Pemban People: Local Ceramics and Changing Identities

Abstract

This chapter examines locally produced ceramics on nineteenth-century clove plantations, analyzing how we might examine these as an artifact of modernity. Locally produced ceramics on Zanzibari clove plantations provide an interesting study because, despite widespread social change and the arrival of massive numbers of immigrants, the style of these wares remains remarkably stable from the eighteenth century through to the present day. This chapter argues that this conservatism in production is actually indicative of the manner in which immigrants were able to incorporate themselves into Zanzibari society. Drawing on oral historical evidence to interpret the archaeological data, it discusses the way in which pottery manufacturers, usually women working on a relatively small scale, may have been part of communities of practice. The ability to apprentice immigrant women into the skill of making Zanzibari wares may have been one way in which immigrants made themselves into integral members of coastal society. As ceramics are also an important part of foodways, this chapter argues that this shift was not only linked to production; the use of these specific coastal forms and designs was also a way in which newcomers were able to signal their cultural incorporation into local norms. Food was central to this transformation, as was discussed by several elderly residents of Pemba in oral historical interviews.

Unlike other areas of material culture and landscape discussed in this book, such as imported commodities and stone houses , locally produced ceramics garner no discussion from historians studying nineteenth-century Eastern Africa. These ceramics gain little discussion in documents from the time, which were of course largely written by European men, who were little inclined to write of cooking pots. However, this lack of historical discussion is set against their ubiquity in the past and in the present. Today, most Zanzibari households have at least one of these pots in their kitchen, normally coupled with a lid, for the purpose of cooking rice . Indeed, the popularity of locally produced ceramics appears to be growing due to fears about leeching from aluminum cookware. Historical scholarship of nineteenth- and early twentieth-century Zanzibar largely focuses on the ways in which Zanzibaris were entwined with broader processes of modernity. Even scholars focusing on gender relations during this period (e.g., Fair 1998, 2001) do not mention locally produced ceramics.

These objects seem perhaps somewhat of an anachronism in a nineteenth-century context in which Zanzibaris were cosmopolitan and liked to global processes of capitalism . Historical studies, which so wholly ignore these artifacts, would suggest that they are of little importance in the materialities of identity in nineteenth-century Zanzibar. But as I explore in this chapter, these objects were a dominant object of production in clove plantation areas, and have remained important for female producers on into the twenty-first century. Their continued importance seems to suggest that they are actually central to understandings of gendered and ethnic identities through production, exchange , and use. Rather than being an afterthought to the way in which enslaved Africans claimed their place as Zanzibaris, I argue that local ceramics were actually central to this claim within the plantation context.

Ceramics and Clove Plantations

The ceramics recorded from excavations at Mgoli and from the survey work were remarkably homogeneous. Even more remarkable was the fact that during the nineteenth-century, locally-produced ceramics—almost certainly made on a very small scale—were largely homogeneous in form and decoration along the Swahili coast, with significant changes inland. Locally-produced ceramics formed the bulk of artifacts overall from excavations, as discussed in the last chapter. Ceramic analysis discussed here is drawn largely from Trench C and D from the site of Mgoli.¹ In Trench C, locally produced ceramics comprised 91 percent of the total ceramic artifacts. From the slightly later contexts in Trench D, this figure was a little lower, 85 percent.

Analysis of diagnostic ceramic sherds from all contexts demonstrated that 98 percent were open bowls with rounded bases and a carination (Fig. 7.1).² A total of 58 sherds in lip, ledge, or handle fragments were taken to represent lids, most likely used on the rounded open bowls. Only a very tiny number of sherds were diagnostic for forms other than these two. Of the total assemblage, five sherds were recognizable as base sherds (all were flat disc bases). These comprised < 1 percent of the diagnostic sherds for form from the site (n  =  929). Obviously, the data from Mgoli cannot be taken as a representative for all clove plantation sties. However, there are clear indications that varied ceramic forms and styles would be easily visible from survey data. In western Tanzania, where locally produced ceramics in the nineteenth-century clearly took a diversity of forms and styles, surface survey easily located varied sherds with burnished, roulette, and appliqué decorated sherds (along with a couple of incised arcs, similar to coastal styles; Wynne-Jones and Croucher 2007). The fact that surface survey and excavation in other areas of East Africa have easily identified locally produced ceramics from the nineteenth-century with various forms strongly suggests that the Mgoli data fit into a clear pattern for Zanzibari clove plantation ceramics. The Zanzibar Clove Plantation Survey did not produce any evidence of variance in this pattern and, as I have discussed extensively, below, this type of carinated open-bowl form is still manufactured on Zanzibar and elsewhere on the East African coast today.

Fig. 7.1

Carinated ceramic, Mgoli

Contemporary cooking practices on Zanzibar today make understanding the likely function of these vessels fairly easy. Much cooking today utilizes aluminum pans, which gained popularity from the early twentieth century, but rice is still regularly prepared in ceramic vessels. Anecdotal evidence suggests that growing health concerns about the leeching of aluminum from metal pans is increasing the popularity of cooking rice in ceramics, along with the fact that rice cooked in this way supposedly tastes better (from my limited number of samples, it is delicious). To cook rice in such a vessel, the grain and water is placed into the dish, the lid is placed onto the vessel, and hot coals are piled onto this lid while the pot sits on a charcoal stove, slowly cooking the grain inside. The contemporary picture of ceramics on Zanzibar is therefore homogeneity, round-based carinated open bowls , and a much smaller number of lids, both known to be utilized for cooking. Although only a small proportion of sherds from the Mgoli assemblage had visible sooting, potentially hinting that some may have been used to serve food, a point I shall return to later.

In decoration, homogeneity was also visible. A total of 18 identifiable ceramic designs, all in incised and impressed forms, were recorded from across the excavated contexts at Mgoli . Of these, only nine were represented in contexts selected for analysis. Six of these types showed some form of representation of an arc form above the carination (Fig. 7.1), whereas the three others had impressions or incisions repeated along the carination. These arcs were commonly made by the incision of single or multiple lines in a repeated pattern around the circumference of the vessel. Occasional derivations included impressed arcs, and nonarc incised and impressed decoration. From analyzed contexts in Trench C, representing mid-nineteenth-century material, out of six identifiable decorative incisions or impressions recorded, four included some form of arc decoration.³ Sherds from analyzed contexts in Trench D, representing late nineteenth-century material showed a similar trend toward arc designs. Three out of five recognizable designs showed arc elements, with one of these variations (single arcs) found in both Trench C and D.⁴

Fig. 7.2

Nineteenth-century ceramics from the Kilwa region, drawn by S. Wynne-Jones

Other elements of the open-bowl form also showed some minor variations. The rim types were as varied as decorative elements on carinations. A total of 10 different rim types were identified from a total of 150 rim sherds recorded at the site.⁵ These showed variation, and while one type was most common across the site overall, this did not hold true for both units.⁶ Although minor variations in these styles were visible, there seems to be no clear progression through time, but rather a suite of variations within a generally definable style. The final decorative element in locally produced ceramics came from surface treatment. Vessels were commonly burnished, with a quarter or more of sherds from each context found to have this treatment. Red paint was also found on several sherds, but this totaled < 1 percent of ceramics by count from analyzed material in Trench C and D. This paint was largely applied to the outside of the vessel, from the carination upward, sometimes running over the rim and into the interior.

Ceramics from Mgoli show an overall style, with almost all conforming to open carinated bowl forms or lids, these latter most likely intended to fit the bowls vessels. Decoration also broadly fits into a style, with incised or impressed elements above the carination, often taking the form of arcs. This was then occasionally supplemented with red paint, and vessels were commonly finished with burnishing. In addition to decorative variation, rim forms varied, but in ways which are only discernible when carefully breaking down stylistic elements. It would have been obvious that these vessels still fit the common ceramic pattern. However, to makers and users, these variations in rims and decorations may have provided subtle differences allowing them to choose patterns by preference, or as the result of the individual style of potters .

Regional Style

To analyze the meaning of these ceramics requires some kind of broader context beyond a single clove plantation site. Survey data from the Zanzibar Clove Plantation Survey were problematic. The majority of sherds visible on the ground surface were undiagnostic for form and decoration. Nevertheless, several sherds with variations on arc motifs seen in ceramics at Mgoli were recovered, as were others with incised designs, punctuates and red paint. These allow me to suggest, tentatively, that the trend toward arcs on open carinated bowls was a common one for ceramics across clove plantation contexts. What did differentiate the sherds recovered from survey was a higher proportion of base sherds, four in total from across 64 sites. This may imply that the almost total lack of base sherds at Mgoli is an aberration, and that perhaps vessels with bases were more common in the past.

As discussed in Chap. 2, archaeological studies in Eastern Africa tend to cut off prior to the nineteenth-century . However, this situation is changing with significant studies from Walz (2010), Marshall (2011), and Biginagwa (2012) providing detailed treatment of later historical sites. On Pemba , no other later historical archaeology has been taken so far. However, Fleisher’s (2003) extensive survey of northern Pemba did record later sites and included these within his ceramic typology. This work allows for an understanding of the local ceramic forms found at Mgoli in relation to earlier periods in Pemba. On Pemba, as urban centers began to grow in size, two key changes took place. One was the use of rice as a crop and its adoption as a staple grain of the islands. This began in the eleventh century AD, from which time rice rapidly came to dominate the Pemban diet as demonstrated by the archaeobotanical record and early historical documents, such as a mention by the fourteenth-century traveler Ibn Battua (Fleisher 2010, p. 205; Walshaw 2010). Changing ceramic patters accompanied this shift in diet, now understood in part to relate to the importance of feasting as a means of cementing the power of new elites in urban centers on Pemba. As discussed in Chap. 2, the earliest ceramics found here were Tana tradition , which have been associated with the indigenous roots of urbanism in the East African coast (Chami 1998). By the fourteenth century, new bowl forms were commonly used, often large and highly decorated with hematite and graphite burnishing, a technique completely absent from nineteenth-century ceramics in Zanzibar. These had flat bases, and Fleisher has argued that the predominance of bowl forms—forming > 70 percent of ceramic assemblages in northern Pemban towns by the fifteenth century, up from just 16 percent of vessels in the eighth to tenth centuries—was related to new practices of feasting from the domestic to the public ritual level (Fleisher 2010, p. 210).

These wares are very different from those of the nineteenth-century , but Fleisher’s analysis lays the groundwork for understanding a specifically Pemban culture in which locally produced ceramics, in a relationship with forms and decorations of imported ceramics, were at the heart of the production of local identity (via rice consumption) and power relations (via feasting) from as early as the first centuries of the second millennium AD. The relationship with later ceramic forms becomes difficult to trace at this point. The work focusing on feasting has a clear relationship only to periods up to around 1700. The hiatus of 200 years or more before a detailed level of analysis can be undertaken via clove plantation data is problematic and relates to a lack of interest in the archaeology of the recent past in the East African coast, coupled with relatively few archaeologists researching in this region. However, Fleisher’s (2003) earlier survey work did produce a complete ceramic typology which ran up to later historical periods. The rounded carinated pots described above match those of Fleisher’s (2003, p. 258) “Type 13,” described as “simple restricted vessels.” This particular form is discussed as being characteristic of local ceramics from the sixteenth to eighteenth centuries on Pemba . Evidence from Mgoli and contemporary observations show that this form is actually characteristic of the sixteenth through twenty-first centuries, providing a long genealogy for the formal and decorative practices of contemporary craft producers on Zanzibar today (although the information on decoration found on sixteenth- to eighteenth-century Pemban ceramics is almost nonexistent).

Ceramic evidence also exists to understand a little of Mombasa through excavations at Fort Jesus, carried out by Kirkman (1974). Abdalla bin Jabir, the plantation owner at Mgoli, was said to have come to Pemba from Oman via Mombasa. The city was only 100 km to the north of the island, with known long-term trade links between the two. Mombasa was the entry point for serious Omani interest in the area from 1698. Similar to Zanzibar, Mombasa was deeply impacted by the political and economic shifts occurring in nineteenth-century Eastern Africa, and land near the city and in the area around nearby Malindi was also turned into grain plantations (Cooper 1977). A large volume of material was excavated from the fort in the mid-twentieth century, but the artifact analysis was very much of its time. Detailed descriptions of various types of forms and decorations are provided, but with no accompanying quantitative information, we learn only that types are “common” or “rare.”

Mombasa was a hub of colonial activity for the Portuguese from the sixteenth century and Omanis from the late seventeenth century until occupation by the British. Fort Jesus was a central point for colonial powers, and residents of the Fort would have had ample access to imported goods. However, material excavated by Kirkman demonstrates that throughout the seventeenth- through nineteenth-century residents of the Fort made ample use of local ceramics. Findings from Fort Jesus dating to the seventeenth century were similar to ceramic forms and decorative types Kirkman had encountered at the Swahili town site of Gede, located only 65 miles to the north of Mombasa. In the earliest levels at Fort Jesus, the local ceramic finds included the rare use of lamps and “horned bowls,” likely used as cooking stoves. It is hard to ascertain precisely the relationship between the ceramics described for the nineteenth-century and those of earlier periods (Kirkman 1974, pp. 90–91). The increase in large carinated pots with red necks, incised ornament and appliqué wares appears to have a close formal relationship with nineteenth-century Zanzibari wares, as does the increase in rounded and shouldered cooking pots. In decoration, charts of design elements (Kirkman 1974, pp. 259–261) show definite matches with those types found on Zanzibar, as well as demonstrating a diversity of patterns not identified within the Mgoli assemblage.⁷

While parallels to all ceramics found within the relatively simple range at Mgoli can be recognized with the Fort Jesus assemblage, there are also significant divergences between the two sets of data. One is in terms of form; the vessels available at Fort Jesus seemed to have offered almost a bewildering diversity in contrast to the almost homogeneous round-based open pots and lids in use at Mgoli. “Eating bowls” in a form dating back to the late seventeenth century were in use during the nineteenth-century at Fort Jesus , as were a new form of “sturdy bowls” with flat bases and a thick bottom, and some locally made ceramic lamps. In addition to these differences in form, some of the decoration while having parallels to Mgoli in terms of its elements—particularly in the use of repeated arcs—was sometimes applied to the body of the pot in an appliqué mode of decoration, where extra clay was applied in order to make the decoration, producing almost the inverse effect to incised decoration.

This variety of forms appeared to be matched by Marshall’s findings at watoro and Giriama sites in Kenya . Here, she found open bowls, restricted bowls and jars, and lids. The forms here showed considerably more variation than was represented at Mgoli , and seem to be closer to some of the vessel types described by Kirkman for Fort Jesus . Marshall’s study also identified a brushed surface treatment and decorative elements not present in any of the Mgoli ceramics. Some of the decorative styles visible in Mgoli ceramics such as incised lines running vertically through the carination were also found by Marshall. But the types of arc decorations so common at Mgoli were absent from her material. Comparisons between Makoroboi and Koromio showed high inter-household variability in the form of ceramics, potentially “establishing earthenware production (or trade) as probably organized at the household level rather than settlement level” (Marshall 2011, p. 166). In discussions of local ceramic production and trade (below), this will become an important point for comparison. In addition to the greater diversity of form, Marshall (2011, p. 170) identified particular kinds of stylistic links between watoro sites and the local ceramics from Mgoli. Beveled rims (identified as grooved rims at Mgoli) were found only at Koromio, the watoro site closest to the coast, potentially acting as “a marker for closer economic or cultural connections to the coast.” Marshall also found cooking lids with folded angles only at Koromio, another link with the Mgoli assemblage. Despite these connections however, what emerged from ceramic findings at all three of the Kenyan sites excavated by Marshall was a high incidence of intra-site variability. As she pointed out, this was markedly different from the type of homogeneous material patterning found on Zanzibar, and, indeed at other coastal sites (Marshall 2011, p. 195).

The heterogeneous nature of ceramics studied by Marshall is thrown into sharper contrast by findings from the Kilwa area. The town of Kilwa further south in the Tanzanian coast is now a World Heritage Site, nominated as such for the fluorescence of the town of Kilwa Kisiwani and associated Songo Mnara during the mid-second millennium AD. By the nineteenth-century , the town had changed dramatically, with wealth transferring to the growing settlement of Kilwa Kivinje. A caravan route from Lake Nyasa meant that the town was very important within the nineteenth-century economic and political relations of the coast, but this was not accompanied by an influx of new population and the transformation of the agrarian sector to the north as was the case with Zanzibar and Mombasa (Alpers 1975). Comparison discussed here came from a small number of sherds which were recovered as part of an extensive survey project in the region designed to investigate the development of urbanism in centuries prior to the later historical transformations discussed here (Wynne-Jones 2005, 2006). Although the sample of later material was small, similarities in ware types were remarkable (Fig. 7.2). Nineteenth-century Kilwa ceramics exhibit slight differences to the design types that are found on Zanzibar, such as smaller arcs, a double row of parallel arcs, the use of appliqué decoration (matching the style of Fort Jesus), and a slightly more rounded shape to the majority of carinations on the pots. Nevertheless, these vessel forms and decoration styles very clearly relate to those of Mgoli and with those at Fort Jesus (Croucher and Wynne-Jones 2006, p. 119). What makes this similarity all the more remarkable is the fact that decoration types at the watoro sites of Koromio and Makoroboi are so different. The similarity between ceramic styles found at coastal sites (especially between material from Zanzibar, Kilwa, and—perhaps in a more limited way—Fort Jesus ) is all the more extraordinary when we consider that around the eleventh to thirteenth centuries AD, local ceramic styles (such as the bowls recorded by Fleisher on Pemba and Husuni Modeled ware found at Kilwa and discussed in Chap. 2) appear to have diverged. Just as towns along the East African coast began to assert a more clearly “Arab” origin, locally produced ceramics appear to have shifted toward a generally more homogeneous range of forms.

Moving inland, during the nineteenth-century, locally produced ceramics were quite different in their form and decoration than those of the coast. The majority of comparative material comes from sites which were very closely connected to Zanzibar through caravan trading routes, and yet a sharp divide seems to exist between coast and inland areas. Kwa Fungo was a small settlement that served as a caravan halt during the late nineteenth-century, located near the Mrusa River, close to Mount Tongwe in northeastern Tanzania. It was founded in the late 1860s by a Zigua man named Fungo (the name of the site literally means “Fungo’s Place”), and by the late 1880s it was described as having around 30 huts (Lane, personal comment 2005). As discussed in Chap. 2, the caravan routes of the nineteenth-century were areas along which many porters and slaves travelled, providing interactions between both coastal and inland groups. Local leaders on these caravan routes drew on aspects of Arab identities in order to create themselves as powerful figures in the regional context of nineteenth-century East Africa (Glassman 1995). Archaeological research was carried out at the site over two fieldwork seasons in 1991 (Lane 1991, 1992). Survey was carried out over the settlement area, including subsurface testing with shovel test pits, as well as some small-scale excavations. Analysis of local ceramics from this site provides an interesting comparison with which it is possible to assess how ceramic traditions at Mgoli and along the East African coast were connected to those inland.

Of the total locally produced ceramics recorded from the site, 80 percent were undiagnostic. From the remaining 19 percent, only 2 percent were carinated bowls —the dominant ceramic form at Mgoli and across the coastal region during the nineteenth-century. A comparably small proportion, 2 percent were spherical bowls, a form which may be represented at Mgoli , although the sherds representing these are fragmentary, and the presence of spherical bowls on nineteenth-century Zanzibar needs confirming by further excavations. These may be similar to the restricted bowls found by Marshall on Kenyan watoro and Girama sites (Marshall 2011). The remaining 15 percent of Kwa Fungo ceramics are forms which have not been recorded at nineteenth-century coastal sites. Rather than the carinated cooking pots that dominate the Mgoli assemblage, forms at Kwa Fungo include flared neck pots (the most common form , 9 percent overall by sherd count); open bowls (5 percent); and shallow bowls, plates, and platters (1 percent). The decorative forms found in Kwa Fungo ceramics also differ from those of coastal ceramics, with no arc motifs visible. Graphite burnishing, a finish unknown in any nineteenth-century coastal ceramics, was present in 3 percent of the Kwa Fungo sherds. Although there are a very small number of carinated bowls and pots found at Kwa Fungo, the vast majority of material from this inland site is very different from that of contemporary clove plantation sites and that from the southern Tanzanian coast in the Kilwa region. The forms of these ceramics suggest that while some forms of material culture from the coast, such as house forms and styles of dress, were being adopted at sites along caravan routes , localized production practices continued to differ between coastal and inland regions. This also suggests that the practices such as food preparation, storage, and consumption surrounding these local ceramics would also have differed between coastal and inland areas, as these uses would have impacted on the forms of the vessels.

Our understandings of later historical patterns of materiality inland have been greatly expanded by the work of Thomas Biginagwa (2009) at a series of sites associated with the caravan trade in northeastern Tanzania. This research has been conducted in conjunction with a larger project investigating historical ecologies of landscape related to the nineteenth-century ivory trade in the region (Lane 2010). Ceramic analysis of the sites is not yet complete, but data available thus far provide another glimpse into the sharp differences found in local ceramic practices away from the coast. These sites were part of “a chain of Zigua- abandoned villages” along the Pangani River which date from the seventeenth to mid-twentieth century, and demonstrate that these Zigua communities were active participants in long-distance trade during this period (Biginagwa 2009, p. 58). No ceramics matching coastal styles have been recovered thus far (Biginagwa, personal comment 2011), but material of the “Usambara Mountains pottery group D” style, characterized by open pots, decorated mostly with “raised ‘pimples’ and patterns of dots,” along with graphite-coated and comb-stamped open-bowl forms were recovered from the sites (Biginagwa 2009, p. 54).⁸ These sites were only 90 km from the Indian Ocean , and trade evidence from beads, discussed in the previous chapter, shows close connections with the coast. As with Kwa Fungo, the differences in ceramic assemblage in comparison to the nearby, and well-known, coast are startling. Clearly, there is something of a break in practices of ceramic production—and we might presume cuisines and eating habits—between the coast and inland areas, despite the numbers of travelers from the coast moving up and down these caravan routes .

The final set of ceramic comparisons comes from further inland, drawn from surface survey and excavation at two towns which formed only as a result of the caravan trade . Both were sites which became population centers for the purpose of trade and respite on the long trails. Survey work at the two sites (Wynne-Jones and Croucher 2007) and excavations at Ujiji in 2009 demonstrate that the styles of wares found here are markedly different from those of the coast—yet these were towns located in regions that were major centers for the nineteenth-century slave trade .

Analysis is ongoing, but roulette decorations predominate in the assemblage. This is a classic form of ceramic decoration across much of sub-Saharan Africa (Haour et al. 2010), but not a single sherd of roulette decorated ware has thus far been documented in Zanzibar or in any other coastal region. The striking findings of the excavated assemblage at Ujiji is the diversity of decoration, a sharp contrast to the very limited range of styles being applied to the sherds found at Mgoli . Two types of roulette are present: cord wrapped and knotted strip (this latter type being typical of the mid- to late-second millennium AD—the late iron age—in the interlacustrine region of East-Central Africa (Haour 2010, p. 178).⁹ Along with diversity in decoration type, ware type is also diverse here, in contrast to an almost entirely homogeneous local ware found at Mgoli. The knotted strip roulette and appliqué decorated sherds shown in Fig. 7.3 clearly came from distinct clay sources, showing that the practices surrounding ceramic production and the movement of ceramics (whether by trade or other means) were very different at these Omani caravan halt towns than was the case at the Omani plantation of Mgoli.

Fig. 7.3

Ceramics from Ujiji. Drawing by S. Wynne-Jones

This leads us to a problem that I have been grappling with now for several years (Croucher and Wynne-Jones 2006). This involves a paradox: that the style of coastal ceramics appeared to be homogeneous, but the populations of these areas were radically heterogeneous, as discussed in Chap. 2. Recent material from Marshall’s research in Kenya further complicates this picture, as watoro clearly did not simply purchase or manufacture ceramics in the same way as the communities they had been taken from, coastal communities, or sites located along caravan trade routes. But at Mgoli , unlike these watoro sites, locally produced ceramics were broadly homogeneous in the overarching types being utilized. In the face of massive social change, local ceramics were most stable material forms through time, even as plantations were settled in new areas and large numbers of enslaved Africans moved in alongside Omani plantation owners . The ceramic styles used in different areas did not match up to ethnic boundaries on the coast, although inland at sites such as Kwa Fungo there were clear trends toward ceramic styles that appear to be differentiated by geographical locale. More difficult to explain was the homogeneity of Zanzibari ceramics. Open carinated bowls were used right across the islands. While there were clearly variations in rims and decoration, these were remarkably stable through time and space. How could it be that so many immigrants and changing social relations allowed for such stability of local production ?

Contemporary Production

On the outskirts of the city of Zanzibar in a neighborhood called Kiembe Samaki, ceramics are produced intensely, with women making large quantities of cooking pots, lids, and incense burners, and men producing ceramic cooking stoves, these latter being a new innovation in the islands. Air in the neighborhood is sometimes thick with the smoke of multiple open fires being used to fire pots, mostly tended by women. Orders to women in Kiembe Samaki are called in by cell phone from shopkeepers across the island who regularly buy quite a few vessels at once. Women also come to purchase pots for their daughters’ weddings. A typical set for a wedding would be two cooking pots, two lids, and two incense burners. One 40-year-old potter told me about the nature of her work, which she had been taught as a young woman by her grandmother.¹⁰ She had been taught to make pots in a few different forms: kikaango, a vessel normally used to cook curry and chungu, larger forms of which were used for cooking rice , with smaller versions used to cook vegetables. She left vessels undecorated, but her grandmother had decorated lids by notching around the raised edge. She told us that some of the other potters in her neighborhood also decorated their wares. This particular woman had always worked as a potter, and even in the past, when there were fewer orders, the trade had provided enough for both her and her grandmother to live on.

This was not the only area in which pots were made, although Kiembe Samaki today is a site of intense production. The potter I spoke with in 2011 knew of at least two more areas in which potting took place. A woman in her sixties who lived in the Kikwajuni neighborhood of Zanzibar had discussed with us the origin of pots she had used to buy in the market. She knew that these sometimes came from Kiembe Samaki, but she was also aware of potters who bought their wares in from the countryside to sell in town. This woman was interesting as she had also lived on Pemba for a time and had moved back to Zanzibar. She had bought and used cooking pots on both islands and was certain that there were no differences between each area.

This manner of production was not limited to Zanzibar. On Pemba, I interviewed a potter called Fatuma Mohammed, who was the head of a relatively large operation (Fig. 7.4). She had been taught to make pottery by her grandmother, as had been the case with the potter in Kiembe Samaki. In the 1990s, with government encouragement for women’s cooperatives, Fatuma Mohammed began to expand her business. This began with just two other women from the village , as demand was relatively low and they continued to make the styles her grandmother had taught her. These were the same types of cooking pots as were manufactured on Zanzibar. In the early stage of their business, the women earned only a little from producing ceramics, but now Fatuma Mohammed and the other women living in her village dominate the Pemban market for local ceramics through the work of their cooperative. This growth in business means that they now produce the common carinated bowls and lids every day. The carinated bowls made by Fatuma have a small single arc decoration above the rim, which she includes because this was the way in which her grandmother had taught her. Her grandmother also used to finger impress some of the rims of pots, but Fatuma no longer does this to the vessels she produces. She also sometimes paints red slip to decorate her wares, although only occasionally on cooking pots. Shopkeepers and market stallholders come to this village in order to purchase wares, with peak buying periods during the clove harvest season in July to August, and during Ramadan.

Fig. 7.4

Ceramic production in Pemba

These women have also been creative in broadening their market. The Pemban women have also begun to make new styles which broaden the repertoire taught by an older generation. As on Zanzibar, Binti Mohamed and her colleagues are now making incense burners. These are small clay pots, often with red slip decoration and holes carved into the side of the clay body. They hold hot coals and powdery udi (incense) together to generate fragrance for bedrooms and scenting women’s clothing. As was noted by potters on Zanzibar, while there are a range of alternatives to clay cooking pots, mostly in the form of aluminum pans, business in these locally produced wares is booming. In the 13 years which had elapsed since the formation of the cooperative, these women had managed to carve out a wide regional trade in their wares, even sometimes selling some to Zanzibaris (Croucher 2006, p. 511).

Specialized neighborhoods, including the cooperative scheme on Pemba , seem to be increasingly dominating the market for locally produced ceramics, but they are not the only producers working on Zanzibar today. Visiting a small village in northern Pemba, I interviewed two women who only occasionally make pots, selling only to women from neighboring villages. As with the other potters I interviewed, both of these women learned their craft from their grandmothers. Both left their pots plain, with no decoration because, they told me, this was how they had learned from their grandmothers. Although they had seen decorated cooking pots, they chose to leave theirs plain, as this was the manner in which they had been taught. As well as the undecorated cooking pots, these women had just finished an order for a man from a neighboring village. This consisted of several small pots, which the women thought would be used for “making medicines” or giving offerings to spirits (Fig. 7.5; Croucher 2006, p. 513).

Fig. 7.5

Medicine pots

Along with the ceramic data, these four women who manufactured ceramics provided valuable evidence on the nature of potting on the islands. These interviews were taking place over a hundred years after the ceramics I am discussing archaeologically and it would be impossible to gloss over the social changes that have occurred during this period. Nevertheless, they, and other interviews carried out with potters in the East African coast (Wynne-Jones and Mapunda 2008), provide possible insights into the social relations that may have been involved in the production of local earthenwares during the nineteenth-century .

Drawing on the interviews as a group, it is clear that the manufacture of pottery operates with a significantly gendered division of labor on Pemba today. Although men may contribute labor in digging clay and collecting firewood, they never actually manufacture pots (with the exception of the new large clay stoves manufactured at Kiembe Samaki). All of the women I spoke to had learned to make pots from their grandmothers. This suggests that kinship links may have been an important route through which women were able to take on an apprentice role in learning their trade. Designs used to decorate pots (or not used, some left their pots plain) appear to have been taught alongside the general lessons of craft production, these younger potters usually reproduced decoration in the style which they had been taught with the only potters to leave their vessels plain doing so because it was the style of their teacher (Croucher 2006, pp. 511, 513).

Zanzibaris who were not potters differed on their views about decorative style . One elderly man claimed that all ceramic designs were the same all over Pemba and that it was not important where the pots came from. An elderly woman said the decoration was put in ceramics just as decoration is put on kangas ¹¹ “for attracting people”; consumers of these goods selected them on the basis of these styles, with particularly “nice” ones chosen for weddings. She also remembered more variation in the design on ceramics, having seen triangles and arcs which matched the designs on waterpots etched into vessels (Croucher 2006, pp. 515, 529).

The purchase of ceramic vessels was discussed uniformly as occurring on a small scale (in contrast to Fatuma Mohammed’s current operations). Many more women in the past used to make pots it was claimed, but this would always be a part-time activity, carried out in conjunction with work on farms or similar kinds of labor. Most ceramics were bought directly from the potter herself, who would be located in the same village or a neighboring village as her customer. Having one potter in every village, or at least having one located in a neighboring village, used to be the common pattern. Even with the changes in production that have expanded potting in some locations, there are still some potters today only selling to neighboring villages, with no traders from far afield coming to buy their wares. Where settlements were larger, several women might be able to manufacture ceramics, such as the large village of Tumbe where an elderly lady told us that there used to be three potters in the village and there continue to be several today, all working with their daughters. This small scale of production and sales facilitated close relationships between potters and their consumers, with ceramic vessels often only made in response to a direct request. The purchasers of these wares were then also able to ask for specific designs. One man remembered that there used to be many local potters, some of which used to sell them to Indian traders. Another elderly lady from Pemba recalled that in the village that she grew up there used to be a potter. If a person wanted a pot they could go and order them in advance, although if the potter fired a lot of pots for which there were no customers “they might take the extra pots and sell them in a store and town” (Croucher 2006, pp. 511, 513, 515, 517, 520). Such potters would be able to take direction from some customers but would also produce vessels for women with whom they had no direct contact.

Although it is possible only to use these histories in a heuristic manner (Brück 2005, p. 151), they provide a starting point with which to begin to understand the community relations within which ceramic manufacture took place on Pemba . Clearly, ceramic manufacture was a gendered activity, and was carried out widely across the island. Virtually no documentary history can link us back to nineteenth-century ceramic production, for it was most certainly no concern of the explorers and missionaries who made up the majority of those writing about the islands. But a single consular letter from the late nineteenth-century confirms that at least one potter was a formerly enslaved woman. It discusses the case of a woman named Mia, who had originally been enslaved, and was owned by Tippu Tip . She was freed at Bagamoyo (presumably the mainland coastal town) in 1883 by Tippu Tip, but was then kidnapped by an “Arab slave dealer” who set sail for Pemba. She managed to escape from the slave dealer, whose boat was almost captured by a British ship trying to prevent the slave trade , as although slavery was still legal on Zanzibar, the trade in slaves from the mainland had been banned following a British treaty in 1873. Having fled:

“Mia” made her way to Chaki Chaki [Chake Chake] and sought the protection of the then wali, Basuk bin Bedwi…who on hearing the story gave orders that no one was to enslave her…. About three years ago [1893] she went to live upon a small plot of land which she rented from one Binti Hamadi, al Banani, a spot situated upon the southern coast of Chaki Chaki bay. She supported herself by making and selling earthen pots. (ZNA, AC 5/2: 1896, PCVC, Letter from BVCP)

It is a tentative claim, but the patterns of local production on Pemba today and their historical recollections seem to suggest that it would have been likely that women resident in plantations may have been making vessels. The scale of ceramic production is, and was, localized for the most part, and thus we might expect that women living far from large indigenous Swahili coastal villages would have taken up this trade. Clay sources are readily available on the islands, and three of the four potters I interviewed were in clove plantation areas of Pemba. In addition, the one archival mention of a potter makes it clear that at least one enslaved woman was able to learn the local patterns of ceramics to a degree where she was able to manufacture a few pots that were saleable to local communities in order to help support herself.

Further interview evidence collected by Stephanie Wynne-Jones and Bertram Mapunda from contemporary potters on Mafia Island, located not far from the southern Tanzanian coast, also supports the idea that immigrant women may have readily been able to learn the skills required in order to make pots in styles local to the areas in which they now lived. Out of the 28 potters interviewed on Mafia, 27 were immigrants, quite a contrast to the Pemban-born potters I interviewed. These potters however, made their wares in almost exactly the same forms as is seen in the coastal repertoire of the nineteenth-century , and of contemporary Pemban wares. Despite having knowledge of coil built pots, Makonde immigrants adopted the method of drawing pots out from a lump, as characterizes coastal ceramic technology, as well as copying the form and design of coastal wares (Wynne-Jones and Mapunda 2008, p. 5). They learned this technology from female kin and nearby neighbors, but not in an inherited pattern, rather, these networks were an “enabling factor” which meant that Makonde immigrant women were able to learn the new practices of a ceramic technology and style in order to make cooking pots in forms that conformed to the tastes of the wider Mafia (nonimmigrant) community (Wynne-Jones and Mapunda 2008, p. 10). Since in questioning, Makonde potters replied that they made ceramics in this style simply because “this is what pots look like on Mafia,” their adoption of almost the entire chaîne operatoire of ceramic production in a coastal style was analyzed by Wynne-Jones and Mapunda as a reaction to the “market” for pots on Mafia. As such, Makonde immigrant potters functioning as “rational actors who exploit the economic opportunity provided” by particular types, while simultaneously signaling “their inclusion in a society defined by a particular pattern of consumption” (Wynne-Jones and Mapunda 2008, p. 14, my emphasis). The way in which immigrant women on Mafia were able to learn to make pots in a new style, ceasing to use the technologies they had learned in their homeland, provides an example of a potential route through which women such as Mia may have learned their craft during the nineteenth-century .

Food and Identity on Zanzibar

To interpret the reasons for the homogeneous coastal ceramics of Zanzibar, in contrast to those found inland, I want to think about the practices with which they were engaged after their production . It is easy to get caught up in production, and clearly the potters themselves and their social networks were a crucial reason for the material signature viewed archaeologically. But ceramic vessels were, and are, made to be used. Taking up new ceramic styles was not like putting on a new pair of clothes ; food itself is culturally defined and is a corporeal act of incorporation involving senses, feelings, and emotions, the preparation and serving of which can be a crucial site of social engagement (Counihan 1999; Goody 1982; Hamilakis 1999, p. 39; Stoller and Olkes 1986; Weismantel 1988). To begin to understand the role that locally produced ceramics may have played beyond simply their style of production, a brief foray into understanding food practices on Zanzibar is essential. As was discussed in the previous chapter, Zanzibaris are proud of the dishes they cook, particularly rice-based dishes of pilau and biryani. The heritage of eating rice on Pemba goes back to the early first millennium AD, where ethnobotanical work has shown a striking move from a subsistence economy based on pearl millet (an African cultigen) to one based upon Asian rice (Fleisher 2010, p. 204; Walshaw 2010). Thus, the consumption of rice would seem to have long-term historical antecedents as an important cultural practice on Pemba. This may have distinguished the area itself, since Pemba, with wet valleys between the rolling hills which are so good for growing clove trees, presents an ideal landscape for rice growing .

In the area of Mombasa , Kirkman (1974, p. 81) noted that the common foods of the mid-twentieth century were millet, rice , and bananas. Crucially, he describes the fact that “The carinated bowls , particularly those with inward sloping necks and sharp carination, would seem to be less suitable for millet porridge…which must be stirred.” He suggests that these vessels were used for cooking meat or fish. On Zanzibar, millet was mentioned only in coastal areas as an important food. In plantation areas no one mentioned eating millet; the problems Kirkman identified with the use of these vessels for cooking millet were probably not an issue, as rice is placed into a pot with water and sometimes spices and then left to cook. It requires no continual stirring. Ceramics are the preferred medium for this process, even today, as they allow cooks to be able to tell whether their rice is becoming dry as it cooks (Croucher 2006, p. 517). The ceramic pattern leaves us with a record of the day-to-day food preparation and consumption patterns of one plantation.¹² Millet and other African grains seem to have been an unlikely element of the diet of the plantation household . Instead, rice accompanied by fish, meat, vegetables, and sauces would most likely have been the predominant foods for plantation households such as Mgoli, and likely for the majority of the population of the islands back into the nineteenth-century . Evidence is provided for this by the stability in ceramic forms through to the present day used to cook curry, vegetables, and rice . Certainly, the foods being eaten at watoro sites such as Koromio and Makoroboi would have been prepared and served in significantly different ways than at Mgoli (Marshall 2011). The similarity in ceramic style between Kilwa, to some extent Fort Jesus, and Pemba, suggests that these coastal areas may have shared at least some practices of preparing and serving foods. Moving inland drawing on ceramics from sites along caravan routes , foodways also seem to have differed significantly from those on Zanzibar and in the Kilwa region (Biginagwa 2012a, b) .

Within contemporary Zanzibari society, the foods immigrants eat, the ways they cook them, and the vessels they cook them in may be a crucial arena of expressing or minimizing difference. One man interviewed recounted his changing place in Pemban society through changes in foodways and tastes. When he arrived as a young man in the mid-twentieth century, he and other Nyamwezi immigrants would cook in the metal pans that they had brought with them from their home, and would eat ugali (maize meal porridge, now the staple food across much of East Africa) directly from these, a practice they regarded as part of their traditional culture. When he first arrived, he found that for himself and other Nyamwezi immigrants “it wasn’t difficult to eat ugali twice a day, because it’s easy to cook and we grew up with it.” By contrast, other Pembans would not eat ugali, explained simply as the fact that they “didn’t like eating ugali, that’s why they didn’t eat it” (Croucher 2006, p. 302). Over time, the tastes of this man changed, and now it’s rare that he ever eats ugali, for he told us that far prefers rice , so long as he can afford to buy it. This change in taste has taken place alongside his conversion to Islam, and a growing identification with other Pembans . Thus, it seems that with the shift away from ugali, a food commonly identified by Zanzibaris as typical of mainlanders, and toward rice—the longtime food of Pemba —this man and his fellow Nyamwezi immigrants began to see themselves as more like Pembans. While he did not reject his origins, he was clear that he had become a Muslim, had become comfortable with life on Pemba, and that his food preferences had shifted with this change. In Tumbe, located in northern Pemba near the coast, an elderly man whose family had a “long history” discussed the contrast understood to broadly exist between mainland immigrants and Pembans. In Tumbe, unlike other areas, he told us that millet had been the most popular crop grown, followed by rice. In Tumbe, people did sometimes eat ugali , but it was “made from cassava, not maize.” He continued to tell us about differences he understood between immigrants and mainlanders, sometimes struggling to quite capture actual demarcations in food between different groups:

The people of Tumbe don’t like ugali from the mainland. Also, it wasn’t easy to get the flour for ugali from the mainland [maize meal ugali], so even immigrants from the mainland would eat cassava ugali. But the food preferred by Arabs and Pembans was millet and rice . Fish was the main food eaten with these in Tumbe, although sometimes chicken and beef might also accompany it. (Croucher 2006, p. 514)

As was discussed in Chap. 6, food was an important way through which Zanzibaris came to understand ethnicity. In the case of migrants from the African mainland, the divide was not so sharp as that perceived to divide Arabs and Zanzibaris from Indians . Mainland immigrants might effectively assimilate into Zanzibari society, as with the Nyamwezi man we interviewed who had come to marry a woman from Pemba . Likewise, “Arabs” were discussed as a particular separate category, but within conversations about food and daily practice, their similarity to Zanzibaris was always stressed. Coming to use local ceramics to cook with therefore appears to have a particular salience for the way in which foreigners could be seen to be assimilating into Zanzibari society .

Communities of Practice

Despite the importance of coming to eat local foods, conservatism in ceramic style seems strange in coastal communities so enmeshed in rapid social change. Why would enslaved plantation laborers come to so wholly take up the style of their Swahili neighbors? Why would Omani plantation owners come to use ceramics in styles that were so typical of the coast? Aside from the form of stone houses , almost no other material culture of daily life seems to give any indication of the large numbers of Arab immigrants to plantation areas. Just 19 sherds of a light buff earthenware with incised design were recovered which are most likely from a water jar imported from the Middle East .¹³ Locally produced ceramics tend to be an object that in studying colonial societies we place with “traditional” practice. Unlike mass-produced imports such as the brightly printed spongewares used on plantations, the fact that individual women were making these wares seems to leave open the idea that enslaved and immigrant women would innovate with ideas from their earlier experiences. Such innovations and small changes have been suggested to be one route through which we might locate in captive women in particular societies (Cameron 2011), a point I will return to later in this chapter. But here innovations are startling for their absence. It seems unlikely that potters were acting simply as “rational actors” in making vessels in the traditional forms of Zanzibar, since immigrants by far outnumbered indigenous residents in plantation areas. Why would they use ceramics in coastal forms, when they knew that other options were available, and when women who were cooking—or possibly directing cooking, in the case of freeborn plantation wives —knew of other viable forms which could have served as alternates. What might be the importance of the stability visible in ceramic traditions throughout nineteenth-century Zanzibar?

I am not the first to note that the place of the “traditional” in the field of historical archaeology is not always straightforward. At the nineteenth-century colonial site of the Rancho Petaluma in California, amongst mass-produced goods such as glass beads, bottle glass, and ceramics were found a range of “traditional” native artifacts; lithics, shell beads , and worked bone, with an apparent continuity of precolonial lithic technology. Since such a wide range of mass-produced material culture was available, and no expedient tools were made from glass, Stephen Silliman (2001), the excavator of the site, suggested that the stone tools were a continuity of practice not for functional reasons, but as an active materialization of native identity. The practices of these technologies were important in this context, with men needing to gather stones from particular sources, in contrast to the fact that glass—a highly functional material for making tools—was readily available. Since other aspects of indigenous material tradition were not continued though this period (women seemingly rapidly adopting imported goods such as scissors, needles, and thimbles into their daily tools), Silliman argues (2001, p. 215) that the continuity of “tradition” here became a discursive practice in the activation or solidifying of a nineteenth-century indigenous identity. Cases are also documented where “traditional” production expands in relation to colonial and capitalist relations, particularly where locally produced goods may also find a new market amongst collectors, alongside continued local use of such artifacts (Harrison 2002).

In both of the cases cited above, interpretations of the importance of local production were based upon the relationships formed through the actual practice of production. This could be framed by the arguments of Bourdieu (1977, p. 164), whereby new norms of repeated practice had to be created to appear to produce the world of tradition . In this manner, the social structure of plantation life on Zanzibar could naturalize “its own arbitrariness” (Bourdieu 1977). But, despite the possibility of improvisation within Borudieu’s theory of practice, the rigidity of the way in which social structure and reproduction was envisaged by his theory seems to fall short of explanatory potential for the kinds of rapid social change occurring in Zanzibar (Joyce and Lopiparo 2005). Imagining an ongoing social structure in which small changes are possible, as posited by Bourdieu, still does not seem to leave space for the actions of nineteenth-century Zanzibaris, as “traditional” ceramic forms and foodways of Pemba were taken as traditional practice for all immigrants in plantations, elite Omanis and enslaved Africans alike.

Making the move into utilizing practice theory helps us to break out of interpreting material culture only through its formal and stylistic qualities. Ordering archaeological material culture into taxonomic units based on resemblance and visual observation is at the heart of the archaeological project and has been a part of the process of creating our objects of study as objects, whether they be material or living (Fabian 2002/[1983], p. 33; Meskell 2004, p. 40). Yet the use of “style” within archaeology has a troubled history, to the point that challenges have been raised as to whether style is capable of capturing the complexities of materiality captured within the artifacts we study (Boast 1997). Instead, styles do not come into being wholly formed, but are created as the production of a network of social relations (Dietler and Herbich 1998; Dobres 2000). Archaeologists have recently turned to the theory of communities of practice in order to examine the way in which networks of knowledge are materialized through the production of goods such as ceramics (Gosselain 2008; Gosselain et al. 2010; Cordell and Habicht-Mauche 2012). The idea of communities of practice draws out of theories of learning as an ongoing process. Negotiation of meaning within a community of practice is a diachronic process, as such negotiation “always generates new circumstances for further negotiation and further meanings,” and therefore members of a community of practice do not have a fixed set of meanings, but instead participate in “a continual process of renewed negotiation” (Wenger 1998, p. 56). We may also consider the argument put forward by Appadurai as to how to retain practice theory within a contemporary frame, where it becomes increasingly difficult to view the structural continuities of practice. He suggests that we retain the importance of habitus , but that we should also place greater emphasis on Bourdieu’s idea of improvisation which “no longer occurs within a relatively bounded set of thinkable postures but is always skidding and taking off, powered by the imagined vistas of mass-mediated master narratives” (Appadurai 1996, p. 55). On nineteenth-century Zanzibar, potters would not have had “mass-mediated master narratives,” but their work would have been able to move beyond the “thinkable postures” learned from their grandmothers, or for enslaved potters , learned from their childhood or early lives.

Taking it as a given that ceramic production was widespread among Pemban women on nineteenth-century Zanzibar, both free and unfree, and that new techniques of ceramic production and of food preparation and foodways were learned by immigrants through communities of practice, our idea of what it might mean to see a continuation of precolonial ceramic traditions on the coast changes. Rather than the reproduction of a static set of meanings, what it meant to make and use these artifacts was part of participation in new communities by immigrants. Making and utilizing new ceramics and foods would have been mediated within the “dynamics of interpersonal consideration,” where immigrants entering new communities were aware of how others around them would react to their learning, rejection, or transformation of community practices (Farnell 2000, p. 405). Conforming to coastal styles and practices may therefore have been a way of appearing to accept that varied immigrants now shared in new Zanzibari community relations.

Such participation depended on being able to enter into a community in order to learn these ongoing sets of shared knowledge. A potter, for example, had to learn the steps of shaping clay, decorating it, and firing it, according to Pemban conventions, as well as having knowledge of what the users of her pots desired. These relations of apprenticeship into a craft, given the close kinship or neighborly ties that are historically known to exist for contemporary ceramic production on the East African coast, suggest that this may also have been a way in which kinship bonds were able to emerge in a society where many so many immigrants were coming together.¹⁴ That ceramic styles are so broadly shared along the coast, and yet stop so suddenly inland, may suggest to us ways in which communities of practice formed into constellations of practice on a wider social level, where face-to-face community relations did not exist, but where related participation in foodways and ceramic production coexisted and shaped one another (Wenger 1998, p. 131).

Within this interpretation, adoption into the community may have been a tacit process. Members may not have been rational actors, consciously taking on the practice of Zanzibar so as to better their situation. As individuals entered a bewildering social context, in which new social relations had to be forged, communities could be newly formed and/or expanded through participation in practices such as the production, exchange, and use of ceramics. In this manner we see that society is shaped by practice, where practice is an ongoing and changing process in which meanings are not fixed, and is a social structure in which enough flexibility exists for the stark changes wrought by the capitalist colonial relations of clove plantations to still seem to offer a continuity of everyday life for residents. The role of these ceramics as part of the new process of understanding of Pemban and Zanzibari identities cannot be fully comprehended from an archaeological vantage point. That shifts in the arrangements of different communities who viewed themselves as having shared identities was occurring on clove plantations is certain. As potters made their wares for neighboring women, perhaps in response to specific requests, ceramic style may have become a reification—if only temporarily—of a notion of shared identity through knowledge of the acceptance of shared tastes (Wenger 1998, p. 58).

The materiality of ceramics is important; they were meaningful as goods that were made, exchanged, and utilized for everyday tasks, and that existed in a repertoire of shared tastes (Stahl 2002). Such materialities do not only reflect social relations but are also constitutive of such relations and of subjectivities themselves (Dant 2006; Meskell 2005). Objects may, in particular contexts, be powerful and have effects beyond those which we might immediately recognize (Gosden 2005). In this colonial context, the continuities of ceramic style on Zanzibari plantations are a representation of a powerful object. They differ from the way in which coastal culture was used by watoro (maroons) in Kenya . As discussed in Chap. 5, Kenyan watoro utilized coastal styles for building, as plantation slaves did, seeming to demonstrate visible ways in which enslaved and free mainland Africans presented themselves as part of Swahili societies. But the ceramics we find at sites such as Mgoli were used by all who lived in plantation households. Nineteenth-century Zanzibari ceramics were embedded in lives through the practices of production, and the effects that they created in terms of the food preparation, of consumption , and via their role in foodways, in helping to shape a new corporeal realm of tastes where immigrants came not to like ugali anymore and where Omanis cemented the fact that they ate the same food as other Zanzibaris through their use of the same cooking techniques.

Slavery and Ceramics

Ceramics do not simply matter for interpreting social relations on Zanzibari clove plantations. They have been an important marker for understanding the ways in which plantation slaves in the Americas may sometimes have found ways to continue African traditions. This topic has been explored through the materiality of colono wares , a broad category encompassing a range of locally produced ceramics found by archaeologists in the colonial Americas, including the Atlantic seaboard, the U.S. South, Southeast, and Southwest, several islands in the Caribbean , and in many locations in Latin America. I am not trying to give a complete summary of these artifacts, and am particularly interested in those which are found in plantation contexts which are in some ways comparable with Zanzibari clove plantations. What flows through the discussion of colono wares is their identification as manifestations of the hybrid nature of colonial relations in plantations. They are seen to be hybrid as preexisting cultural practices (i.e., the African knowledge of enslaved laborers) are identified as continuing on in changed forms. Hybridity emphasizes the fact that the social relations of colonial contexts fall “in-between” any binary of colonizer/colonized (van Dommelen 2005, p. 117). Despite this acknowledgement, colono wares remain a common location for identifying cultural continuity in enslaved populations of African descent.

Not all colono wares are assumed to have been made by slaves in plantations. But within plantation studies there tends to be a broad consensus on this group of material. As Mark Hauser summarizes in the introduction to his recent work on a specific Jamaican iteration of these wares:

These kinds of earthenware have received considerable attention in the archaeology of the African Diaspora and have been called “colono-ware,” “colonoware,” “Afro-Caribbean ware,” “yabbas,” and “Criollo ware.”… They are one of the few types of material culture that were predominantly made, used, and traded by enslaved laborers, probably women, and that survive the archaeological record. As such they can speak to the silences implicit in the documentary record about the social relations and networks that the enslaved created as they negotiated their lives and worlds. They became a material way through which the archaeologist can glean insight into the ways in which the enslaved transformed the world around them. (Hauser 2008, p. 94)

This definition provides us with a general bracketing of how these ceramics have come to be understood within African diaspora archaeology in particular although the term “colono ware” also spreads into non-African diaspora contexts. As Hauser (2008, p. 94) goes on to note, even within African diaspora archaeology the traditions subsumed under the label of colono ware show considerable “idiosyncrasies,” meaning that it is “not a single type, except as a useful device for archaeologists to categorize ‘other ceramics.’”

When first identified on plantations in the U.S. wares like this intrigued historical archaeologists. Working on functional interpretations, these wares were first assumed to be associated with enslaved populations as they were of an inferior form to the glazed wares more commonly purchased by Europeans (Noël Hume 1962, p. 7, cited in Ferguson 1992, p. 6). By the 1990s, African origins were clearly understood to inhere in colono wares. Fundamental to this transfer was the fact that Africans were understood to carry from Africa “craft skills essential to daily life” (Ferguson 1992, p. 16). The materiality of African American produced material culture was now seen to embody (static) elements of African culture through a variety of decorative and formal attributes (Emerson 1999; Deetz 1993, Chap. 4; Fennell 2000; Ferguson 1999).

Parallel developments had taken place in the archaeology of Spanish colonialism . Early advances in understanding locally made wares as part of a process of colonial hybridity (termed, in this case, mestizaje) were made by Kathleen Deagan (1973, 1983) as she developed a theoretical framework for interpreting archaeological remains from Spanish Florida and early Spanish colonial sites in the Caribbean . In the early 1970s, while to the north archaeologists of Anglo-America had slotted low-fired earthenwares to the production of Native American populations, Deagan recognized that these were the product of intercultural relations, specifically those mediated by gender. Deagan’s work is remarkable for the early attention that she called to gender in the archaeology of colonial relationships, even if this was of a fairly normative and rigid framework (Voss 2008). Deagan (1973, p. 57) pointed to the marriage or concubinage of indigenous women with Spanish colonizers as fundamental to understanding the archaeological record of colonial sites in St. Augustine, arguing that it would be in the domestic assemblages that the archaeological presence of these women would be expected to be visible. Thus, the “admixture of Spanish and material cultures…was most strongly reflected in the ceramic assemblage…of English earthenwares, Spanish Majolica and aboriginal pottery. San Marcos ware was virtually the only utilitarian ware present. This was an Indian-made ware, no doubt easily accessible for the inhabitants of St. Augustine to use as kitchen pottery” (Deagan 1973, p. 62). Sooting on San Marcos ware was utilized to demonstrate its use in cooking, which Deagan argued would have been carried out by indigenous women in these mixed households, thus making “food preparation techniques and technology” “predominantly aboriginal, with a minimum amount of Spanish influence” (Deagan 1973). This work was then carried into the archaeology of the Spanish Caribbean (Deagan 1995), where similar patterns of the presence or absence of indigenous persons was traced through the ceramic evidence. In the archaeological record of Puerto Real, the changing labor force from the indigenous Caribbean population to that of enslaved Africans was also argued to be visible in changing ceramic forms (Smith 1995). Deagan’s legacy has been hugely important in drawing attention to locally produced wares as potential continuities in cultural practice. However, apparent in all of these interpretations is the easy equation of ceramic style (and by extension food preparation and/or serving techniques) with ethnic and cultural identities. Thus, the mestizaje households of the Spanish colonial world can be identified as such precisely because of the merging of Spanish/male and Indigenous/female material culture, creating a problematic gloss over the complex relationships of colonial relations and materiality (Voss 2008).¹⁵

The tying in of locally produced wares specifically to a female domestic sphere was to prove influential in the developing archaeology of colono wares in general, as increasing attention was paid to the fact that colono ware forms seemed to relate to the necessity of enslaved populations to make their own foods. The cultural dimension of colono wares in this situation shifted to the domestic realm; African Americans made foodstuffs that often conformed to the prior practices of African cuisines, particularly cooking food in shared dishes rather than the individual portions of European foodways. Particular African American cuisines developed in the American context that were structured by the requirements of enslaved laborers to perform their regular labor tasks, and in the new foods available to them (Battle-Baptiste 2007; Franklin 2001a, b; Steen 1999; Wilkie and Farnsworth 2005, Chap. 7). This has marked an important change in allowing colono wares to become part of a dynamic cultural tradition rather than being trapped in the fixed Africanisms characterizing earlier approaches.

Despite the turn toward practice theory over cognitive approaches, and the linkage of colono ware traditions to ongoing transformations in enslaved foodways and quotidian practice, in discussion about precisely why it is so difficult to attribute ethnicity to colono wares , we still see some of the original assumptions about the nature of these artifacts. One recent and highly sophisticated analysis, which highlights the complexity of the routes via which enslaved Africans came to the Americas, notes that “part of the problem with attaching ethnic identities to material culture has to do with the state of knowledge about archaeological analysis in West Africa by North Americanists and Caribbeanists” (Hauser 2008, p. 98). Archaeologists seem to flip between calling attention to the need to be attuned to the complexity of difference in West Africa¹⁶ and the idea that still, figuring out the colono ware (and pipe) question would depend for its success “on access to better comparative material from Africa” (DeCorse 1999, p. 152).

The separation of locally produced items as a particular kind of object seeming to mark a developmental status is a common one for historical archaeologists where objects are often differentiated into categories of Native/Indigenous, European/colonist, and hybridized forms (Silliman 2009, p. 213). In the field of plantation studies, these distinctions are shifted a little as dichotomies are often implicitly understood between slave and free, African and European descent, and plantation owner and laborers. The majority of artifacts in plantation sites in the Americas are of mass-produced origin, but there has been a search to find the type of hybrid or crossover objects (such as materials placed in ritual caches, as discussed in Chap. 4) which can be attributed to enslaved individuals or communities of African descent. As discussion on ritual practice made clear, it is difficult to categorize this material as a direct result of cultural practices transmitted from Africa. The majority of anthropologists now agree that such practices were a hybridized. Utilizing particular material objects as part of ritual or religious practices, or using space in particular ways, were not simply passed down from first generation Africans to their descendants. In the Americas, the intermingling of different cultures often resulted in new forms of practice which drew on prior traditions and yet were different from them in sometimes significant ways.

Two decades ago, cultural anthropologists came to terms with the fact that a “savage slot” haunted their field, caused by the boundaries set up in the Enlightenment age between the West and others (Trouillot 1991). Archaeologists are continually grappling with the fact that they deal with multiple temporalities and geographies within a single discipline, often causing particular eras to be blocked into “historic,” “prehistoric,” “colonial,” or “contact” periods. The disciplinary drive toward delineating categories has, I am arguing, caused colono wares to be placed outside of the temporality of modernity in which plantations actually existed. The Zanzibari material discussed in this chapter is a perfect example of the way in which “traditional” ceramic manufacture is continually part of a dynamic cultural context. On Zanzibar, the use of locally manufactured ceramics is not simply an index of the degree to which Zanzibaris progressed forward into modernity. Instead, they were part of complex social dynamics, with their manufacture and use shaped by relations of gender and ethnicity.

Pemban Identities

Locally produced ceramics on clove plantations are perhaps in some ways less an artifact of tradition than were the mass-produced goods discussed in the previous chapter. While spongewares were a new item for East Africans in the nineteenth-century, they seem to have been used to help create and maintain older systems of power. Mass-produced ceramics allowed for the continual recreation of client–patron relations and networks of reciprocity and dependence. By contrast, the “traditional” locally produced ceramics in the open-bowl forms allowed for women in particular to engage in new ways of cooking and eating. These allowed them to begin to participate in new forms of Zanzibari identities. As discussed in Chap. 2, by the early twentieth century a new kind of identity was emerging on the coast. A specifically Swahili identity was often taken by former slaves so as to demonstrate the fact that they belonged and were fully a part of coastal culture. As was demonstrated by the almost complete absence of Zanzibaris who identified their own ancestors as having been enslaved, the process of assimilating into a new identity appears to have been remarkably successful.

Women were at the forefront of claiming this identity. New forms of fashion such as the kanga allowed women to dress in ways that showed that they were pious Muslims (Fair 1998; McCurdy 2006). This was similar to the process of building houses in coastal styles and the broader adoption of Islam , as discussed with reference to plantation landscapes in Chap. 4. Cooking has been absent so far from this discussion, but it was clearly a vital element of the ways in which traditions became newly (re)created so that immigrants could find ways of placing themselves as legitimate members of coastal society. Clearly, this was not a straightforward process, as the struggles over dress and religious practice make clear. But through forging bonds with neighbors and learning how to produce and cook with ceramics, it may have been possible for women to find ways to place themselves into the ever-changing definition of what it meant to be Zanzibari. As one elderly lady told us, in the past “All people ate the same” (Croucher 2006, p. 520). Clearly, for many who arrived in the nineteenth-century, rice was a new food, or the ways in which Zanzibaris had traditionally cooked rice and curry dishes were new. And yet, women and men rapidly seemed to assimilate to ways of cooking and eating that allowed for a certain kind of homogeneity to be understood between a large mass of people, a shared identity that would allow many enslaved people to argue that they were Pemban or Zanzibari.