Dissertation Title:
Petrographic and Chemical Analysis of Ceramics from Douentza and Tongo Maaré Diabal
Abstract …………………………
Acknowledgments ………………………
Table of Content ……………………
List of Tables ………………………..
List of Figures ………………………
CHAPTER 1: INTRODUCTION
CHAPTER 2: BACKGROUND CHAPTER
CHAPTER 3: MATERIALS AND METHODS
CHAPTER 4: RESULTS
CHAPTER 5: DISCUSSION (INTERPRETATION OF THE DATA)
CHAPTER 6: CONCLUSIONS
REFERENCES CITED
APPENDICES
-what you will study, why it is of interest (general research problem) specific aims.
-brief introduction to the temporal period(s), study region, study site(s), materials,
-brief introduction to materials and methods,
-brief explanation of how the dissertation is organised
-detailed information about the physical environment, climate, cultural period(s), and if appropriate, culture change, the site(s) under study, and previous studies done in the region and/or the site that relate to your study
Excavation and Survey of Tongo Maare Diabal and the Surrounding Sites.
-Samples:what are they? how were the collected? Who did the collecting? how many? general size; context or excavation information etc.
-Methods of analysis
This dissertation will discuss the petrographic and chemical analysis of samples of ceramic from two predominant pottery types found at TMD and sites surrounding DTZ.
I am using thin section petrography to attempt to answer the following questions:
The town of Douentza is located on the Eastern edge of the Mopti Region of present-day Mali. It rests on a flat alluvial plain between two imposing sandstone massifs, the Bandiagara and Dyounde plateaus, orientated roughly SW-NE.
It lies in a crucial gap between the towering Bandiagara and Dyounde escarpments.
To the north of Douentza is the ancient settlement mound of Tongo Maare Diabal (“the Ju-Jube tree covered dune”). A 9ha “tell”.
Two hundred kilometres due south of Timbuktu lies the Malian town of Douentza, a polyglot mudbrick settlement founded by Bambara and Dogon peoples. (MacDonald, Arch International)
Tuareg camel and donkey caravans make a weekly trek from the Sahara carrying salt to excahnge for millet at Douentza and Bankass. The dirt road between the two towns is the principal overland route to Timbuktu (MacDonald, Arch International).
A survey of the area for evidence of the Late Neolithic in 1993 led to a 1x4 metre test-excavation at TMD. An aberrant radiocarbon date led MacDonald And Togola to return to the site in 1995 to undertake more substantial excavations. (MacDonald, Arch International)
In 1993 Togola and MacDonald conducted a 2x2m test excavation which revealed mud-brick structures with multiple occupation horizons.
In 1995 MacDonald carried a further excavation. Three 4x4m unit reached natural soil and the earliest extent of the occupation. This excavation was completed in 1996 and revealed five horizons of superimposed earthworks (Gestrich, 2014).
Archaeologists have been aware of the emerging urbanism of the Inland Niger Delta since McIntosh and McIntosh’s excavations of Jenne-Jeno in 1977. (MacDonald, Arch International)
The primary focus of this study is TMD where excavations were carried out in 1993, 1995, 1996 and 2010.
In 2010 Gestrich undertook an excavation of horizon 5 with a 9x5m unit, a 7x5m unit and a 7x4m unit. As they were exploring the most recent occupation horizon the depth of the units reached between 50cm and 120cm.
The 5 further sites from which my samples have been selected are amongst 11 identified by Gestrich in his 2010 survey of sites within 5km of TMD.
The pottery found at these additional sites represent surface finds. Gestrich asserts that the morphology of the pottery and the nature of the stratigraphy of these additional sites makes it likely that they were contemporaneous with TMD.
A consistent approach to recording sherds was taken across these excavations. Togola and MacDonald based their system on the one used by Susan McIntosh at Jenne-Jeno. This was replicated by Gestrich so that the data sets could be used as a better resource for comparing ceramics in the wider region.
“To begin the sorting process, the bags of washed pottery were divided into four categories: rim sherds, body sherds, sherds from the bases of pots, and so-called feature sherds , which mainly include handles, lugs and lids. The body sherds were grouped according to temper and decoration and the number of sherds in each group was then recorded. The protocol followed for the rim sherds was far more exhaustive, and each rim sherd was individually recorded. On the recording sheets were noted:
There are two predominant pottery types, pounding in a concave form over a mat; and pounding out the base over a convex form. These are perceived to represent different ‘potters lineages’ and are both found throughout each occupation horizon at TMD.
“Pounding in a concave form; using a mat nowadays made from baobab (adanso-nia digitata) fibres and a tamper such as a stone. Coils can be added to further build up the rim. This technique leaves distinct traces on the body of the pot meaning that many sherds stemming from pots made in this manner can be clearly distinguished.Two decor types never co-occurred with the mat impressions. One was the most common decorative motif, made by a complex braided cord roulette, while the other was painted or red-slipped decor.” (Gestrich 2014)
“Pounding out of the base on a convex form. This usually comprises the use of a pre-existing upside-down pot or a wooden pre-form as a mould for the base, with the subsequent building up of the body with coils. The pots made this technique show several distinctive traces, which allowed Keita to identify them on the complete pots. These are firstly a thin base prone pot where the transition from beating on the mould to coil-building was made. Often, the traces where subsequent coils were mounted can also be seen on the interior of the pots.”
The sherds relating to this latter tradition make up the bulk of the pottery assemblage, at 53.7% whereas the mat-formed sherds only amount to 17.2%.
The vast majority, 94.4%, of all sherds recovered were tempered with grog. Gestrich notes that the prevalence of this single grog type is significant as it “cross-cuts all vessel types.” There are a minority of notable exceptions however.
Found in association with iron workshops a high percentage of pottery tempered with plant material. This is consistent with some of the technical ceramics recovered from TMD and sites nearby which were also tempered with plant material. Gestrich observed that the tendency for the organically tempered pots to differ morphologically from the grog tempered pottery may suggest they were made by a different group.
Indeed, it is possible that this represents male pottery associated with iron production.
Gestrich undertook a correspondence analysis using published pottery datasets from sites in the region. This compared the different decorative motifs, the co-occurrence of motifs and the proportions and percentages of each motif at site level.
“The temporal frame of the analysis allows only the inclusion of datasets from AD 400 - 1200, limiting it to a period that lies between a marked change in pottery and economy at Jenne-Jeno (McIntosh, 1995) and an often cited period of instability in the 13th and 14th centuries (Gallay et al., 1998), as well as being the temporal extent of TMD.” (Gestrich, 2014)
Pottery decoration in many cases is regarded as the most ephemeral aspect of production which is subject to change and spread quickly. McIntosh (1995) has remarked that there is a conservatism in the pottery of the Middle Niger region.
This has been observed by Gestrich where his correspondence analysis showed clear regional differences in pottery decoration. This boundedness of aesthetic diffusion appears to be in contrast to Gosselain’s ethnographic studies which showed that decorative forms transgressed cultural and linguistic groups and diffused by propinquity.
This however may offer insights in kinship structure. Peter G Roe suggests that patrilocality can lead to stylistic uniformity. Roe addresses the question of when a female potter learns her craft. If it is learned when the woman changes residence due to marriage then she will be taught by her mother-in-law. This model would suggest that the Deetz-Longacre hypothesis (Longacre, 1964)(Deetz, 1965) of propinquity is correct (Roe, P, 2013 in (eds.) Car,c; Neitzel, J, 2013).
When Gestrich split the pottery finds by mat-formed and non-mat formed pottery in the correspondence analysis each type indicated a relationship with groups from different “geographic spheres”. The mat-formed pottery “fits neatly within the Oudalan sites (vin Czerniewicz, 2004, 2011)”.
The non mat-formed pottery shows similarities with pottery found in the Eastern IND and the Lakes region. At this time TMD would have been in the eastern margins of the IND and the south-eastern extent of the Lakes region “lies not far to the North”(Gestrich).
Gestrich offers two tentative conclusions about the mix of pottery:
Tal Tamari discusses the rise of the caste system in Mali possibly happening with the rise of the Mali Empire. This is in contrast to McIntosh who contends that the blacksmith caste system was fully established by the 8th century CE.
On the Eastern side of the Bandiagara Escarpment it is known that some agriculturalists practised iron smelting alongside farming.
If all the clay is the same, the temper is the same, the technology is different. Woman gets married, mother-in-law gets her to make a pot in a specific way. Her son gets married, her and mother-in-law teach daughter in law to make a pot.
Pot gets traded, gifted.
Chaine-operatoire - women collect the clay. Women then temper the clay with sand, possibly?, and grog. Women buy the clay or take the clay back to their own compounds/domicile where they produce the pottery in the style according to their lineage.
OR - people are only marrying into their castes and therefore the agriculturalists are required to find wives in Oudala.
Rural complexity - producing large amounts of iron - distance parity model against Ghana Empire?
Standard thin sections were taken from 29 ceramics from 6 sites on the Gandamia plain lying between the Bandiagara and Dyounde massifs, Douentza, central Mali. The samples were taken from pottery finds collected over 3 seasons of fieldwork carried out in 1995, 1996, 2010.
The samples were picked initially to reflect the two predominant pottery types found at TMD and sites surveyed within 5km of TMD.
The two pottery types which are discussed in further detail in Chapter??? have been analysed macroscopically and can be differentiated by surface decoration and forming methods.
In addition to these two types a limited number of less predominant or exceptional examples were picked to be sampled. This includes 3 samples which have been burnished or slipped; 1 sample of a sherd which was exceptional for containing organic temper; and 1 sample of a sherd which contained a large amount of iron, possibly slag.
This sampling strategy aims to identify and explore the relationship within and between the readily established stylistic and technological ceramic categories established by Gestrich.
19 of the samples were taken from pottery found at TMD. 5 of these represent pottery moulded over a convex form. 2 of the TMD samples had burnished/slipped surface treat. 1 was organically tempered. 1 showed particularly high amounts of iron or slag inclusions.
3 samples of sherds from TMS were taken. These represented 1 mat-formed, 1 non mat-formed and 1 slipped sherd.
4 samples of sherds recovered from Ngassa were taken. (2 from Ngassa, 2 from Site 09). These all represent mat-formed pottery types.
2 samples of sherds recovered at Lumoorde Hoore Nyiwa were taken. The both were taken from mat-formed pottery types.
2 samples of sherds recovered from the surface of sites at Tangu. The 2 samples represented one mat formed and one non mat-formed type of pottery.