Abstract
In this article we examine the role of informal settlements inhabited by Europeans, Africans and, potentially, indigenous people in the eighteenth-century insular Caribbean. Rather than simply being frontier settlements established in anticipation of formal colonization, in many cases settlements on and beyond the margins of colonies represent alternative possibilities and facilitate ways of life, modes of production, and means of trade and exchange that are at odds with expected norms of colonial society. We view such settlements as holdouts, practicing what James Scott refers to as the ‘art of not being governed’. To make this argument we compare ethnohistorical data related to settlement patterns in St John and Dominica and archaeological data retrieved from household excavations of plantation settlements dating to the eighteenth century. Examining such settlements allows us to map the range of variation in colonial life during the apogee of plantation-based slavery.
Keywords
Introduction
What do we call a settlement that is outside the territorial limits of a colony and what are the implications for residents whose material worlds are separate from yet linked to broader networks of colonization? In the context of the Caribbean, research has focused on colonial settlements marked on cadastral maps (e.g. plantations, urban centers, and port towns) or frontier settlements established for planned colonial expansion (e.g. missions, forts, and trading posts). Yet, a range of settlements exist established outside the direct political control of colonial administrators. These settlements are variable in nature and their inhabitants came from a range of backgrounds. Today it might seem common sense to view such settlements as consequences of imperial efforts. But, what if we were to try to understand the material record outside a narrative in which colonialism is a fait accompli? Following James Scott’s (2009) recent insight on the geographies of colonialism and resistance, we argue that inhabitants were practicing the Art of not Being Governed. Rather than depicting such settlements as frontier colonies, we argue that inhabitants built structures, grew provisions, and engaged in trade between colonies as a tactic to redefine the terms under which they engaged or did not engage with colonial administrators.
We compare two such settlements in the insular Caribbean to locate some of the tactics that residents employed to mediate the relationships between existing colonies. On St John, we explore a coastal sugar plantation at Cinnamon Bay with discrete archaeological contexts that contain evidence for occupation nearly a half century prior to formal Danish colonization in 1718. On Dominica, we examine Bois Cotlette, a plantation with evidence of settlement 20 years prior to formal colonization by Great Britain (Figure 1). To make this comparison we utilize ethnohistorical data related to island settlement patterns and archaeological data retrieved from plantation settlements documented immediately after formal colonization of each island. Due to their difficult terrain, islands such as St John and Dominica were colonized relatively late. Yet early on, European settlers from colonial centers such as Jamaica, Saint-Domingue, Barbados, and Martinique established semi-permanent settlements to extract timber, fish, or grow crops.
Location of St. John and Dominica in relation to international boundaries established after the treaty of Paris in 1763. The location of Bois Cotlette and Cinnamon Bay is indicated on each island.
In this analysis, we examine the degree to which archaeology provides data that permits the exploration of individual action in the face of seemingly overpowering external constraints. First, we explain the geographies of alternative possibilities that existed in the Caribbean for groups of people who chose to live ungoverned. Second, we show the practical knowledge employed by common people to establish and sustain settlements in lands beyond the territorial limits of their homeland or the homeland of their captors. Third, we suggest some general strategies for studying authority and practical knowledge in archaeological contexts.
Alternative possibilities
This article explores the possibilities and alternatives for those living in the penumbra of colonial states. Michael Dietler defines colonialism as: ‘the projects and practices of control marshaled in interactions between societies linked in asymmetrical relations of power and the processes of social and cultural transformation resulting from those practices’ (2010: 18). For Dietler, colonialism is a history written backwards. In this case the writers, administrators, and historians employ categories of analysis that were inherited by people colonizing a particular place. The archaeological record becomes a mechanism to locate such dynamics and chart their material effects in the kinds of things settlers and indigenous people made, used and consumed (2010: 13). Our difficulty with developing a label or category for the settlements described above highlights the difficulty in writing such histories backward. In adopting categories from the written record, we can often miss the fact that long-term residents had different political, economic, and technological projects in mind. For the residents of such settlements one could debate whether they even thought of themselves as colonists, rather than opportunists or something of the sort. Indeed, colonialism was not a homogeneous historical force, but an ambiguous assemblage of local traditions and trajectories. We can therefore read the archaeological record as uneven accounts of the lives of long-term residents and their local engagements with colonialism’s threats and promises.
As many historians have noted, long-term residents of the Caribbean had to confront the increasing centrality of export-oriented markets and constant changes in the regulations and rules of colonial governance (Hall, 1985; Higman, 2000, 2005; Pérotin-Dumon, 1991; Shepherd, 2002; Trouillot, 1988). We posit that some individuals could take advantage of geographies of alternative possibilities. While settlements in these geographies of possibility are from different periods of a very turbulent eighteenth century in the Caribbean, they are linked by shared features. The settlements were occupied before formal colonization in lands just beyond the territorial limits of colonial states. A diverse cluster of Europeans, Africans, and, probably, some indigenous people inhabited the settlements. Many of these settlers were either creole (born in the Americas) or long-term residents of the region. Some of these people were enslaved, many would have been poor, and most would have been without property in established colonies and not fully invested in colonial structures.
We take as our point of departure a recent insight by James Scott about peoples living in the penumbra of agrarian states in Southeast Asia. He argues that peoples described during the nineteenth and twentieth century as less complex, primitive or even backwards were none of these things. Rather, according to Scott, they were practicing the ‘art of not being governed’ and adopting strategies of settlement and subsistence to evade the apparatus of the state (Scott, 2009: 9). This insight enables us to privilege the analytical potential of frontiers, and focuses our analytical attention not only on how power works but also on how it is received. Scott’s work, including this latest incarnation, has been criticized for its focus on institutions rather than the pragmatic choices of bureaucrats (Herzfeld, 2005), the tendency to overgeneralize in the anticipation rather than the recollection of evidence (Ferguson, 2005), and its sometimes overly optimistic view of human autonomy in light of market forces (see Coronil, 2001). Yet the alternative perspective offered by his work is both appealing and consistent with the data we are examining here.
By attending to the ways in which physical geography and power intersect, ‘the art of not being governed’ acknowledges that long-term residents might have known things that colonial administrators did not. Central to Scott’s argument are the regions in Southeast Asia referred to as Zomias, to which ‘much of the population in the hills has, for more than a millennium and a half, come … to evade the manifold afflictions of state-making projects in the valleys’ (2009: 22). Certainly Scott is not the first to assert that geography and power are interlinked. Yet while Braudel (1972) and Wolf (1982) have done so before, Scott’s analysis provides realms of possibility in which common people can employ pragmatic knowledge that enables them to mediate asymmetrical relationships of power in colonial societies.
As many historians have noted, long-term residents of the Caribbean had to confront dramatic changes in the political and economic landscape of the Caribbean and some of these peoples were described as backwards or less complex by their contemporaries. However, as a mosaic of island colonies, the expression of colonial authority was an uneven topography creating geographies of alternative possibility. A brief survey of the archaeological literature would reveal a range of settlements in such geographies. Africans who escaped slavery established Maroon settlements in settings like Jamaica, Surinam and French Guiana in lands that remained uncontrolled yet ‘internal’ to the colony (Agorsah, 2002; Goucher and Agorsah, 2011; Price, 1976; Price and Price, 2001). Inhabitants formed flexible communities that could change identity for practical advantage almost at will, and absorb a huge range of disparate incomers. Long-term residents of European and African extraction established seasonal camps to extract resources from islands where there was no clear commercial value under the emerging cash-crop-centered and labor-intensive agro-industrial system (for logwood cutting in Roatan and the Miskito Coast see Finamore, 2006; McBride, 2006). Participants were at least marginally outcasts by way of their economic or ethnic position. Finally, a class of settlements existed in territories where colonial authority was ambiguous due to an entrenched indigenous population (see Hicks, 2007, for St Lucia; Lenik, 2011, for Dominica). Inhabitants of these settlements took advantage of geographies very similar to the Zomias described by Scott.
Settlements and neighboring colonial centers.
As colonies, they were noted for their difficult terrain, entrenched creole population, and difficulties to govern (Table 1). St John was located between the Danish West Indian colony of St Thomas and the British Virgin Island colony of Tortola in the northeast corner of the Leeward Islands. The mountainous island was colonized by Denmark in 1718 and was the smallest of the less economically developed of the Danish colonial holdings. The mountainous island of Dominica is located in the Eastern Caribbean between the French Departments of Martinique and Guadeloupe. As an island Dominica remained a neutral territory reserved for the Kalinago until it was ceded to Britain in 1763.
In a Danish history published in 1791, Georg Hóst reported that when Erik Smith arrived to attempt an early and ultimately aborted Danish settlement on St Thomas in 1665, he found the island inhabited by ‘Hollanders and Caribs’, the former referring to Dutch and the later a term used to define renegade indigenous peoples (reported in Knox, 1852). This was a pattern repeated in St John 20 years later. While identified by Columbus in 1493, the indigenous people living on Dominica did not suffer the total devastation that the indigenous population did in the Northern Caribbean. In fact, while France laid claim to the island in 1623 and England in 1627, through a patent granted to the Earl of Carlisle, they both signed a treaty at the cessation of the Anglo-Spanish war in 1660 that set aside the islands of Dominica and St Vincent for the ‘Carib Indians’ inhabiting the Windward Islands (Boromé, 1972). This made the island a buffer zone refuge for fleeing native peoples and others who found their way to the island for either the resources or refuge that it provided them (Honychurch, 1995, 1997).
A range of settlements exists in the penumbra of colonial authority. Physical geography, rather than the homeland, cultural dispositions, or economic dispositions, appears to have been an important element in the location of these settlements. The geographies of possibility, and the people who inhabited those geographies, fell ‘betwixt and between’ the standard definitions and outlines of history, but may in fact represent a significant trend in Caribbean interactions on the margins of formal colonial societies. Inhabitants had to take advantage of practical knowledge in order to exploit these geographies of alternative possibility.
Practical knowledge: Historical and archaeological evidence for long-term residents
In the late seventeenth century, islands like St John or Dominica were considered to be too small, dry, wet, or steep for agricultural production or military monitoring. They were not necessarily difficult to make a living in, yet they did require a degree of practical knowledge. Residents of such islands, whether indigenous or foreign, generally had some depth of knowledge of the region that they acquired with time and skill. This practical knowledge can be examined through the documentary and archaeological record.
Present-day extant architectural and archaeological features of Cinnamon Bay and Bois Cotlette represent only the last incarnation of a landscape dramatically changed over the past 500 years. Cinnamon Bay is located on the north coast of St John, USVI (Figure 1). Cinnamon Bay Plantation (initially called Caneel Estate) has been characterized as a fairly typical St Johnian sugar estate with a plantation house, sugar works and slave village located 300 meters inland. A road led down to a paved jetty on the shoreline. As a plantation settlement it contains extant architectural features consistent with late eighteenth-century sugar estates in the Caribbean. It is similar to other contemporary estates on St John including Annenberg Estate in Maho Bay and Adrian Estate on the Center ridge of the island (Figure 2). Landscape features significant to this study were a shoreline warehouse thought to have been used for the storage of sugar while awaiting transfer to ships for export, and evidence of a village located directly west of the warehouse on the shoreline.
Site maps for Bois Cotlette (inset) and Cinnamon Bay (bottom). At Cinnamon Bay the sugar works are 300 meters inland. The top map is focused on the early eighteenth-century settlement.
Bois Cotlette lies in a saddle between Morne Vert and Morne Patate in the southwest of Dominica (Figure 1). It was named after the Bois Cotlette tree (Citharexylum spinosum). A road running in a roughly north–south orientation with intermittent cobble stone patches splits the estate and currently links it to estates on the southern tip of Dominica and the village of Soufriere to the north. A sugar and coffee estate that also processed cocoa during its almost 300-year history, Bois Cotlette today includes a windmill (moulin à vent), the planter’s main house (maison de maître), coffee racks (boucan), a refurbished pulping mill and estate kitchen, and the ruins of the boiling house and enslaved laborers’ quarter (cases). Between the maison de maître, the pulping mill and the boiling house, trays of coffee were laid out on racks inside the drying shed (séchoir) and were brought out into the sun on the paved stone glacis for drying coffee during the day (Figure 2).
Historical evidence
When St John was formally colonized in 1718, the new, sanctioned settlements leapfrogged the shoreline. Patents for new estates initially omitted the valuable shoreline of Cinnamon Bay, and defined parcels to the west, east, and inland, leaving what might initially appear to be an unsettled and unchartered parcel on the shore. However, the lack of a land patent (deed) did not mean a lack of occupation. The newly settled estates and their associated social and economic activities took place further inland from the people who already lived on the shoreline estate. Lands were cleared for sugar cane fields with an array of buildings serving for agricultural production and workers’ quarters. The buildings of this new estate were constructed about 250 meters inland from the shoreline site (Figure 3).
Left: detail of the Oxholm Map (1774) indicating the new sugar works and slave village at Cinnamon Bay. Courtesy Rigsarkivet, Denmark. Right: detail of the Byres Map (1775). Plots near ‘K’ are the properties leased by Louis and Adrien Le Ferrier Constance that would become Bois Cotlette. Permission to reproduce the Byres Map is courtesy of the Dominica Museum.
During the 1680s, Danish Governor Esmit, himself a former indentured laborer with a background in Barbados, attempted to settle St John with a group of disenfranchized, formerly indentured, individuals from Barbados. A letter written by Esmit in 1684 to the Danish directors describes his attempt to utilize the financial resources of two Barbadian merchants to set up ‘works’ on St John. Esmit arranged for 40 men to establish the settlement. However, this effort at settlement was countered by William Stapleton, governor of the British Leeward Islands. When the Barbadians arrived, two British sloops were dispatched to St John to drive off the new settlers. The confrontation resulted in a flurry of correspondence that addresses the issue of settlement on St John, including a complaint by Governor Esmit to his directors saying, ‘this is the third time that he has driven our people [from St. John]’ (Westergaard, 1917: 127 – Esmit to Directors, 26 January 1684).
The group certainly had knowledge of the island and they were likely to have stayed on in the Danish colony after their ‘eviction’ from St John was ordered by the British. They are likely to have been among those who returned to populate this shoreline landholding in the years preceding formal colonization by the Danes. A number of the early Dutch settlers in the Virgin Islands became part of the core of the successful resettlement of St John. The Dutch and their constituent settlers stayed when the British preempted Esmit’s attempt at settlement in 1667, and ultimately they were incorporated into a renewed Danish occupation of the island beginning in 1672. Knox, in his 1852 book A Historical Account of St. Thomas, acknowledged the continued Dutch presence, but discounted the notion that indigenous populations were extant on the island during this period. However, as Knight (1999: 11) points out, it would have been logical for the Dutch settlements, prior to 1672, to have utilized indigenous labor to support their plantations. African descendants are documented as present in 1718, and the material record projects their presence in the initial period of settlement. Prior to 1718, ungoverned settlers on St John probably provided sea- and land-based provisions to neighboring islands, including the growing port and plantation-based island of St Thomas. After 1718, provisions from the shoreline estate were expanded upon as formal settlement brought capital investment to newly created sugar estates located inland from the shore at Cinnamon Bay and throughout the island of St John. In order to encourage rapid settlement, the Danish government established a 10-year tax exemption for new estates (Larsen, 1991: 19). This facilitated rapid settlement and kept development costs down while plantations were constructed and labor acquired, but it also meant that detailed lists of settlers and production found in later tax records would not be recorded until the end of the first 10-year period, in 1728.
The estate now known as Cinnamon Bay Plantation was formed out of three independently settled plantations that accounted for a combined total of 300 acres (SJLL 1728). The three holdings were initially known by the name of the respective owners, and the term ‘Cinnamon Bay’ does not occur in estate records until its appearance in a mortgage in the 1760s (MP 1799–69). Historical records document an early settlement, like a negative photographic image, as viable lands that were left unclaimed by the newcomers since they were settled by long-term residents. Formal application for patents to these lands would only be filed at a later date, either by the existing settlers or persons who acquired the shoreline estate from them. In any case, patents for the shoreline were filed at a later date. The first tax rolls in St John described the Cinnamon Bay property as being located ‘inde flacten af Caneel Bay’ (in the flat of Cinnamon Bay) (SJLL 1728). Adjoining properties were granted patents in 1718. It is not until 1722 that an official patent is granted to William Gandy, a master builder of English descent. The fact that this parcel was already a well-established cotton plantation by the time the first St John tax accounts were compiled in 1728 suggests that it was already in operation before William Gandy received a formal land patent in 1722 (Knight, 1999; SJLL 1728).
Between 1722 and 1724, William Gandy (SJLL 1728) obtained the patent for the flat in Cinnamon Bay and took up residence of his waterfront storehouse with his wife Elizabeth and his two stepdaughters, Anna Marie and Rachel Mauor (STBP 1724–32). For this early period, records indicate that the shoreline estate had 10 slaves, but the records for this period do not provide us with names or details concerning age, condition, or specific trades that they may have engaged in. By 1731, both the shoreline settlement and an adjacent sugar estate had become the property of Adriana Jansen, widow of Daniel Jansen (SJLL 1731). That same year, 1731, William Gandy’s stepdaughter Rachel Mauor married Daniel and Adriana Jansen’s son Johannes (STBP 1723–33). The couple took up residence on the Gandy shoreline property after Johannes’s death a decade later (SJLL 1732). Though tax rolls continued to record the shoreline property as distinct, economic focus had already shifted to sugar and the workforce for the combined sugar estate included 27 adult male and female slaves (SJLL 1739). The documentary record is partially helpful in establishing the presence of settlements that were occupied prior to formal colonization. While the exact contours of the settlement were unknown, it is certain that the property in 1718 was not terra nullis and that it was composed of free and enslaved people. This is a pattern we can see repeated nearly half a century later in Dominica.
By the mid-eighteenth century the neutral islands of Dominica, St Vincent, Tobago, and St Lucia were among the last uncolonized Caribbean territories. Through tacit agreements and official treaties these islands had been deemed neutral, not to be inhabited by Europeans, and left to the indigenous Carib or Kalinago. Designating mountainous Dominica as neutral had strategic significance for Britain and France, as the island lies between Guadeloupe and Martinique, French colonies since 1635. Preventing Dominica from becoming a colony would avoid a block of French-controlled islands, or a British colony in between two French colonies. Only during the Seven Years War did France and Britain begin occupying territories in anticipation of using these islands as bargaining chips in post-war negotiations (Pares, 1936). During Dominica’s long period of neutrality it was not an empty land. Rather it was a non-aligned island settled by Caribs, maroons, and French-speaking families. These communities lived under a constant perceived threat of violence – invasion, expulsion, or enslavement – which the neutral status was supposed to ameliorate.
Between the 1690s and the 1740s, various European squatters, coffee planters, lumbermen, and fishermen settled Dominica in the hope of finding a stake in an increasingly crowded and profitable Caribbean. By 1699, the English attempted to establish a fort just south of Roseau (in an area called Castle Comfort today) and by 1727 the Governor of Martinique appointed a commander to oversee the French subjects on the island. Under informal French settlement, cultivation was kept to a minimum for several reasons. First, the informal and tenuous nature of settlements led to an ever-present fear of evacuation that prevented much in the way of formal investment. Second, patrons in Martinique and Guadeloupe did not want competition and thus restricted investment in some enterprises such as sugar. Most of the agricultural activities took the form of cotton [261,000 pieds], cacao [952,000 pieds], plantains [253,000 pieds] and coffee [1585 pieds] (Boromé, 1972: 68).
When Britain absorbed Dominica, administrators promoted the establishment of sugar plantations. In 1763, 94,345 acres were apportioned, mapped and sold (Byres, 1776). This apportionment was documented in what has been referred to as the Byres Map, a document published by John Byres under the commission of Sir William Young (1725–1788), the President of the Commission for the Sale of Lands in the Ceded Islands. It was linked to an index of ownership that detailed who the freeholders of the land were and who the leaseholders of the land were (Byres, 1777). Potential British owners were restricted to 300 acres in an attempt to mitigate the growing planter interest. The French were not allowed to purchase property under the new English administration. Instead they leased land from the Crown, which was unevenly distributed on plantations in fertile valleys and marginal coastlines in northeast and southern-most parishes.
Dominica was a peripheral, but important, colonial holding for the British Empire in comparison to islands such as Barbados or Jamaica. While the ambitions for this island were great, it was at the end of infrastructural networks. At the same time, it was strategically necessary. The diversity of people living in Dominica cannot be underestimated. Boromé estimates that by the time England took control of Dominica there were 7892 non-Kalinago inhabitants: 5872 were enslaved Africans, 300 were free people of color, and 1718 were white (Honychurch, 1995: 75). In Dominica, and also in St Vincent and the Grenadines, the Garufuna (Kalinago, Kalipona, Kalina), an Arawakan-speaking native group mixed/allied with the black population (so-called ‘Black Caribs’), were expelled by the British around 1793 to what is Belize today. After 1793 these ‘Black Caribs’ engaged in resistance and warfare that most certainly would have had major but variable consequences in the operation of plantations.
The property known today as Bois Cotlette Estate has been inhabited since the 1750s, if not earlier. It was identified on the Byres Map as ‘3*’ and was listed on the associated property list as a leasehold. An indenture conferred by the British Crown in 1765 leased the land to Louis and Adrien le Ferrier Constance, both creoles of Martinique (DBOD B1 1765). By 1775 both individuals had died and their daughter Therese La Ferrier Constance, a minor, inherited rights to lease the land to be administered by power of attorney by Jospeh Serrant of Point Michel (DBOD 1760). In 1775, a marriage was contracted between Therese and Joseph Bellot of Morne Rouge at which point he became controller of the property and the 98 slaves tasked to work the property (DBOD M2 1775). The vast majority of enslaved laborers living at Bois Cotlette were born in Dominica (75%). The average age of creoles was 25.5 and of Africans 45 years old. The oldest individual living on the site was born in 1725 in Martinique – three years after the owner of the property. The oldest Dominican slave was born in 1752. The youngest Martiniquais slave was born in 1749 (DTSR SM 1817). This provides a fairly firm date for the founding of the estate somewhere between 1749 and 1752.
In St John, it is unlikely that Gandy was the first person to live on the land occupied by the warehouse. The negative image of lands presented by land patents and the continued use of the property seem to suggest the presence of long-term inhabitants. We see a similar pattern repeated nearly half a century later in Dominica. Existing settlers and long-term residents appear in the documentary record through an analysis of land acquisition. Ways of making a living continued, yet there is a dramatic shift in the archaeological landscape that accompanies the investment into sugar agriculture. There are some significant differences that emerge in the archaeological record.
Archaeological record
The archaeological record of Bois Cotlette and Cinnamon Bay demonstrates that both sites are multiphase and complex, complementing a documentary record which suggests extending settlement before formal colonization. Differences in structures and artifacts seem to suggest that, while poor whites settling these properties before formal colonization might have sought alternatives to the colonial strategies in places like Barbados and Jamaica, the economics of such places were still dependent on asymmetrical social relationships. Thus not everyone was there by choice. The discarded artifacts of these structures indicate that the people who lived at these two settlements relied on a diverse set of economic and social networks.
Multiphase sites exist throughout the Caribbean and probably the best documented example is St Ann’s Bay, where in a single site there are archaeological manifestations of a prehistoric settlement, Maimi Bay (for a report on the pre-colonial occupation see Cotter, 1970); a Spanish colonial (1503–1655) seat of government and sugar hacienda, Sevilla Nueva (Woodward, 2011); and a British period (1694–1860s) sugar estate, Seville (Armstrong and Kelly, 2000). Each phase is located in spatially discrete areas of the site and areas of overlap are generally disturbed. Materials from each phase were distinct, including chronology of local coarse earthenware (Hauser, 2008; Woodward, 2006). The nature of the site formation process and the continuity of material culture from one phase to the next distinguish Cinnamon Bay and Bois Cotlette from Seville. The number of subsequent occupations indicates a preference for these locations by successive settlers or occupiers even though the populations of each settlement were quite low. We speculate that the overlap in material culture has as much to do with interaction between various settlers and inhabitants as do the various taphonomic processes. In other words, we think that new residents learned from long-term inhabitants.
Comparisons of settlements with contemporary sugar estates.
That said, some general transitions were observed in the archaeological record with which we can coordinate historical dates (Figure 2). At Cinnamon Bay, we argue, the date of formal colonization was a non-event. Settlement was established before 1718 and not modified until 1733. Underlying a burn layer associated with the 1733 slave rebellion at Cinnamon Bay were a series of post-holes that indicated an earthfast house that was most likely the residence of enslaved laborers. Underneath the builders’ trench for the associated eighteenth-century warehouse was indication of an early occupation that used a stone foundation. The archaeological record indicates the deepest arbitrary levels associated with the planter’s house and the post-holes for the earthfast house contain porcelain, slipware, onion bottles, tin-glazed earthenware, and local coarse earthenware. This is material you find in the seventeenth and eighteenth centuries. Materials diagnostic to the eighteenth century (n = 3) appear only in the uppermost levels of the pre-1733 deposit of the planter’s house context. This house was burned in 1733. As such, the materials are far more similar to a seventeenth-century context than an eighteenth-century context.
At Bois Cotlette, there is a similar pattern where formal colonization punctuates existing settlement patterns. Underlying the boiling house built in the 1770s were a series of post-holes that were also most likely the remains of an earthfast house once occupied by enslaved laborers. Underneath and at a 4-degree angle to the foundations of the extent maisan de maître was a mortared course of stonework – most likely, the foundation of a building associated with the pre-British occupation. Materials recovered from both contexts include Dutch-made pipes, local and European coarse earthenware, French-made faience, European slipware, and White Salt Glazed Stoneware. This is a fairly standard mid-eighteenth-century assemblage. In all likelihood it was the residence of Louis and Adrien le Ferrier Constance. After formal colonization, the village was moved about 100 meters north and a new estate house was built. Materials recovered from these archaeological contexts include creamware, pearlware, faience, dip-molded glass bottles, and English pipes.
While the people occupying and reoccupying these settlements were living in a way that was different from formal colonies, the settlements still relied on asymmetrical social relations. In the case of both settlements, long-term residents cleared land and undertook major building projects after formal colonization. Gandy built a new warehouse at Cinnamon Bay and Jensen had enslaved laborers clear forested land and build a state-of-the-art industrial works to process sugar. At Cinnamon Bay, though poorly documented, it is likely that 10 laborers lived in the earthfast house that was burned in 1733. Gandy most likely lived in the adjacent two-storey structure that served as residence (second storey) and warehouse (ground floor). The separation of the two areas, as well as the differences in goods recovered from each house, indicates that there were differences between both sets of residents.
Similarly, Louis Ferrier Constance had laborers clear and terrace mountainous land and build a new maison de maître, a windmill and a boiling house at Bois Cotlette. These were dramatic changes to the archaeological landscape and were observed in stratigraphic profiles of test units at both settlements. Yet we can observe continuity in the material practices. Before colonization, structures and the artifacts recovered from their living surfaces indicate differences between free and enslaved residents. The enslaved lived in an earthfast house, separated by some distance from the planters, and they utilized a set of goods that would have been considered to be of lesser quality by the planter. These were differences that continued to manifest themselves in occupations dating to immediately after colonization. While colonization did not necessarily affect the material assemblages left by those who were living on each of the properties, it certainly shaped the very contours of how those properties looked. Although administrators took advantage of new colonies by introducing sugar and securing investments for existing plantations, the degree to which changes in field management and industrial infrastructure altered the material practices of free European settlers and enslaved Africans is not entirely clear. This calls into question how much Gandy and the Constances were invested in the Danish and British colonial projects. Moreover, it is probably the case that the laborers living in these settlements rejected the new colonial authority or the formality of their being enslaved.
A second difficulty in studying these sites is the use of associated material culture – buttons, coins, pipes and ceramics – to ascertain the dates of occupation. In some cases the problems associated with developing firm dates for occupation result from the imperfect parallels between different European manufacturers. The Mean Ceramic Formula (South, 1977) tended to artificially skew the dates for occupations. In the case of Bois Cotlette (mcd = 1749), dates appeared earlier than indicated on the documentary record. While attempts have been made to assign locations and date ranges to the manufacture of faience based on stylistic variations (Walthall, 1991a, 1991b; Waselkov and Walthall, 2002), more recent treatments have shown that there was much greater variation in French faience industries at Nevers and Dijon than can be attributed to specific times (Rosen, 1995, 2009). An implication of this for Cinnamon Bay and Bois Cotlette is on the end dates of manufacture. Waselkov and Walthall mark the end of manufacture of faience at 1763 – the same year as the conclusion of the Seven Years War. French ceramic manufacturers continue to produce faience well after 1763, with changes in ceramic technology and style (Kelly, 2011).
There are ways to use such diagnostic materials to coordinate dates (Figure 4). The absence of White Salt Glazed Stoneware in the earliest occupations of Cinnamon Bay, and Cream Colored Wares in the earliest occupations of Bois Cotlette, allows us to place each early settlement before formal colonization. While Salt Glazed Stoneware is produced as early as 1720, it does not gain in popularity until after 1744. This corresponds to frequencies observed at Cinnamon Bay. Similarly, while creamware is produced in the late 1750s, it is not until after the Seven Years War that it becomes a popular place setting. This also corresponds nicely with archaeological contexts observed at Bois Cotlette. In addition, ceramics identified as faience have a tin-glaze in contexts associated with the early British period.
Distribution of ceramic types before and after formal colonization.
It is also important to remember that the Caribbean was a mosaic of island colonies linked to various mercantile networks whose political boundaries were constantly shifting during the course of the eighteenth century (Pares, 1956). As such, the use of datable European-made ceramics introduces a series of considerations. For example, on many islands, French, English, Spanish or Dutch controlled, it is not uncommon to recover a mixture of English- and Dutch-made pipes. At Cinnamon Bay the diversity and small number make the Mean Ceramic Formula untenable. Dutch-made delft, French-made faience and coarse earthenware, and English cream-colored ware (1750s–) can also be recovered from islands whose imperial center is technically at war with the country who made the pottery. For example, early cream-colored wares dating to the 1760s can be found on plantation sites and colonial settlements in Guadeloupe – which was occupied by the British during the Seven Years War (Kelly, 2011).
Even though there are issues with dating, such diversity can help us understand changes in economic networks over time. If we were to examine in detail the broad classes of porcelain, tin-glazed earthenware, and coarse earthenware, complex social networks emerge. More importantly, formal colonization does not seem to shape these networks in significant ways (Figure 5). The residents of both the warehouse and the earthfast house at Cinnamon Bay used porcelain originating in Pacific Asia both before and after colonization. They also used Dutch tin-glazed earthenware, English slipware and French cooking pots (Vallaurise) both before and after formal colonization. After formal colonization inventories grew to include English White Salt Glazed stoneware and Moravian (from Winston Salem, NC) Slipware. For a relatively minor settlement this is a rich set of economic networks.
Distribution of ceramic types with known provenance before and after formal colonization.
Residents of the maison de maître and the earthfast house at Bois Cotlette employed an equally rich set of networks. Porcelain is a small component of the archaeological assemblage. Tin-glazed earthenware, however, was recovered in a variety of decorative styles. In contexts associated with the settlement before and after colonization, colors and motifs found on Dutch delft were recorded. Also recorded were palettes and decorations similar to the varieties Brittany Blue on White, Normandy Plain, Rouen Plain, Rouen Polychrome, Seine, and St Cloud Polychrome, both before and after formal colonization. French cooking, washing, and storage pots (Vallauris, Huveaune, and Biot) were also recovered in contexts dating to both before and after colonization. After colonization there was an increase in creamware and pearlware. Rather than shifting economic networks, this might more accurately reflect the varied networks of individuals.
Finally, if we look at locally produced ceramics, including wheel and hand-thrown varieties, complex networks emerge. Included in assemblages are pots identified by Kelly et al. (2008) as industrial ceramics (type 1), wheel-thrown water pots (type 2), and coco neg (type 3). These ceramics would have been purchased by free settlers and enslaved laborers. Further compositional analysis (Hauser and Kelly, 2011) indicates that the residents of Bois Cotlette got any such ceramics from one of three places both before and after colonization (Martinique, Guadeloupe, or St Lucia) (Hauser, 2011). The local ceramics were anything but local. The lack of change in both pre and post colonization indicates a durable set of networks that the enslaved and free settlers tapped into.
This evidence suggests that both settlements had a social and economic infrastructure that enabled them to obtain a variety of goods both before and after official colonization. This variety of goods is not necessarily that common on archaeological assemblages associated with sugar plantations in the last half of the eighteenth century. In many ways it is bucking a trend seen elsewhere, where there is a greater increase in the homogeneity of goods free and enslaved residents are using. Rather than such networks acting as a colonial starting point for empires such as Britain or Denmark, we argue that these residents were seeking out alternatives to the mercantile regimes of the time period. Such alternatives proved vital in times of crisis such as wars, earthquakes, and hurricanes, when mercantile networks orchestrated in London, Paris, or Copenhagen broke down. Administrators of new colonies might have also found such networks vital.
The depositional history presented in Table 2 would be consistent with the range of natural transformation processes that both sites would have been exposed to. Hurricanes and floods (Cinnamon Bay) and the deposition of pyroclastic ash from volcanic eruptions and landslides resulting from earthquakes (Bois Cotlette) have led to soil accumulation. Such depositional history is also indicative of changes in the political and economic landscape of the islands. The change in political affiliation in Dominica and St John led to an investment in sugar agriculture. The deforestation and land modifications such as terracing and water catchments that accompanied this transition led to soil destabilization and accumulation. We can summarize two observations. Changes in political sovereignty did lead to transformation of the archaeological landscape in the form of industrial buildings for sugar production and the building of new estate residences. That said, changes in material culture did not lead to a change in the suite of goods people used.
We can think of two implications for these observations. First, after formal colonization, administrators relied on long-term residents to convert existing shoreline settlements into sugar estates that would make quick profits. In St John, this amounted to deeding land already in use by existing owners. On Dominica, administrators, fearful of a French presence, leased lands to long-term French residents. One can only imagine why such a strategy was found expedient, but we would like to suggest that administrators wanted to rely on the knowledge of people who were familiar with the land. The second implication is that residents continued to do things, provision themselves, and outfit their houses after the political transition in ways they did before formal colonization. This is the case despite the changes in mercantile regimes people were expected to follow. Residents at Bois Cotlette continued to get goods that ultimately originated in Holland, France, and England after British colonization.
We believe that both implications are linked. It is probably the case that long-term residents – enslaved and free, indigenous, European or African – did not view themselves as part of a colonial project by a motherland. After all, as noted historian Jack Greene describes, England ‘was, for much of the two centuries after 1560 and more especially after the union with Scotland, a composite state characterized by indirect governance, fragmented authority, an inchoate theory of national sovereignty, and limited fiscal, administrative, and coercive resources’ (2000: 93; see also Greene, 2001). It is possible to imagine that such settlers and residents viewed the new political structures, mercantile regimes, and legal codes that accompanied formal colonization as an imposition to work around. It is therefore possible to consider such material as evidence of alternative strategies that facilitated ways of life, modes of production, and means of trade and exchange that are at odds with expected norms of colonial society.
Conclusion: Colonial authority and Caribbean settlements
In this article we have asked the question: what do we call settlements that are inhabited by European, African and indigenous peoples prior to the formal colonization of a place? We have no answer. Yet to think of such settlements as frontiers, or proto-colonials, is to tell a story from a certain perspective privileged by knowledge of the present. It is important to note that during the eighteenth century not every European considered him/herself a colonist, not every African was living in conditions of enslavement, and finally not all indigenous Americans living in the Caribbean were from the Caribbean. The people who lived in these settlements were, to a certain extent, seeking out a different realm of possibility from the kinds of conditions open to them in colonial contexts such as Martinique, Barbados, or Puerto Rico. Neither were such settlements utopias. They were complex societies with asymmetrical relationships of power in which some of the residents did not choose to be there.
That said, we would like to insist that while such settlements were different, they were not outside of a larger context of colonial interaction. Inhabitants relied on external economic networks to furnish their households and provide a market for the cotton grown at Cinnamon Bay or the provisions grown at Bois Cotlette. Similarly, these same settlements were crucial to the operation of sugar estates in Martinique, Puerto Rico, and St Thomas. They grew experimental crops like coffee, they provided staple goods such as manioc and yams, and residents fished surrounding water to supply an ever-increasing number of laborers enslaved on formal colonies. They were off the books but they were essential for balancing those same books.
By the later nineteenth century, Bois Cotlette became something akin to a corporate headquarters for the Bellot family. Similarly, the plantation at Cinnamon Bay was eventually abandoned as an agricultural enterprise. The ultimate failure of these settlements did not come from their marginal location. Rather, it came from trying to impose the prescriptions of people who had never been to the Caribbean trying to make money off newly colonized lands. This came in the form of growing sugar at both Bois Cotlette and Cinnamon Bay.
In the case of this study we have argued that there is a prehistory to the formal colonization of two sugar plantations in the Caribbean. That prehistory has informed a particular colonial relationship in the last quarter of the eighteenth century in which residents were difficult to govern from the perspective of colonial administrators. This does not mean that such settlements were not without their own relationships of power. We are fully cognizant of the irony in requiring our readers to question the centrality of colonization during a time of colonialism. In such relationships of power, practical knowledge helped circumvent the new political and economic regimes which accompanied formal colonization. A question for future research would be to explore whose knowledge this was that of the Kalinago, European settlers or the enslaved? And how much did formal colonization affect the lives of those whose land (Kalinago) or lives (Africans) were appropriated by long-term European settlers (the French) and newcomers (the British)?
Footnotes
Acknowledgements
Research on both settlements has been conducted through the assistance of Lennox Honychurch and the Commonwealth of Dominica, the United States National Park Service, Syracuse University, University of Notre Dame, and, since 2009, Northwestern University. The primary author is grateful for the material support provided by these organizations as well as the National Science Foundation and the Direction Régionale des Affaires Culturelles du Centre (DRAC), Guadeloupe. The authors are especially grateful for research collaborations with Kenneth G. Kelly and Ken Wild, which ultimately enabled baseline data for this research. We would like to thank Kenneth G. Kelly, Réginald Auger, Kalyani Menon, Francois Richard, Cynthia Robin, Matthew Johnson and the anonymous JSA reviewers for insightful comments on earlier drafts of this manuscript.