnoarchaeology: A Discussion of Methods and Applications. Man, 12(1):87-103.

Schmidt, P. R. and B. B. Mapunda 1997. Ideology and the Archaeological Record in Africa: Interpreting Symbolism in Iron Smelting Technology. Journal of Anthropological Archaeology, 16:73-102.

Stirnimann, H. 1967. Zur gesellschaftsordnung und Religion der Pangwa. Anthrophos, 62:800-812.

Schmidt, P. R. 1997. Iron Technology in East Africa. Bloomington, Indianapolis: Indiana University Press.

Stirnimann, H. J. 1969. Iron making in Upangwa, unpublished manuscript (with kind permission from Dr. John Sutton).

Shinnie P. L. 1971. The African Iron Age. Oxford: Oxford University Press.

Stirnimann, H. 1976. Existensgrundlagen und traditionalles Handwerk der Pangwa von SW. Tansania. Studia Ethnographica Friburgensia 4. Freiburg: Universtitätsverlag Freiburg.

Sigaut, F. 1990. Preface en Haudricourt et la technologie, La technologie Science humanaine Recherche d’historie et de Technologie de tecniques. Paris: Edition de la maison des Sciences de l’homme Paris.

Stirnimann, H. 1979. Die Pangwa von SW-Tansania, Soziale Organisation und Riten des Lebens. Studia Ethnographica Friburgensia 7. Freiburg: Universitätsverlag Freiburg.

Sigaut, F. 1994. Technology, in Ingold, T. (ed.). Companion Encyclopaedia of Anthropology. London: Routledge.

Stirnimann, H. 1983. Praktische Grammatik der Pangwasprache (SW.Tansania). Studia Ethnographica Friburgensia 10. Freiburg: Universitätsverlag Freiburg.

Simensen, J. 1996. Afrikas historie. Halden, Norway: Cappelens Akademiske forlag.

Strathern, A. J. 1996. Body Thoughts. Michigan: University of Michigan Press.

Smith, E. and Dale. A. 1968 (1920). The Ila-Speaking Peoples of Northern Rhodesia. 2.Vols. New York: New Hyde Park University Books.

Strathern, A. J. and Lambek (eds.). 1998. Bodies and Persons: comparative perspectives from Africa and Melanesia. Cambridge: Cambridge University Press.

Speth, J. D. 1992. Foreword, in Lemonnier, P. Elements for an Anthropology of Technology. Anthropological papers, Museum of Anthropology, 88. Michigan: University of Michigan Press.

Sutton, J. E. G. 1971a. The Interior of East Africa, in Shinnie P. L. (ed.). The African Iron Age. Oxford: Oxford University Press.

Spiegel. A. D. 1994. Struggling with tradition in South Africa: the multivocality of images of the past, in Bond, G. C. and A. Gilliam (eds.), Social Constructions of the past. 185-202, London: Routledge.

Sutton, J. E. G. 1971b. Local industries in Tanzania. unpublished seminar paper, History dept. University of Dar es Salaam.

Spier, R. F. G. 1973. Material Culture and Technology. Minneapolis: University of Missouri Press.

Sutton, J.E. G. 1975. Archaeology at Kalambo: terminology and historical interpretation Review Article. Journal of African History, XVI:137-141.

Stanislawsky, M. B. 1973. Longacre: Archaeology as Anthropology: A case study. American Antiquiety, 4:117121.

Sutton, J. E. G. 1985. Temporal and Spatial Variability in African Iron Furnaces, in Haaland, R. and P. Shinnie (eds.). African Iron Working, Ancient and Traditional. Oslo: Norwegian University Press.

Stannus, H. S. 1914. Nyasaland: Angoni Smelting Furnace. Man, 64-65:131-132.

Sutton, J. E. G. 1990. A Thousand Years of East Africa. Nairobi: British Institute in Eastern Africa.

Stark, M. T. 1993. Re-fitting the ‘cracked and broken facade’: The case for empirism in post-processual ethnoarchaeology, in Yoffee, N. and A. Sherratt. (eds.) Archaeological theory: Who sets the agenda? New directions in Archaeology, Cambridge: Cambridge University Press.

Sutton, J. E. G. and A. D. Roberts, 1968. Uvinza and its salt industry. Azania, III.45-86. Tambiah, S. J. 1993. (1990). Magic, Science, Religion and the scope of Rationality. Lewis Henry Morgan Lectures, Cambridge: Cambridge University Press.

Sterner, J. 1992. Sacred pots and ‘Symbolic Reservoirs’ in the Mandara Highlands of Northern Cameroon, in Sterner, J. and N. David (eds.) The African Commitment. Calgary: Calgary University Press.

173

Tambila, A. 1981. A History of the Rukwa Region (Tanzania) ca. 1870-1940 :Aspects of Economic and Social Change from Pre-Colonila to Colonial times. unpubl. Ph.D. Dissertation/ zur Erlangung der Wurde des Doktors der Philosophie der Universitat Hamburg.

Wainwright, G. A. 1942. The Coming of iron to some African Peoples. Man, 42:103-108.

Tew, M. 1950. Peoples of the Lake Nyasa Region. Ethnographic Survey of Africa, London: Oxford University Press.

Waldbaum, J. C. 1980. The First Archaeological Appearance of Iron and the Transitions to the Iron Age, in Wertime, T. and D. J. Muhly (eds.). The Coming of the Age of Iron. New Haven: Yale University Press.

Wainwright, G. A. 1954. The Diffusion of uma as a name for iron. Uganda Journal, XVIII:113-136.

Tessmann, G. 1913. Die Pangwe. 2. vols. Berlin. Weiss, B. 1996. The making and unmaking of the Haya lived world. London: Duke University Press.

Thomson, J. 1881. To the African lakes and back. Vol.1, London.

Weiss, B. 1998. Electric vampires: Haya rumors of the commodified body, in Strathern, A. J. and M. Lambek (eds.). 1998. Bodies and Persons, Comparative perspectives from Africa and Melanesia. Cambridge: Cambridge University Press.

Tilley, C. 1990. Claude Lévi-Strauss: structuralism and beyond, in C. Tilley (ed.). Reading material culture. Oxford: Blackwell Publishers. Tilley, C. 1991. Material Culture and Text: the art of Ambiguity. London: Routledge.

Wembah-Rashid, J. A. R. 1969. Iron Workers of Ufipa, Bulletin of The International Committee of Urgent Anthropological Research,11:65-72.

Tilley, C. 1993. Introduction: Interpretation and a Poetics of the Past, in Tilley, C. (ed.). Interpretative Archaeology. Oxford: Berg Publisher.

Wembah-Rashid, J. A. R. 1973. Iron Working in Ufipa. A Record of traditional processes of iron smelting and forging among the Fipa of Tanzania. Dar es Salaam: National Museum.

Tilley, C. 1994. A phenomenology of Landscape: Places, Paths and Monuments. Oxford: Berg Publisher.

Wertime, T. 1980. The Pyrotechnological Background, in Wertime, T. and J. D. Muhly (eds.). The Coming of the Age of Iron. New Haven: Yale University Press.

Tilley, C. 1999. Metaphor and Material Culture. Oxford: Blackwell. Todd, J. A. 1985. Iron Production by the Dimi of Ethiopia, in Haaland , R. and P. Shinnie (eds.). African Iron Working, Ancient and Traditional. 88- 101, Oslo: Norwegian University Press.

Wertime, T. A. and Muhley J. D. 1980. Introduction, in Wertime, T. A. and J. D. Muhly (eds.). The Coming of the Age of Iron. New Haven: Yale University Press.

Trigger, B. G. 1989. A History of Archaeological Thought. Cambridge: Cambridge University Press.

White, R. 1993. Foreword in Leroi-Gourhan, A. Gesture and Speech. Cambridge Massachusetts: MIT Press.

Trigger, B. G. 1995. Expanding middle range theory. Antiquity, 69:449-58.

Whyte, R. S. 1990. The Widow’s Dream: Sex And Death in Western Kenya, in Jackson, and I. Karp (eds.). Personhood and Agency. Uppsala,14:95-114. Sweden: Uppsala Studies in Anthropology.

Tringham, R. 1978. Experimentation, Ethnoarchaeology, and the Leapfrogs in Archaeological methodology, in Gould, R. (ed.). Explorations in Ethnoarchaeology. Albuquerque: University of New Mexico Press.

Widding, J. A. 1990. Subjective Body , Objective Space An Introduction, in Widding, J. A. (ed.). Body and Space,16:15-51. Uppsala: Uppsala Studies in Cultural Anthropology.

Turner, V. 1967. The Forest of Symbols, Aspects of Ndembu Ritual. Itacha: Cornell University Press.

Winch, P. 1958 (1988). The Idea of a Social Science and its relation to Philosophy. London: Routledge.

Turner, V. 1969. The Ritual Process. Itacha: Cornell University Press.

Wiessner, P. 1984. Reconsidering the behavioural basis for style: a case study among the Kalahari San. Journal of Anthropological Archaeology, 3:190-234.

Vansina, J. 1990. Paths in the Rainforest. Toward a History of Political Tradition in equatorial Africa, London: James Curry.

Willis, R. G. 1964. Traditional history and Social Structure in Ufipa. Africa, 37:425-52.

Wagner, R. 1986. Symbols that stand for themselves. Chicago: Chicago University Press. 174

Willis, R. G. 1966. Fipa and Related Peoples. London: Ethnographic Survey of Africa by D. Forde.

Wise, G. 1958a. Some Rituals of Iron working in Ufipa. Tanganyika Notes and Records, 51:232-238.

Willis, R. G. 1967. The Head and The Loins, LéviStrauss and Beyond. Man, 2(3):510-34.

Wise, G. 1958b. Iron Smelting in Ufipa. Tanganyika Notes and Records, 51:106-111.

Willis, R. G. 1968. Changes in mystical concepts and practices among the Fipa. Ethnology, 2:139-57.

Wiredu, K. 1996. Cultural Universals and Particulars, An African Perspective. Bloomington and Indianapolis: Indiana University Press.

Willis, R. G. 1970. Kaswa: Oral Tradition of a Fipa Prophet. Africa, 3:248-256. Willis, R. G. 1972. Pollution and Paradigms. Man, 7(3):369-378.

Wobst, M. 1977. Stylistic behaviour and information exchange, in C. Cleland (ed.). Papers for the director: research essays in honour of James B. Griffin. Ann Arbour: University of Michigan Press.

Willis, R. G. 1973. The indigenous critique of Colonialism, a case study, in Asad, T. (ed.). Anthropology and the Colonial Encounter. London: Itacha press.

Woodhouse, J. 1998. Iron in Africa: metal from nowhere, in Connah, G. (ed.). Transformations in Africa. 160-186, London: Leicester University Press.

Willis, R. G. 1974. Man and Beast. London: MacGibbon.

Wrigth, M. 1980. Iron Working and the old regional economy of South Rukwa. History Dept. University of Dar es Salaam, Tanzania, unpubl. paper.

Willis, R. G. 1978a. There was Certain Man. London: Oxford University Pres.

Wright, M. 1982. Towards a critical history of Iron Makers in Sumbawanga District, History Dept., University of Dar es Salaam, Tanzania, unpubl. paper.

Willis, R. G. 1978b. Magic and ‘Medicine’ in Ufipa, in Morley, P. and R. Wallis (eds.). Culture and Curing. London: Peter Owen

Wyckaert, R. P. 1914. Forgerons païens et forgerons crétiens au Tanganyika. Anthropos, IX:371-380.

Willis, R. G. 1981. A State in the Making. Bloomington: Indiana University Press.

Wylie, A. 1985. The reaction against analogy. Advances in Archaeological Method and Theory, 8:63-111.

Willis, R .G. 1989a. The ‘Peace Puzzle’ in Ufipa, in Howell, S. and R. Willis (eds.). Societies at Peace. Anthropological Perspectives, London and New York: Routledge.

Wylie, A. 1989. Archaeological cables and tacking: the implications of practice for Bernstein’s ’options beyond Objectivism and Relativism’. Philosophy of the social sciences, 19(1):1-18.

Willis, R. G. 1989b. Power begins at Home: The Symbolism of Male-Female Commensality in Ufipa, in Arens, W, and I. Karp (eds.). Creativity of Power. Washington: Smithsonian Institution Press.

Wylie, A. 1993. A proliferation of new archaeologies: “Beyond objectivism and relativism, in Yoffee, N. and A. Sherratt (eds.). Archaeological theory: who sets the agenda? Cambridge: Cambridge University Press.

Willis, R. G. 1991. The Body as Metaphor: Synthetic observations on an African art, in A. Jacobsen-Widding, (ed.). Body and Space. Uppsala: Acta Univerisitatis Uppsalensis, Uppsala Studies in Cultural Anthropology 16.

Yates, T. 1990. Archaeology through the Looking-Glass, in Bapty, I. and T. Yates (eds.). Archaeology after Structuralism. 154-202, London: Routledge. Yates. T. 1993. Frameworks for an Archaeology of the Body, in Tilley, C. (ed.). Interpretative Archaeology. Oxford: Berg Publisher.

Wilson, M. 1957. Rituals of kinship among the Nyakyusa. London: Oxford University Press. Wilson, M. 1958. The Peoples of the Nyasa-Tanganyika Corridor. Communications from the Schools of African Studies, New Series, 29, October: Cape Town: University of Cape Town. Wilson, M. 1959. Communal Rituals of The Nyakyusa. London: Oxford University Press. Wilson, M. 1977. For Men and Elders. London: International African Institute. 175

CAMBRIDGE MONOGRAPHS IN AFRICAN ARCHAEOLOGY No 1 No 2 No 3 No 4 No 5 No 6 No 7 No 8 No 9 No 10 No 11 No 12 No 13 No 14 No 15 No 16 No 17 No 18 No 19 No 20 No 21 No 22 No 23 No 24 No 25 No 26 No 27 No 28 No 29 No 30

BAR S75, 1980 The Niger Delta Aspects of its Prehistoric Economy and Culture by Nwanna Nzewunwa. ISBN 0 86054 083 9 BAR S89, 1980 Prehistoric Investigations in the Region of Jenne, Mali A Study in the Development of Urbanism in the Sahel by Susan Keech McIntosh and Roderick J. McIntosh ISBN 0 86054 103 7 BAR S97, 1981 Off-Site Archaeology and Human Adaptation in Eastern Africa An Analysis of Regional Artefact Density in the Amboseli, Southern Kenya by Robert Foley. ISBN 0 86054 114 2 BAR S114, 1981 Later Pleistocene Cultural Adaptations in Sudanese Nubia by Yousif Mukhtar el Amin. ISBN 0 86054 134 7 BAR S119, 1981 Settlement Patterns in the Iron Age of Zululand An Ecological Interpretation by Martin Hall. ISBN 0 86054 143 6 BAR S139, 1982 The Neolithic Period in the Sudan, c. 6000-2500 B.C. by Abbas S. Mohammed-Ali. ISBN 0 86054 170 3 BAR S195, 1984 History and Ethnoarchaeology in Eastern Nigeria A Study of Igbo-Igala relations with special reference to the Anambra Valley by Philip Adigwe Oguagha and Alex Ikechukwu Okpoko. ISBN 0 86054 249 1 BAR S197, 1984 Meroitic Settlement in the Central Sudan An Analysis of Sites in the Nile Valley and the Western Butana by Khidir Abdelkarim Ahmed. ISBN 0 86054 252 1 BAR S201, 1984 Economy and Technology in the Late Stone Age of Southern Natal by Charles Cable. ISBN 0 86054 258 0 BAR S207, 1984 Frontiers Southern African Archaeology Today edited by M. Hall, G. Avery, D.M. Avery, M.L. Wilson and A.J.B. Humphreys. ISBN 0 86054 268 8. £23.00. BAR S215, 1984 Archaeology and History in Southern Nigeria The ancient linear earthworks of Benin and Ishan by P.J. Darling. ISBN 0 86054 275 0 BAR S213, 1984 The Later Stone Age of Southernmost Africa by Janette Deacon. ISBN 0 86054 276 9 BAR S254, 1985 Fisher-Hunters and Neolithic Pastoralists in East Turkana, Kenya by John Webster Barthelme. ISBN 0 86054 325 0 BAR S285, 1986 The Archaeology of Central Darfur (Sudan) in the 1st Millennium A.D. by Ibrahim Musa Mohammed. ISBN 0 86054 367 6. BAR S293, 1986 Stable Carbon Isotopes and Prehistoric Diets in the South-Western Cape Province, South Africa by Judith Sealy. ISBN 0 86054 376 5. BAR S318, 1986 L'art rupestre préhistorique des massifs centraux sahariens by Alfred Muzzolini.. ISBN 0 86054 406 0 BAR S321, 1987 Spheriods and Battered Stones in the African Early and Middle Stone Age by Pamela R. Willoughby. ISBN 0 86054 410 9 BAR S338, 1987 The Royal Crowns of Kush A study in Middle Nile Valley regalia and iconography in the 1st millennia B.C. and A.D. by Lázló Török.. ISBN 0 86054 432 X BAR S339, 1987 The Later Stone Age of the Drakensberg Range and its Foothills by H. Opperman. ISBN 0 86054 437 0 BAR S350, 1987 Socio-Economic Differentiation in the Neolithic Sudan by Randi Haaland. ISBN 0 86054 453 2 BAR S351, 1987 Later Stone Age Settlement Patterns in the Sandveld of the South-Western Cape Province, South Africa by Anthony Manhire. ISBN 0 86054 454 0 BAR S365, 1987 L'art rupestre du Fezzan septentrional (Libye) Widyan Zreda et Tarut (Wadi eshShati) by Jean-Loïc Le Quellec. ISBN 0 86054 473 7 BAR S368, 1987 Archaeology and Environment in the Libyan Sahara The excavations in the Tadrart Acacus, 1978-1983 edited by Barbara E. Barich. ISBN 0 86054 474 5 BAR S378, 1987 The Early Farmers of Transkei, Southern Africa Before A.D. 1870 by J.M. Feely. ISBN 0 86054 486 9 BAR S380, 1987 Later Stone Age Hunters and Gatherers of the Southern Transvaal Social and ecological interpretation by Lyn Wadley. ISBN 0 86054 492 3 BAR S405, 1988 Prehistoric Cultures and Environments in the Late Quaternary of Africa edited by John Bower and David Lubell. ISBN 0 86054 520 2 BAR S418, 1988 Zooarchaeology in the Middle Nile Valley A Study of four Neolithic Sites near Khartoum by Ali Tigani El Mahi. ISBN 0 86054 539 3 BAR S422, 1988 L'Ancienne Métallurgie du Fer à Madagascar by Chantal Radimilahy. ISBN 0 86054 544 X BAR S424, 1988 El Geili The History of a Middle Nile Environment, 7000 B.C.-A.D. 1500 edited by I. Caneva. ISBN 0 86054 548 2 BAR S445, 1988 The Ethnoarchaeology of the Zaghawa of Darfur (Sudan) Settlement and Transcience by Natalie Tobert. ISBN 0 86054 574 1

No 31

BAR S455, 1988 Shellfish in Prehistoric Diet Elands Bay, S.W. Cape Coast, South Africa by W.F. Buchanan. ISBN 0 86054 584 9 No 32 BAR S456, 1988 Houlouf I Archéologie des sociétés protohistoriques du Nord-Cameroun by Augustin Holl. ISBN 0 86054 586 5 No 33 BAR S469, 1989 The Predynastic Lithic Industries of Upper Egypt by Liane L. Holmes. ISBN 0 86054 601 2 (two volumes) No 34 BAR S521, 1989 Fishing Sites of North and East Africa in the Late Pleistocene and Holocene Environmental Change and Human Adaptation by Kathlyn Moore Stewart. ISBN 0 86054 662 4 No 35 BAR S523, 1989 Plant Domestication in the Middle Nile Basin An Archaeoethnobotanical Case Study by Anwar Abdel-Magid. ISBN 0 86054 664 0 No 36 BAR S537, 1989 Archaeology and Settlement in Upper Nubia in the 1st Millennium A.D. by David N. Edwards. ISBN 0 86054 682 9 No 37 BAR S541, 1989 Prehistoric Settlement and Subsistence in the Kaduna Valley, Nigeria by Kolawole David Aiyedun and Thurstan Shaw. ISBN 0 86054 684 5 No 38 BAR S640, 1996 The Archaeology of the Meroitic State New perspectives on its social and political organisation by David N. Edwards. ISBN 0 86054 825 2 No 39 BAR S647, 1996 Islam, Archaeology and History Gao Region (Mali) ca. AD 900 - 1250 by Timothy Insoll. ISBN 0 86054 832 5 No 40 BAR S651, 1996 State Formation in Egypt: Chronology and society by Toby A.H. Wilkinson. ISBN 0 86054 838 4 No 41 BAR S680, 1997 Recherches archéologiques sur la capitale de l’empire de Ghana Etude d’un secteur d’habitat à Koumbi Saleh, Mauritanie. Campagnes II-III-IV-V (1975-1976)-(1980-1981) by S. Berthier. ISBN 0 86054 868 6 No 42 BAR S689, 1998 The Lower Palaeolithic of the Maghreb Excavations and analyses at Ain Hanech, Algeria by Mohamed Sahnouni. ISBN0 86954 875 9 No 43 BAR S715, 1998 The Waterberg Plateau in the Northern Province, Republic of South Africa, in the Later Stone Age by Maria M. Van der Ryst. ISBN 0 86054 893 7 No 44 BAR S734, 1998 Cultural Succession and Continuity in S.E. Nigeria Excavations in Afikpo by V. Emenike Chikwendu. ISBN 0 86054 921 6 No 45 BAR S763, 1999 The Emergence of Food Production in Ethiopia by Tertia Barnett. ISBN 0 86054 971 2 No 46 BAR S768, 1999 Sociétés préhistoriques et Mégalithes dans le Nord-Ouest de la République Centrafricaine by Étienne Zangato. ISBN 0 86054 980 1 No 47 BAR S775, 1999 Ethnohistoric Archaeology of the Mukogodo in North-Central Kenya Hunter-gatherer subsistence and the transition to pastoralism in secondary settings by Kennedy K. Mutundu. ISBN 0 86054 990 9 No 48 BAR S782, 1999 Échanges et contacts le long du Nil et de la Mer Rouge dans l'époque protohistorique (IIIe et IIe millénaires avant J.-C.) Une synthèse préliminaire by Andrea Manzo. ISBN 1 84171 002 4 No 49 BAR S838, 2000 Ethno-Archaeology in Jenné, Mali Craft and status among smiths, potters and masons by Adria LaViolette. ISBN 1 84171 043 1 No 50 BAR S860, 2000 Hunter-Gatherers and Farmers An enduring Frontier in the Caledon Valley, South Africa by Carolyn R. Thorp. ISBN 1 84171 061 X No 51 BAR S906, 2000 The Kintampo Complex The Late Holocene on the Gambaga Escarpment, Northern Ghana by Joanna Casey. ISBN 1 84171 202 7 No 52 BAR S964, 2000 The Middle and Later Stone Ages in the Mukogodo Hills of Central Kenya A Comparative Analysis of Lithic Artefacts from Shurmai (GnJm1) and Kakwa Lelash (GnJm2) Rockshelters by G-Young Gang. ISBN 1 84171 251 5

No 53 BAR S1006, 2001 Darfur (Sudan) In the Age of Stone Architecture c. 1000 - 1750 AD Problems in historical reconstruction by Andrew James McGregor. ISBN 1 84171 285 X No 54 BAR S1037, 2002 Holocene Foragers, Fishers and Herders of Western Kenya by Karega-MNJnene. ISBN 1 84171 1037 No 55 BAR S1090, 2002 Archaeology and History in Ìlàrè District (Central Yorubaland, Nigeria) 1200-1900 A.D. by Akinwumi O. Ogundiran. ISBN 1 84171 468 2 No 56 BAR S1133, 2003 Ethnoarchaeology in the Zinder Region, Republic of Niger: the site of Kufan Kanawa by Anne Haour. ISBN 1 84171 506 9 No 57 BAR S1187, 2003 Le Capsien typique et le Capsien supérieur Évolution ou contemporanéité. Les données technologiques by Noura Rahmani. ISBN 1 84171 553 0 No 58 BAR S1216, 2004 Fortifications et urbanisation en Afrique orientale by Stéphane Pradines. ISBN 1 84171 576 X No 59 BAR S1247, 2004 Archaeology and Geoarchaeology of the Mukogodo Hills and Ewaso Ng’iro Plains, Central Kenya by Frederic Pearl. ISBN 1 84171 607 3 No 60 BAR S1289, 2004 Islamic Archaeology in the Sudan by Intisar Soghayroun Elzein. ISBN 1 84171 639 1.