Introduction

The houses in Old Dongola in the period from the decline of the Makurian kingdom, through the transition, and finally the Funj domination exhibited a high level of uniformity (Obłuski et al., 2021; Wyżgoł, forthcomingb). Domestic architecture in Old Dongola from the fourteenth century has continued, namely in the well-established form of two-room houses (Edwards, 2018) clustered in compounds organized around a shared courtyard (Obłuski et al., 2021). This homogenization of dwellings may appear as a characteristic for subordinate people, doomed to repetition of house forms, living in a timeless state, indifferent to dynamic political changes occurring in the Middle Nile Valley and the city itself. In the transformative environment of the transition between Makurian and Funj, Christian and Muslim, the dynamics of persistence rather than change may require explanation. To attain this, domestic spaces should be attended as intensive, constituted by actions and relations, rather than extensive, represented by the architectural forms (DeLanda, 2016: 76; Jervis et al., 2021: 232). A useful method to reach this goal is an analysis of micro-residues complementing identification of macro remains of domestic actions. This paper, drawing from the toolkit of the high-resolution archaeology, uses multielemental analysis of floors in order to nuance the seemingly unchangeable Dongolese domestic spaces and to determine how the households emerged though everyday practice and living experience of the house dwellers.

The history of housing in the Middle Nile Valley is commonly perceived either in terms of a linear development of architectural forms (Anderson, 1996) or as rapid changes associated with population shifts (Godlewski, 2018). This paper proposes a closer investigation of a period, during which the persistence of domestic forms was observed. The high-resolution data comprising the results of the analysis of elemental composition of floors, provided a possibility to assess domestic space from the perspective of everyday experience of the house dwellers. It is required for the analysis of changes in housing, as the process of building is inseparable from dwelling (Ingold, 2000: 172–188).

This paper discusses heterogeneous practices of dwelling as well as the persistence and changeability of the household assemblages from the Makurian to the Funj period (fourteenth–seventeenth centuries). To accomplish this objective, the multielemental analysis of floors was used to investigate domestic space of Dongolese houses and the emergence of households through everyday practice.

Historical Settings

Old Dongola (Fig. 1) is located in the territory of modern Sudan, between the Third and the Fourth Nile Cataracts. From the sixth until the fourteenth century, it was the capital of the Christian Makurian Kingdom. The royal court abandoned Old Dongola in 1365 after prolonged conflicts with the Banu Ja’d endangered the city (Vantini, 1975: 699). After the subsequent disintegration of the Makurian state in the fourteenth to fifteenth centuries, Old Dongola became the capital of a small polity called the Kingdom of Dongola, and from the sixteenth century, the city was included into the sphere of influence of the Funj Sultanate. During the fourteenth and the fifteenth century, the inhabitants of the city gradually abandoned Christianity and converted to Islam (Obłuski, 2021: 5). Old Dongola retained its important political and economic role in the Middle Nile Valley until the nineteenth century when the Funj Sultanate was conquered by Egyptian forces of Mohammed Ali. In the Funj period (from the sixteenth to the nineteenth century), the city consisted of numerous clusters of house compounds separated by an irregular network of streets located within the city walls and outside. The walled city center in the Funj period was almost unrecognizable from the rest of the settlement and inhabited by both commoners and aristocracy (Obłuski et al., 2021; Wyżgoł, forthcominga).

Fig. 1
figure 1

A map of the Middle Nile Valley, orthophoto of the so-called Citadel at the archaeological site at Old Dongola with the marked location of the studied house compounds (draw. D. Zielińska, M. Wyżgoł; phot. A. Chlebowski)

Full size image

Changeability and Persistence of Household Assemblages

A household is commonly identified as a social unit defined by the activities and behavior of co-resident groups extending beyond the space of the house (Beaudry, 2015; Bolender & Johnson, 2018: 66), often entangled with other social groups from outside the house, e.g., in the forms of agrarian or craft production (see Jervis, 2022). The houses, as well as people inhabiting them and all the objects and relations holding them together, make up heterogeneous assemblages (DeLanda, 2016: 20) that mutually constitute one another (Hutson, 2010: 112). From the perspective of assemblage thought, the household is a sedimentation of objects, the house and the inhabitants, or it is formed of coagulated flows of objects, people and their mutual relations in space (Jervis, 2019: 92). The components involved in the emergence of the household assemblages, such as food preparation stations, middens, and storage are assemblages on their own, being always on the move. Therefore, the houses are changeable, rather than stable entities.

Building a house and assembling a household are, in fact, repetitions of practices rather than replications of a preconceived form. As a result, household assemblages are prone to change, as imperfection in repetition is unavoidable. The changes occur on multiple scales, thus not always resulting in an alteration of the form of a house, which may only occur when changes reach the critical point causing a phase transition (Jervis, 2019: 65–66; Crellin, 2017: 119). Yet, the vibrancy of household assemblages does not mean that everything is always flowing, but only that the changes occur on various levels, not all manifested architecturally. The complex discussion of change requires, therefore, taking into account not only the formal changes of houses, but also heterogeneous use of uniform domestic space (see Crellin, 2017).

Even though, the household assemblages are in flux, the persistence might be afforded in various ways as both continuity and change emerge from relations (Jervis et al., 2021: 232). The persistence of the household assemblages may take place thorough the process of stratification allowing past relations to circulate in other performances, which is often referred to as the tradition. The persistence can also be afforded by temporalities of materials of some of the components of assemblages. Especially, the slow temporalities of architectural elements encode interactions and allow past relations to resurface (Jervis, 2019: 64–65). Additionally, some domestic space striations (see Deleuze & Guattari, 1988 in Jervis, 2019: 54–55), such as furnished interiors of residential units, make the relations within a household more predictable and therefore less prone to change.

Model House Compound of the Late Makurian/Funj Period Old Dongola

A model dwelling from late Makurian to Funj period Old Dongola was constituted by an enclosed area comprising usually more than one building identified as a separate residential unit and a courtyard. Each residential unit was equipped with a large bench behind a stub wall separating the interior from the entrance and a narrow oblong bench located by the wall opposite to the entrance, used as a base for storage bins (Obłuski et al., 2021: 240). The residential room gave access to a narrow storage space located in the back. The narrow rooms functioned as storage spaces containing storage vessels, as well as items such as basketry and tools (de Lellis et al., 2021: 86–95; Wyżgoł, 2021a: 41–50, 2021b: 187–193). Similar rooms, known as gatee, were still attested in Sudanese architecture in the twentieth century, used as storerooms for staples and house equipment. In addition, they provided suitable conditions for the production of cured meat called sharmout (Dirar, 1993: 389). Additionally, some buildings had vestibules leading to the main room (Obłuski et al., 2021: 242). In one variant, a compound included a separate building equipped with a quern emplacement and a hearth and a clay bin (Wyżgoł, 2021a: 45, 2021c: 193). In another, equipment used for food preparation was located directly in the courtyard (de Lellis et al., 2021: 95; Deptuła, 2021: 208; Maślak & Deptuła, 2021: 136).

Domestic labor, which was identified archaeologically, mainly in relation to food preparation, was located in highly visible spaces with unrestricted access. According to the available written accounts from the Funj period, this work was performed primarily by women (Bruce, 1805: 514; Burckhardt, 1822: 45–46; Hoskins, 1835: 193). The gender association of domestic work, especially cooking, is also clear from more recent ethnographic data (Barclay, 1964: 13; Dirar, 1993: 186; Kurcz, 2007: 92; Osman, 2004: 95). Food preparation areas were consequently shared by women from all houses of a compound enabling integration of household’s female dwellers. At the same time, the spaces of food preparation facilitated transfer of knowledge and negotiation of power, enforcing a leading role of the oldest and most experienced women (Wyżgoł & Deptuła, 2023: 13–14). Moreover, female engagement in food preparation was highly visible to other house dwellers and guests. The only existing spatial dichotomy in the model dwelling separated space for “dirty,” related to the food preparation, keeping animals and storage, and “clean” for leisure activities (Wyżgoł & Deptuła, 2023: 18–19).

During the excavations at Old Dongola conducted under the auspices of the Polish Centre of Mediterranean Archaeology, University of Warsaw, within the framework of the ERC Starting Grant UMMA directed by Artur Obłuski, 24 house compounds comprising 51 houses were uncovered. The earliest dwellings representing the model were attested from the first half of the fourteenth century (Wyżgoł, forthcomingb). This form dominated domestic architecture until the end of occupation of most of the so-called Citadel in the eighteenth century (Obłuski et al., 2021). The same form of houses was widespread in Ottoman Nubia (Osman et al., 2012; Vila, 1977, 1978). Such houses were tentatively dated to the postmedieval period in reference to the pottery types (Edwards, 2018). Although, based on the evidence from Old Dongola, association of its occurrence with the fall of Makuria should be reconsidered.

High-Resolution Archaeology of Domestic Floors

Artifacts found in Dongolese houses, such as pottery (Danys, 2022), tools (Wyżgoł & Deptuła, 2020), organic containers (Cervi, 2022a; Warowna, 2022), and objects related to leisure activities (Cervi, 2022b; Danys & Wyżgoł, 2022), are indicative of various human activities including cooking, grain griding, spinning or tobacco smoking. Faunal remains attest consumption and processing of predominantly domestic animals, i.e., sheep/goats and cattle (Osypińska, forthcoming), while botanical remains, e.g., sorghum, purslane, barley, wheat, cucumber, attest processing of grains and cooking (Nasreldein, forthcoming; Nasreldein et al., 2024). Application of the single-context system allowed for attribution of artifacts and ecofacts to particular depositions and their interpretation as, e.g., middens or occupational layers. It also enabled its association of depositions with certain human activities. However, several factors including large surface area of specific depositions, their complex nature, e.g., combination of middens and occupational layers within a single context, various cleaning patterns, preclude identification of precise spatial positioning of domestic activities. Furthermore, findspots of tools were related with discard outside the workspace, rather than usage in situ (Wyżgoł, M., Nasreldein, M. and Ryś-Jarmużek, A. (in preparation). The high-resolution archaeology of shared courtyards at Old Dongola (14th–sixteenth century, Sudan): an intensive approach to domestic open spaces.). Finally, the collection/sampling strategy for botanical and animal remains (Dzierzbicka, 2021a), collected from depositions covering often large areas, also led to identification of activities with certain spaces, such as kitchen or courtyards, but not with the exact location within such spaces.

Considering the aforementioned factors concerning artifacts and ecofacts, the multielemental analysis was applied in this study of domestic floors to identify small-scale activity areas. The analysis of multiple elements, that includes elements associated with anthropogenic residues (Fleisher & Sulas, 2015; Middleton et al., 2010; Oonk et al., 2009; Wilson et al., 2008), provides an opportunity to compare datasets from different elements and is sensitive to subtle changes in chemical composition allowing for distinctions between a wide range of human activities with high precision (Middleton et al., 2010; Wilson et al., 2008, 2009). The samples (Table 1) were collected from well-established contemporary surfaces of floors and occupational deposits that had been cleaned to remove any remains of later deposits. The samples from each floor were collected only from the most recent thin layer discernible with the naked eye under field conditions. Multielemental analysis of the floors of the compounds U207/210/211/214/225, U173/174/178, and U32/38/39/87/88 was performed by Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-AES) to determine the concentration of 35 elements (Ag, Al, As, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Fe, K, La, Mg, Mn, Mo, Na, Nb, Ni, P, Pb, S, Sb, Sc, Sn, Sr, Th, Ti, U, V, W, Y, Zn, Zr). Analysis of floors in the compound U242/243/244/246/259 was performed by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) to determine the concentration of 45 elements (Ag, Al, As, Ba, Be, Bi, Ca, Cd, Ce, Co, Cr, Cu, Fe, Hf, In, K, La, Li, Mg, Mn, Mo, Na, Nb, Ni, P, Pb, Rb, Re, S, Sb, Sc, Se, Sn, Sr, Ta, Te, Th, Ti, Tl, U, V, W, Y, Zn, Zr). The analyses used a strong multi-acid extraction. This entailed heating samples in HNO3, HClO4, and HF to fuming and taking them to dryness; the residue was then dissolved in HCl before analysis. Prior to both analyses, the samples were ground in a ceramic mortar and sieved (mesh size 0.1 mm). All samples were analyzed at the laboratory of Bureau Veritas Minerals (Canada).

Table 1 Number of samples collected from each house compound
Full size table

To reliably compare both datasets, results of the ICP-MS analysis with values lower than the level of detection in the ICP-AES analysis were discarded.

The floors were composed of material reclaimed from the rubble of older constructions (Obłuski et al., 2021). Therefore, it was an assemblage of components of past assemblages (see DeLanda, 2016: 37; Jervis, 2019: 48). The deterritorialised components of the past assemblages were reterritorialised forming analyzed floors (see Jervis, 2019:48–54), so they inevitably incorporated molecules associated with past domestic actions. This characteristic of the floor assemblages means that the identification of domestic actions following the construction of analyzed floors is challenging. The legibility of results of analysis of floors in rooms whitewashed with white desert clay (de Lellis, 2021), which had never been a component of any domestic assemblage, and floors made of reused Nile silt (de Lellis forthcominga; Wyżgoł, forthcomingb) was different.

Floor sample analysis results were compared to the background level constituted by the mean value of samples collected from sun-dried bricks, to identify elevated or depleted values of elements. Nevertheless, high values of elements do not necessarily attest anthropogenic input from the analyzed period. In order to reinforce the identification and to mitigate the effect of the aforementioned past anthropogenic input, Moran’s I and local Moran’s I were calculated to identify non-random distribution of high and low values of elements. In the case of Moran’s I p > 0.05, implying random distribution of high and low values; local Moran’s I was not calculated.

Additionally, principal component analysis (PCA) and hierarchical cluster analysis were applied to bring out clustering of samples as well as to identify clusters of elements and their possible association with mineral background or anthropogenic input. All calculations were performed and plotted in R.

Hierarchical dendrograms show (Fig. 2) associations between values of elements. Specific groups can be related to the anthropogenic input (e.g., Mg, P, K, S, Na), while others with the mineral background (e.g., Co, Nb, Sc, Cu, La, Y) (Wyżgoł & Woronko, forthcoming). High values of groups of associated elements were elevated in samples from ethnographic or archaeological activity areas (Table 2), while some individual elements are directly related to elemental composition of residues of domestic activities (Table 3). Nevertheless, since most of them are associated with processing of plant and/or animal tissues, the identification of actions is often ambiguous.

Fig. 2
figure 2

Hierarchical clustering dendrogram of elements' values concentration on all floors

Full size image
Table 2 Observed elevated and depleted values of elements in samples collected from archaeological (Old Dongola) and ethnographical (El Ghaddar) activity areas
Full size table
Table 3 Elements associated with elemental composition of residues
Full size table

House Compounds from the Fourteenth to the Seventeenth Century

The household assemblages in Old Dongola were in flux, constantly changing in time (Deptuła, forthcoming; Wyżgoł, forthcomingb). The number of existing and inhabited houses within the compound often changed in the period between the construction of the first building and final abandonment by its human inhabitants. The multielemental analysis of four house compounds (Fig. 1) captured only snapshots of the flow of matter and relations within the domestic assemblages. Each of them represented a moment in a history of a household, corresponding with an arbitrarily designated archaeological phase constituted by observed major architectural changes in the layout of a house compound, which were positioned in the time period from the fourteenth to the seventeenth century (Dzierzbicka, 2021b; Wyżgoł, forthcomingb). The analyzed house compounds were selected based on the preliminary dating, each representing a century from the fourteenth to the seventeenth. The chronology was subsequently verified with radiocarbon dates. Each of the analyzed house compounds comprised all elements of a model house compound and was, therefore, considered usual.

House compound U242/243/244/246/259

This was the oldest analyzed house compound, dated to the fourteenth century. The radiocarbon date of the sample comprising plant remains collected from the depositions from the analyzed phase, subsequently adjusted to a sequence of dates, brought out a range from 1366 to 1392AD with 95.4% probability (Dzierzbicka, forthcoming). The construction of building U246/259 and the walls of courtyard U244 created a space of intensive interactions between the dwellers of the compound. Another building, comprising rooms with separate entrances—U242 and U243, was built in the corner opposite to U246/259 (Fig. 3). No further architectural changes were introduced until the abandonment of the compound by its human dwellers (Wyżgoł, forthcomingb).

Fig. 3
figure 3

A plan of house compound U242/243/244/246/259 with marked areas of elemental enrichment/depletion based on local Moran’s I (draw. A. Wujec)

Full size image

The coded space of building U246/259, like many others representing this form, had a residential affordance and was connected with a narrow storage. The storage contained fragments of basketry as well as a stone grinder, a pounder, and a wooden mallet. Whether building U242/243, located in the southern corner, was residential or accommodated other activities remains disputable. Similar constructions dated to the fourteenth century were found in an earlier house compound located within the same location (Wyżgoł, forthcomingb) and in the monastery located 2 km north from the walls in the courtyard. These constructions were interpreted either as dwellings, or as part of the workspace serving dwellers of more firm buildings located nearby (Dzierzbicka & Deptuła, 2018). In the case of the partially excavated U243, numerous firedogs found on the floor may point to a workspace (Wyżgoł, forthcomingb). The mean of values of the elements contained in the floor of U242 (Fig. 3) compared with the background level showed significant enrichment in P, 120% above the background, and K, Cu, Zn, Ni, Co, Th, Ca, Mg, all more than 25% above the background. Such high values of these general indicators of human and domestic animal activities (Tables 2 and 3) may suggest that room U242 was the household’s workspace, rather than a dwelling or it joined these two functions. Nevertheless, PC1 vs PC2 of the PCA plot shows that the difference between clusters of samples from U244 and U242 was based mostly on values of elements such as Al, V, La, Mn, Cu, not on P or K. It indicates that the difference lies in the building material of the floor in U242 (Fig. 4). Furthermore, none of the areas of elemental enrichments can be identified within U242 using Moran’s I autocorrelation. Random distribution of values additionally points to the possibility that admixtures to the building material, such as animal dung, were responsible for the enrichments.

Fig. 4
figure 4

Principal component analysis plots PCA 1 vs PCA 2 showing contribution of variables (element’s values) and clustering regarding the possible function of floors in U242/243/244/246/259

Full size image

The composition of the occupational deposition of the courtyard, including large amounts of animal dung pellets, implies the presence of livestock in the area. Furthermore, three clay containers were placed near building U242/243 attesting usage of the southern part of the courtyard for storage. Additionally, isolated finds of stone grinders and spindle-whorls either attest to grain grinding and spinning or disposal of refuse. On the level of micro-residues, three areas distinctive for their elemental composition were identified in the southern part of the courtyard (Figs. 3 and 5). In the south-western part, occurrence of ashes and processing of plant tissues could be suggested by some of the elements, although significant depletion of P and K excludes the latter. At the same time, enrichments in metallic elements such as Fe or Ti may suggest association with craft, e.g., pigment processing, although these enrichments of metallic elements, as well as Al and Mn, can equally derive from the erosion of the adjacent large sun-dried brick wall. The area located close to one of the clay containers in in the south-eastern part of the courtyard was characterized by the elemental enrichments fitting to patterns related to food preparation, i.e., processing of plant and animal tissues, as well as to the presence of livestock. Depletions of metallic elements in the southern part of the courtyard are difficult to interpret as they do not match any pattern related to human activities. A possible explanation is the difference in the background material (Wyżgoł & Woronko, forthcoming).

Fig. 5
figure 5

A map displaying values of elements (P, Ca, Sr, and Zn) compared to the background value (mean value of elements of samples of mud brick) and local Moran’s I calculated for selected elements in the house compound U242/243/244/246/259 (draw. A. Wujec). Prepared in QGIS using Spatial Analysis Toolbox

Full size image

House compound U207/210/211/214/225

The house compound emerged (calibrated radiocarbon date—1456AD-1506AD with 95.4% probability (Dzierzbicka, forthcoming)) when building U210/211 and a building abutting it from the north, as well as a wall of courtyard U214 were built onto already existing walls of older structures. In the analyzed archaeological phase, another two-room building, U207/225, was erected opposite to U210/211. It can be estimated to date to 1468AD-1522AD with 95.4% probability based on the radiocarbon dating of a short-lived plant sample (Dzierzbicka, forthcomingb). The initially smooth space of the courtyard became more striated as two walls running west–east created a north–south division (Fig. 6). After this moment of the life of the house compound, it was inhabited for another at least 150 years. During this period, the components of the household assemblage and their relations often changed as the houses have been alternately built, abandoned, and reinhabited (Wyżgoł, forthcomingb).

Fig. 6
figure 6

A plan of house compound U207/210/211/214/225 with marked areas of elemental enrichment/depletion based on local Moran’s I (draw. A. Wujec)

Full size image

The coded space of buildings U210/211 and U207/225 constituted residential units connected with narrow storages (Wyżgoł, forthcomingb). Distribution of values of elements within room U210 and U211 did not show any statistically significant areas of high values of elements according to local Moran’s I. Comparison of mean values in the floors of house U210/211 and the background level showed elevated values of Pb, Ca, P, and K in U210, and Cu, Zn, Ni, Fe, Ca, P, Mg, Ti, and K in U211 (Fig. 8). Apart from activities related to processing of plants or animals, it may indicate that these enrichments may relate to organic admixtures in the plastering of floors (Tables 2 and 3). Elevated values of Zn, Ni, Fe, Mg, Ti in storage U211, which did not occur in U210, may also attest association with plant and/or animal products (Tables 2 and 3). Alternatively, elemental input may have derived from the presence of rodents feeding on the remains of staples. The evidence of staples stored in U211 include fragments of containers made of unbaked clay, leather, and gourd. PCA shows separate clustering of samples from house U210/211 and from courtyard U214. The most responsible for the difference were elements such as Al, Cr, Mn, Nb, La, Cu. It additionally points to the mineral background of the floors in U210/211 having the main share of the enrichments. At the same time, elements such as plant nutrients K, S, P, Zn, Mg, Ca were more abundant in the courtyard (Fig. 7).

Fig. 7
figure 7

Principal component analysis plot of PCA 1 vs PCA 2 showing contribution of variables (element’s values) and clustering regarding the possible function of floors in U207/210/211/214/225

Full size image

In the courtyard, isolated finds of spindle-whorls, fragments of basketry, wooden bowls, firedogs, but also smoking pipes, a gaming piece and clay figurines either attests to domestic work and/or leisure activities. Human and animal activities were also traced through chemical composition and occurrence of human related components in two distinctive areas. The area located by the western wall of the southern part of the courtyard (Figs. 6 and 8), in the vicinity of two pits, can be associated with disposal of rubbish including ashes, but also with the presence of animals (Tables 2 and 3). Out of elements related to animal metabolism in this area P, K, and Zn are notable (Table 3, Figs. 6 and 8). Depleted values of numerous metallic elements match the locally observed pattern that ashes are poor in most of the analyzed elements (Tables 2 and 3). The second area was identified in the south-eastern part of the courtyard (Figs. 6 and 8). The composition of elements points to activities related to food processing and the presence of livestock (Tables 2 and 3). The interpretation is corroborated by the vicinity of two hearths, and a presumed quern emplacement. This area also bore enrichments in elements associated with plant and animal tissues, especially bones (Tables 2 and 3). The elemental composition indicates further that the central part of the courtyard was deprived of domestic tasks and used mostly for communication purposes (Fig. 8). Another presumed activity area related to cooking was located in the north-eastern part, outside the area of elemental analysis, in the vicinity of a hearth (Fig. 6). Stone tools including querns, grinders, and pounders clustered just outside the area of enrichments related to food preparation in the south-east and in the vicinity of the hearth in the north-east, as well as in the area of enrichment and depletion related to middens located in the west (Wyżgoł, M., Nasreldein, M. and Ryś-Jarmużek, A. (in preparation). The high-resolution archaeology of shared courtyards at Old Dongola (14th–sixteenth century, Sudan): an intensive approach to domestic open spaces.).

Fig. 8
figure 8

A map displaying values of elements (P, Ca, Sr, and Zn) compared to the background value (mean value of elements of samples of mud brick) and local Moran’s I calculated for selected elements in house compound U207/210/211/214/225 (draw. A. Wujec). Prepared in QGIS using Spatial Analysis Toolbox

Full size image

House compound U173/174/178

This house compound was erected in the sixteenth century. The radiocarbon dating of the sample of short-lived plant remains collected from the floor in U178 and later adjusted to a sequence, points to 1504AD to 1591AD with 93.4% probability (Dzierzbicka, forthcomingb). The construction of building U173/178, an unexcavated building located on the other side of the courtyard, and the walls of courtyard U174 created the domestic space (Fig. 9) adapting already existing walls of earlier buildings. This snapshot captured the first architectural intervention in the life of this house compound. In the upcoming one hundred years, it was followed by several renovations including elevation of floor level in U173/178 (de Lellis forthcominga).

Fig. 9
figure 9

A plan of house compound U173/174 /178 with marked areas of elemental enrichment/depletion based on local Moran’s I (draw. A. Wujec)

Full size image

The northern building comprised a front residential unit U178, and a narrow room U173 located behind. U178 was equipped with an oblong bench located in the opposite side to the entrance and a large bench behind a stub wall, which separated the entrance from the interior, as well as a smaller bench abutting it. Among the remains of the objects stored in U173 were bowls, a storage pot of big dimensions and a wooden peg that could have been used as a fire-starter (de Lellis forthcominga). The outline of the walls of the unexcavated building in the south suggests similar spatial arrangement to U178. In room U178, local Moran’s I showed an area located by the entrance which bore enrichments in S, K, Sb, and Th (Fig. 9). However, it does not fit with enrichment patterns of any identified domestic activities (Tables 2 and 3). Therefore, it probably attests to inconsistency in the building material of floors. At the same time, the mean values of certain elements were significantly higher than the background; P (+ 80%) and K (+ 40%) in U178 and P (+ 95%), Sb (+ 80%), and Th (+ 60%) in U173 (Figs. 10 and 11). Only the high values of P and K may be ambiguous with processing of various plant and animal products (Tables 2 and 3). PCA shows separate clusters of samples from house U173/178 and from courtyard U174, differing in the values of most of the analyzed elements (except Pb, K, Th, S, and P), having higher values inside the house than in the courtyard. The enrichments inside the house may have been, however, related to a layer of animal dung accumulated immediately on top of the floor after a shift in use of U173/178 (de Lellis, forthcominga), or to the organic admixture to the building material (Fig. 10).

Fig. 10
figure 10

Principal component analysis plots PCA 1 vs PCA 2 showing contribution of variables (element’s values) and clustering regarding the possible function of floors in U173/174/178

Full size image
Fig. 11
figure 11

A map displaying values of elements (P, Ca, Sr, and Zn) compared to the background value (mean value of elements of samples of mud brick) in the house compound U173/174/178 (draw. A. Wujec). Prepared in QGIS

Full size image

Courtyard 174 was deprived of architectural elements, apart from a bench built next to the entrance to U178. This smooth space stood in opposition to the striated space of building U173/178 and the unexcavated building in the southern part. Though, out of analyzed elemental content, only S showed a significant concentration in the space west of building U173/178 (Fig. 9). According to calculated values of Moran’s I, other elements were distributed randomly, suggesting a lack of hotspots focusing repetitive domestic tasks.

House compound U32/38/39/87/88

This house compound came into being when a small courtyard U88 was separated from buildings located west of it, with building U32/38/39 being built east of it (Fig. 12). The analyzed archaeological phase constituted the last architectural interventions in this area, followed by the abandonment of the compound by its human dwellers (de Lellis, 2021: 114). This phase has no radiocarbon date, but can be dated in accordance to contemporary house compounds accessed from the same street in the period between the seventeenth and the eighteenth century (Dzierzbicka, 2021b).

Fig. 12
figure 12

A plan of house compound U32/38/39/87/88 with marked areas of elemental enrichment/depletion based on local Moran’s I (draw. A. Wujec)

Full size image

Vestibule U38 connected room U32 with a storage U39 and possibly an unexcavated room south-east of U38. U32 had a widespread form of a residential unit and was equipped with benches and a stub wall (de Lellis, 2021: 111). The outlines of walls of the unexcavated room on the opposite side of the vestibule suggested its similar spatial arrangement. The pottery and artefactual assemblage were associated either with the latest activities of the house dwellers or with immediate post-abandonment. It comprised storage and serving vessels, but lacked those used for cooking. Other objects including smoking pipes, headrest, and gaming pieces attest to leisure activities without indicating their exact location (de Lellis, forthcomingb). The flow of matter through the space of the compound can be traced mostly on the elemental level. PCA shows clustering of the samples from particular spaces. Elements such as Al, La, Th, Nb, associated with the whitewash of a desert clay (Wyżgoł & Woronko, forthcoming), were the most responsible for the difference between occupational depositions in U39 and U87/88 and the whitewashed floors of U32 and U38. At the same time, differences between samples in U32 and U38 derive mostly from plant nutrients (Fig. 13). The local Moran’s I showed clusters of high values of elements, mostly plant nutrients (e.g., K, Cu, Mn, Sr, Zn, Fe) associated with food processing, in the central part of the floor in U38 (Figs. 12 and 14, Tables 2 and 3). P was the most abundant in storage room U39, where a cluster of high values of other plant nutrients was recorded in the central-front part of this room while depleted values were found in the back (Figs. 12 and 14). The fact that the values of this P were the highest in U39 was most likely associated with the storage of staples in this space, e.g., it might have originated from bone-derived apatite (Oonk et al., 2009:40) (Tables 2 and 3), although the presence of small rodents and/or bats in a dark room filled with staples should also be considered a possible source of enrichments in elements related to the products of metabolism (Tables 2 and 3).

Fig. 13
figure 13

Principal component analysis plots PCA 1 vs PCA 2 showing contribution of variables (element values) and clustering regarding the possible function of floors in U32/38/39/87/88

Full size image
Fig. 14
figure 14

A map displaying values of elements (P, Ca, Sr, and Zn) compared to the background value (mean value of elements of samples of mud brick) and local Moran’s I calculated for selected elements in the house compound U32/38/39/87/88 (draw. A. Wujec). Prepared in QGIS using Spatial Analysis Toolbox

Full size image

No statistically significant concentrations of high or low values of elements were recorded in the small courtyard U88 (Figs. 12 and 14). Objects, such as tools or pottery fragments, were also not found (de Lellis, forthcomingb). This indicates a lack of hotspots focusing repetitive domestic tasks, suggesting that this courtyard may have had a purely communication purpose. Therefore, the most creative potential for relations between the components of this household assemblage had the shared vestibule U38.

Discussion

Relations Within the Households’ Smooth and Striated Spaces

The discussed snapshots of the lives of the analyzed households captured human and non-human actions identified by their residues from the macro-scale, such as walls, benches, storage vessels, to the micro-scale, represented by elemental traces of domestic activities. It enabled mapping of relations within the households and confronting it with model relations (see above) based on restricted areas of residential units (see Wyżgoł & Deptuła, 2023: 18) and shared work of related women in a single domestic workspace, either in communal courtyards and/or in kitchens (see Wyżgoł & Deptuła, 2023: 13–17). As observed, the flow of things, people, animals, and their relations was affected by heterogeneous domestic space as the straited space of residential units and storage stood in contrast to the smooth spaces of courtyards.

The residential units did not accommodate any actions with detectable residuality apart from storage attested in macro scale by remains of clay bins. The space was therefore limited for sleeping and leisure activities with a low level of residuality. The other possible explanation of lack of micro-residues in the residential units were different cleaning patterns in “dirty” and “clean” spaces (see Wyżgoł & Deptuła, 2023), however, e.g., floor of a “clean” room U38 bore elemental traces of food preparation unlike U32. The striated space of residential units such as U32, U210, U178, and U246 coded relations and actions of their current dwellers. The slow materiality of walls allowed past relations to resurface and trapped dwellers in constant care, which was expressed by numerous refurbishments (Wyżgoł & Deptuła, 2023: 17). It resulted in perpetuation of a house’s form—a residential unit with narrow storage. The residential units constituted an assemblage with limited creative potential for relations; nevertheless, several examples (Deptuła, forthcoming; Wyżgoł, forthcomingb; Wyżgoł & Deptuła, 2023) attest that assemblages often deterritorialised at the moment when they became uninhabited, reterritorialising as domestic workspaces.

Smooth spaces of courtyards, or rooms other than residential units, allowed everyday domestic actions to take place freely as dwellers engaged in relations with such spaces in many ways. The open spaces focused most of the household activities bonding humans, animals, plants, portable objects, such as stone tools, spindle whorls, pots and other containers, pipes, gaming pieces, etc. In compound U242/243/244/246/259 domestic work amassed in the southern part of the courtyard, it was the space where human dwellers met and worked together. This depends on the use of U242, either as workspace or dwelling, the open space separated domestic work carried out in the south from dwelling in the north, or positioning inhabitants of U242 closer to food preparation and possibly in a form of dependency from dwellers of U246/259. In the case of compound U173/174/178, which comprised two residential units and a courtyard, the domestic activities were not identified by their residues. It may contradict the model, in which relations were established while performing domestic work within a communal courtyard. Alternatively, it did not indicate that the courtyard lacked functions, but rather the activities often changed swiftly their location or their residuality or frequency was low precluding accumulation of elemental traces. A single communal workspace was not observed in compound U207/210/211/214/225; instead, the macro- and micro-residues of human actions showed that the space of the courtyard facilitated spatial divisions into two food preparation areas, each serving dwellers of a single house. Therefore, in the captured moment, this household was an assemblage of two assemblages of house dwellers, U207/225 and U210/211, working separately. Finally, the case of compound U32/38/39/87/88 may suggest that not only the open courtyards had a creative potential for establishing and negotiating relations. The use of vestibule U38 instead of the courtyard for food preparation can be perceived as individual innovations. In this way, the vestibule became the space communal workspace used by the dwellers of two residential units. At the same time, courtyard U87/88 did not exhibit any trace of domestic activities, similarly to U174. It appears, that the courtyards often treated as entities with established functions and relations were, in fact, in each case constituted by other vibrant assemblages, such as food preparation stations, middens or animal pens, facilitating relations, thus responsible for creative potential of the courtyards.

The observed fluidity of space proves, that Dongolese households did not have stable form or structure but were changeable. The internal relations were created in the smooth spaces alongside the striated spaces of the residential units. The open spaces with unrestricted affordance had a potential of becoming a space for domestic work, storage, conduit for people and domestic animals. Therefore, the space and relations between dwellers of each house sharing an open courtyard were negotiable, especially in this space. At the same time, internal relations between dwellers of a single house/residential unit were coded by the persistent house form.

Persistence, Change, and Heterogeneity of Households

In the extensive terms, dwellings in Old Dongola exhibit little variation between the terminal phase of the Makurian kingdom in the fourteenth century and the final occupational phases in the city in the Funj period in the eighteenth century. From the simple evolutionary perspective only disappearance of characteristic irregular, constructions (U242/243) abutting one another after the fourteenth century might be considered a change in the organization of households. These buildings presumably accommodated domestic workspace and residential functions (Dzierzbicka & Deptuła, 2018; see compound U242/243/244/246/259). It was the only variation in spatial arrangement observable diachronically. This change in the “building tradition” might have been the last change in the life experience the commoners inhabiting Old Dongola. Effectively, in extensive terms, after the fourteenth century, all the house compounds have followed a model dwelling, with several buildings with residential and storage function surrounding a courtyard. Only this formal change can be considered a phase transition (Jervis, 2019: 65–66; Crellin, 2017: 119), while the more subtle changes observable in the analyzed house compounds did not materialized in an architectural form.

The relational perspective and the application of high-resolution methods attest to the heterogeneous life experiences of house dwellers of various house compounds even though spatially they have all aligned to the model dwelling. The experiences of dwellers of a single residential unit seem to be shared to a certain extent by all discussed households. They were affected by the striated space of the residential units limiting creation of new relations and movement of people and things. The difference, although not diachronic, occurred in the smooth spaces of the courtyards which allowed unrestricted movement of people and creation of relations. In compound U242/243/244/246/249, the courtyard directed domestic production apart from residential unit U246 or even promoted asymmetrical relations between dwellers of U246/259 and U242/243. In the case of two compounds, U207/210/211/214/225 and U32/38/39/87/88, both comprising two residential units, the space outside in the courtyards allowed very different relations between the inhabitants. While in U207/210/211/214/225, two food preparation stations were set up in the courtyard, creating division between each residential unit’s inhabitants, while in U32/38/39/87/88, the vestibule connecting two residential units served as a shared food preparation area, tightening relations between dwellers of each unit. The frequent changes in spatial arrangement of the house compounds (see Deptuła forthcoming; Wyżgoł, forthcomingb) attest that the vibrant open spaces were productive for short-term relations and change, binding people, and things sometimes only in loose relations.

The variability in domestic spaces proves that despite the coding of the space of residential rooms, human, and other-than-human dwellers navigated through the smooth spaces of the courtyards, where the relations between the components of the household occurred, e.g., in the food preparations stations like in those located in compounds U242/243/244/246/259 and U207/210/211/214/225. The changeability in domestic relations resulted in heterogeneity of households in the city. These changes, arguably, did not result from any external large-scale processes like “social” or “cultural” change, which the whole community underwent, but rather derived from household tasks including those with the highest residuality—food preparation. The repetition of mundane actions, unavoidable for the functioning of a household, every time has led to different outcomes, creating the unique character of household assemblages. The households’ emergence through everyday actions, rather than by replication of a preconcept model, asserted the households’ heterogeneity.

Conclusion

Households in Old Dongola in the period from the fourteenth to the seventeenth century coagulated in a single form coding the relations of their inhabitants. Nevertheless, the application of high-resolution methods in the analysis of activity areas within the house compounds showed various life experiences of the house dwellers and use of space, which manifested especially in smooth space of the courtyards. Even though dwellers were bounded with past relations (traditions), they retained capacity to adapt domestic space according to their needs. The lack of formal changes in the dwellings did not contradict the agency of commoners living in Old Dongola, who retained potential for change. The heterogeneity of households predominately resulted from the existence of smooth open spaces having a creative potential, which allowed unrestricted flow of matter and people and relations within the compounds. Changing relations between the human and non-human components of households emerged predominately in these spaces. On the contrary, the striated space of residential units incorporating past relations caused perpetuation of a single form for at least four hundred years, through the period of political and religious shift within the city and the region. Not only the temporality of households did not align with the temporality of the political changes, but the changes in the households were not observed diachronically. Instead, they were dependant on the individual characteristics of households.