*Article* **Landscape and Settlement over 4 Millennia on the South Side of Lake Issyk Kul, Kyrgyzstan: Preliminary Results of Survey Research in 2019–2021**

**Claudia Chang 1,\*, Sergei S. Ivanov <sup>2</sup> and Perry A. Tourtellotte <sup>2</sup>**


**\*** Correspondence: cchang@sbc.edu; Tel.: +1-315-416-7268

**Abstract:** This paper discusses the preliminary results of archaeological surveys conducted in the Juuku Region of north-central Kyrgyzstan on the south side of Lake Issyk-Kul. Our goal was to document ancient and contemporary agropastoral systems over a four-millennia period. During the surveys, about 350 loci were identified as settlements, burial mounds, graves, single artifact finds, and artifact scatters (ceramic). The areas of Juuku Valley surveyed included two discrete polygons: Polygon 1, Lower Juuku at 1750 to 1950 m asl in elevation and Polygon 2, Chak Juuku or Upper Eastern Branch Juuku Valley at 2060 to 2100 m asl in elevation. Three radiometric dates and preliminary archaeobotanical studies were conducted at three exposed profile cuts. The methods included here are: (1) pedestrian surveys; (2) use of digital maps (Google Earth, Nakarte); (3) placing archaeological loci within known chronological time periods; (4) AMS dating of charcoal samples collected from profile deposits; and (5) preliminary identification of plant remains found from archaeobotanical samples. The results of our research represent the first step toward inventorying and interpreting archaeological data in the Juuku Valley derived from field studies.

**Keywords:** archaeological survey; Inner Tian Shan Mountain region; iron age; medieval period; agropastoralism

#### **1. Introduction**

We investigate the impact of ancient farming and herding systems upon the natural landscape of a river valley and alluvial fan of the Inner Tian Shan Mountain region over a four-millennia time period. The Inner Tian Shan region is an important part of the historically known trade and migration routes described as the proto-silk routes and by others as the Inner Asian Mountain Corridor [1] linking the desert-oases of Central Asia, the Eurasian steppe, and the territories of the Chinese Dynasties from the Bronze Age (ca. 2500 BCE to 900 BCE) through contemporary times. This article reports on the preliminary results of pedestrian surveys conducted in 2019 and 2021 in the Juuku Valley on the south side of Lake Issyk-Kul (Figure 1). This landscape ranges from high alpine meadows surrounded by conifer forests to semi-arid grass-covered steppe lands. Our working hypothesis is that ancient farming and herding practices along with human settlement over a landscape were shaped by the local climate and environment, while in turn such human activities altered those natural landscapes. To test this hypothesis, we chose two sample polygons in the Juuku Valley for field survey, one in the upper valley and the other in the lower reaches. Our objective is to examine the long-term effects of land clearing for farming and herding as well as the alteration of land surfaces through the construction of architectural features. The results of our field seasons include an inventory of approximately 350 activity loci. We collected radiocarbon samples and sediment samples from three site profiles. The soil samples have been analyzed by archaeobotanists who

**Citation:** Chang, C.; Ivanov, S.S.; Tourtellotte, P.A. Landscape and Settlement over 4 Millennia on the South Side of Lake Issyk Kul, Kyrgyzstan: Preliminary Results of Survey Research in 2019–2021. *Land* **2022**, *11*, 456. https://doi.org/ 10.3390/land11040456

Academic Editors: Paolo Biagi, Elisabetta Starnini and Francesca Cigna

Received: 13 February 2022 Accepted: 15 March 2022 Published: 23 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

identified the remains of ancient seeds. These preliminary results contribute to the larger discussion of Late Holocene human impact upon the mountainous regions of Central Asia. As such, this is the first step toward creating site inventories that can be used as future planning tools for land development, cultural heritage, and the preservation of natural landscapes in the Issyk-Kul basin.

**Figure 1.** Locator map of Kyrgyzstan. Kizil Suu is the study region.

New forms of evidence for explaining the trade, communication, and migration networks of the intermontane regions of Central Asia and their importance to the Eurasian steppe have included detailed archaeobotanical research tracing the pathways of domesticated plants [2–6]. Ancient DNA studies have traced human population movements [7–9], while isotope analyses of ancient human and animal bone materials have documented mobility and dietary practices [10,11]. These laboratory studies on ancient plant, animal, and human remains augment a rich inventory of material culture from archaeological settlements. Our research turns to a more mundane but important strategy for examining ancient mobility patterns in a single valley area. Since the 1990s, archaeologists have used basic survey methods for finding and inventorying archaeological sites in the Tian Shan mountains and surrounding region for the Iron and Bronze Age materials in the neighboring Republic of Kazakhstan [12–17]. In Western Tian Shan, ethnoarchaeological observations on the seasonal movements of pastoral transhumance have noted the use of winter and spring/autumn camps, thus suggesting that the Bronze Age agropastoral site of Adunoqiaolu was also occupied during winter months [18,19]. Survey research at Juuku Valley serves as an excellent contrast to previous archaeological survey research on the alluvial fans of the northern Tian Shan range [12,14,15].

In 2018 and 2019, archaeological surveys in the Kochkor Valley of the Inner Tian Shan have been conducted [20,21]. Lynne Rouse and her colleagues [20,21] have undertaken UAV surveys and GIS mapping in the Kochkor Valley, also situated on the south side of Lake Issyk-Kul. The objectives of their surveys have been to record upland archaeological features dating from the Bronze Age through Medieval periods in conjunction with archaeological excavations conducted at the upland site of Chap at 2000 m asl. that have deposits dating from 1065 BCE to 825 BCE. Rouse and her colleagues intend to document the Inner Asian Mountain corridors along this important passageway of the Inner Tian Shan range [20]. During the Late Bronze occupational phase, archaobotanists discovered ancient seeds of *Hordeum vulgare* (hulled and unhulled barley), *Triticum* (free-threshing and possible glume wheats), *Panicum miliaceum* (broomcorn millet), *Setaria italica* (foxtail millet), and *Pisum sativum* (pea) [22–24]. Indeed, current research on plant and animal remains and material culture throughout Central Eurasia have demonstrated the importance of agropastoralism from the Bronze through Medieval periods [25].

The goal of our preliminary studies of the Juuku Valley is to develop a set of hypotheses and methods using archaeological survey for examining the long-term evolution of agropastoral systems. From the Bronze Age through the historic period, over four millennia during the late Holocene, people have practiced agriculture and pastoralism in regions of Central Asia [14,23,25]. The purpose of these preliminary surveys are to discover when and where certain groups of people (Andronovo, Saka, Wusun, Turkic, Medieval Qarakhanid, and Ethnographic Kirghiz) were inhabiting this valley. In nearby regions, the changes from the mixed herding and farming systems of the Bronze and Iron Ages to irrigated farming systems in combination with pastoralism during the Medieval and historic periods have had a significant impact. Our goal is to outline how this evolutionary process may have taken place in a single valley of Central Asia.

#### **2. Materials and Methods**

#### *Study Area*

The environmental setting and physical landscape features of the Juuku Valley (see Locator Map, Figure 2) are also important for documenting both anthropogenic and natureinduced changes in the Juuku Valley during the late Holocene. The Juuku Valley is a small intermontane valley formed by the mountain streams flowing southward to Lake IsskyKul. The lake itself is fed by 102 streams and rivers and fluctuates 20 cm in water level due to glacial melt [26]. The main glacier peak of this valley is It Tash (elevation 4808 m) and the entire valley extends 50 km north towards the southern littoral of Lake Issyk-Kul. The geology of Juuku Valley is similar to that of Dzhety-Ogyuz valley to the east [27]. The Paleozoic granites and metamorphic rocks are the foundation for the Dzhety-Ogyuz valley and neighboring valleys. Overlaying these granites are Jurassic quartzites. The Eocene and Pliocene deposits consist of a series of red sandstone formations. The surface alluval deposits of indeterminate age include gravels, pebbles, sand, and loam. The valleys and gorges on the south side of the Issyk-Kul basin have been subject to frequent earthquake disturbances, many impacting Medieval settlements [28,29].

**Figure 2.** Locator Map of Juuku Valley. The lower cluster of points is Upper Juuku, and the upper cluster of points is Lower Juuku.

The vertical zonation below the glaciers consists of rocky terrain with some traces of desert-like vegetation. From 3400 to 3000 m there are grassy meadows and a sub-alpine climate and at elevations of 3000 m to 2000 m, Tian Shan spruce trees surround meadows of perennial grasses and shrubs. This is the upland zone (2000 to 3000 m) currently used by Kyrgyz herders for grazing sheep, goats, cattle, and horses and for cultivating small fields of barley and fodder crops. In our study area, this upland zone is demarcated as the Upper Juuku Valley. From here, the gorge opens into the Lower Juuku, an alluvial valley consisting of terraces and benches above the streams and rivers (an area of about 10.5 sq km). Below this alluvial valley is a large alluvial fan (44 sq km) where the Juuku River empties into Lake Issyk-Kul. The Lower Juuku area today has large, irrigated fields of wheat, barley, oats, alfalfa and hay; flocks and herds of animals graze along the edges of the fields and in stubble areas. Along the far reaches of the alluvial fan near the shoreline of Lake Issyk-Kul (ca. 1600 m) there are marshlands and rich pasture areas.

Two sample polygons in the Juuku Valley were chosen for intensive survey because they appeared to have a high density of burial mounds constructed of stone and/or earth (kurgans) and architectural features (Figure 2). We chose these two polygons because they represent two different vertical zones. Polygon 1 in the Lower Juuku Valley, an area of 6.4 sq km, is situated in the productive zone of wheat, barley, oats, fodder plants, and winter grazing in 1750 to 1900 m asl. The 6.4 sq km area is approximately half of the total alluvial valley of the Lower Juuku Valley. Below this alluvial valley is a large fan that reaches the southern shoreline. Polygon 2 in the Upper Juuku Valley, with an area of 0.5 sq km, consists of a series of terraces and a narrow floodplain with steep colluvial deposits along the eastern branch of the Juuku River. Polygon 2 is situated in the zone of summer pasture area, tourist camping, fishing spots, and forest service reserves at elevations from 2000 m to 2100 m asl. The pedestrian surveys conducted by a team of three field archaeologists were aided by inspection of imagery from Google Earth, Soviet maps, and other digital maps (Nakarte). The loci were recorded using Garmin GPS units. Each locus was recorded by coordinates, described, and photographed in the field. From detailed notes, Excel spreadsheets were created for all site and artifact loci. During 15 field days in 2019 and 30 field days in 2021, we amassed an inventory of over a total of 1000 loci from the Kizil Suu, Saruu, and Juuku Valleys.

In addition to surface survey, we also recorded archaeological features such as pits, house structures, storage pits, and fire pits found in exposed stratigraphic profiles. Many of the exposed stratigraphic profiles were erosional or river cuts or the result of road construction and farming activities. Three stratigraphic profiles found at settlement sites were selected for more detailed analyses. Radiocarbon samples and soil samples for archaeobotanical analyses were taken from these three exposed profiles, one in the Lower Juuku polygon and two in the Upper Juuku polygon. In Polygon 1 (Lower Juuku), at Loci 387, large burnt wood samples for radiometric dating and an archaeological soil sample of 17 L were collected from a house pit. In Polygon 2, two stratigraphic profiles from two different erosional cuts were identified as archaeological house pit fills containing mudbrick remains, plastered floors, and midden deposits. At the Settlement 1 profile, 14.5 L of soil was collected for archaeobotanical analyses, and a small charcoal sample was removed for radiometric dating. Then, at the Settlement 2 profile, 11.5 L of soil was collected for archaeobotanical analyses, and a charcoal sample was removed for radiometric dating.

The wood charcoal samples (species unknown) were collected from the three profiles by the field archaeologists. Radiometric analyses were conducted at the Beta Analytic Laboratory (Coral Gables, FL, USA) using AMS methods. The samples taken from the three profiles were wood charcoal pieces. The reporting on these results includes conventional radiocarbon dating and the d13C ratio that can be used by other researchers in the future according to conventional standards [30]. Wood charcoal was dated because the field researchers did not recognize the seed material in the initial sample collection. The authors are aware of old wood effects and in future publications will date carbonized seeds as well as wood charcoal.

The soil samples were processed in the Republic of Kyrgyzstan using a SMAP flotation machine with mechanized agitation to wash sediments. The agitation process washes the archaeological sediments so that organic materials are caught in the overflow spout in geological sieves of 0.344 mm mesh. These organic materials are referred to as the light fraction, and then sieved with mesh sizes of 2.00, 1.40, 1.00, and 0.50 mm. When sorted, this fraction contains carbonized seeds, pips, leaves, and other remains. The heavy fraction was collected and sieved from 1.4 mm to 1.0 mm mesh. Both heavy and light fractions were sorted using atlases for seed identification [31,32] by two archaeobotanists at the Max Planck Institute for Human History, Archaeobotany Laboratory [33]. Preliminary archaeobotanical and radiometric analyses of these three stratigraphic columns represent very preliminary data that shall be used for refining a regional chronology for settlements and for designing a more comprehensive research study of human land use along a vertical gradient.

#### **3. Results**

During the 2019 and 2021 surveys, we registered about 350 loci (single artifact finds, sherd scatters, graves, burial mounds (kurgans), house foundations, and house depressions in the Lower Juuku and Upper Juuku). These loci were also placed in chronological sequences based on local typologies for burial monuments, settlements, and artifacts (ceramics). Our initial observations suggest that the large Medieval settlements (fortresses, citadels, caravanserais, and proto-urban towns) found in Juuku and the neighboring valleys of Sutti Bulak, Chichi Khan cover over and obscure earlier Bronze and Iron Age settlements. When both Medieval and Iron Age ceramic sherds are found on the same land surfaces and loci this indicates that later Medieval deposits cover over and obscure earlier Iron Age settlement features. The prominent appearance of above-ground mortuary features Iron Age burial mounds (kurgans), often 5 to 50 m in diameter and 0.1 to 3 m or more in height [20,21]. These burial mounds are often located near Medieval walls, farmsteads, and dwellings, therefore, marking clear boundaries between Medieval settlements and Iron Age mortuary complexes. Logically, this also means that the places where Iron Age burial mounds exist today are landscapes that were not used by later Medieval populations for house or settlement construction. These landscape palimpsests are essential to developing a deeper and more nuanced approach to site and non-site archaeology in regional contexts [34–36].

#### *3.1. Chronology*

During the survey, we established a local historical chronology based on archaeological and historical sources from the Tian Shan Mountain and surrounding regions. These phase designations are based on archaeological research conducted in Kyrgyzstan over the past one-hundred years on settlements, burial mounds, graves, and artifact collections throughout north-central Kyrgyzstan and the Semirech'ye region of southeastern Kazakhstan [36–38]. More recent kurgan imagery from UAVs from the Kok Sai area of the Kochkor Valley have been documented [21] (p. 12). Gino Caspari [39,40] has used Google Earth, Worldview2, and Ikonos imagery to record the looting and destruction of Iron Age burial mounds in Xinjiang. During the Soviet period, Vinnik identified about 17 Medieval period settlements along the SW coast of Lake Issyk-Kul [41]. Medieval sites are most easily identified by standing mudbrick walls, large enclosure walls, and standard measurements for tortuls or caravanserais [42]. Literature searches assisted us in placing our survey findings into these chronological and phase designations [38,41], see Table 1.

In this section we discuss the survey results and include some preliminary descriptions and results of three stratigraphic profiles. Each survey polygon is a self-contained unit of analysis so as not to confuse the reader. We chose each polygon on the basis of its location on the vertical gradient.

#### *3.2. Polygon 1*

Polygon 1 is a survey area (ca. 6.4 sq km) located on the Lower Juuku alluvial valley where the terraces rise about 30 m above the entrenched stream bed and is situated about 6 km south from the shores of Lake Issyk-Kul (Figure 3). Today, the upper alluvial valley is farmed by tractor and heavy equipment where large, irrigated fields are cultivated in crops of wheat, barley, oats, and fodder crops. The 323 loci documented in the 2019 and 2021 surveys are found between 1750 m asl and 1950 m asl. There is a density of 50 loci per sq km.


**Table 1.** Time Periods, Phase Designations, and Dates used for the Juuku Valley Survey.

**Figure 3.** Lower Juuku, Google Earth Image. Survey Data.

#### 3.2.1. Site Types Found in Lower Juuku

Most loci found from survey were mortuary remains: 192 mortuary remains (burial mounds, graves, and 1 mausoleum) were found, a total of 31 settlement and architectural features (17 settlement sites), and 21 artifact finds (Figure 3).

#### 3.2.2. Settlements in Polygon 1

A total of 17 settlements have been identified in Lower Juuku (Figure 4). The settlements assigned to specific time periods according to ceramic and artifact finds in addition to architectural features such as the citadel (shakristan), the surrounding residential areas (rabat), stone foundations, and room or house depressions.

**Figure 4.** Histogram of site types in Lower Juuku: mortuary features, settlements and architectural features, and artifact finds. The *x*-axis represents types of loci, and the *y*-axis or Field 2 is the number of loci.

#### 3.2.3. Locus 387: Historic or Ethnographic Kirghiz House

One site (Locus. 387, Table 2) is a large settlement located on the eastern bank above the lower Juuku stream at an elevation of 1892 m asl. The settlement is located above the road cut. On the surface, there is one visible single room foundation that is probably part of a much larger settlement. The single row of rocks is rectangular in form, about 4.5 m east–west by 5.5 m north–south. Inside the stone wall foundations is a depression. The profile section, located along the south bend of a road cut, is 4.5 m in length and 1.4 m in depth from the present ground surface (Figure 5). In the profile is the house fill of a structure built of mudbrick that was destroyed by a bulldozer cut when building the dirt road. The humic layer covers a thick layer of mudbrick (about 20 cm thick). Underneath the humic and upper mudbrick layer is a thick midden deposit about 40 to 60 cm thick. This midden layer consists of a thick lens of cultural material including animal bones (cattle) and chunks of charcoal, some as large as 1 cm in diameter. No artifacts were found in the profile, although coarseware ceramic sherds were found on the ground surface near the road cut. The flotation samples were taken from this thick cultural level about 60 cm from the present ground surface along with over 10 g of burnt wood (charcoal) for radiometric dating. At 60 cm to 1 m below the present ground surface was a thick layer of mud brick foundations, yellow buff in color. This mudbrick layer may have been the original floor level because below the mudbrick were large river cobbles probably used as foundation stones. Upon initial inspection, we identified the site as a Medieval settlement. The reason we believed the site to be from the Medieval period was because of the redware sherds found near the road cut. However, the radiometric dating places it within the ethnographic Kirghiz period = (*floruit* 1682–1932 cal CE). In the opinion of the researchers, it is doubtful that the wood charcoal samples were contaminated by either natural or cultural forces due to the intact stratigraphy found in this road cut. In stratigraphic profiles where there are not actual index fossils (diagnostic sherds and metal artifacts) it is easy to misjudge the dating of a house pit, especially when ethnographic houses are also constructed of stone foundations and mudbrick walls, the same materials used to construct Medieval houses.

**Table 2.** Results from Radiocarbon Sample of Locus 387, Lower Juuku, Settlement 1.


**Figure 5.** Photograph showing the Profile of the Kirghiz Ethnographic Period Settlement.

These are AMS (accelerated mass spectrometry) dates. The accuracy is at 95.4%, and each portion or percentage represents the range (or ranges) with an associated probability of an identifiable timescale [30]. IRMS is the method of measurement used at the Beta Analytic Laboratory. The results were calibrated using INTCAL20 [43]. This measurement found in Column 4 (d13C o/oo) can be used by future investigators to recalibrate these results should new calibrations be established.

#### 3.2.4. Preliminary Archaeobotanical Results from Locus 387

A caveat is in order here. Most archaeobotanical studies conducted at sites such as the Chap site, a Late Bronze Age site and Paykend, or a Qarakhanid Medieval site depend on the collection of large samples of archaeological sediments, sometimes entire house fills or pit fills [22,44]. Our soil samples are very small and can only be considered as preliminary in nature. Thus. the results from the flotation of these samples must be interpreted as preliminary results and shall be reported upon in greater depth in another publication. From the 17 L sample, the archaeobotany team discovered a total of 41 seeds, the majority being wild plants. The field crops included barley (*Hordeum vulgare*), wheat (*Triticum aestivum*), and peas (*Pisum sativum*), along with the major component of carbonized chenopods and weed seeds of wild *Fabaceae* and grasses (*Poacae*) [33].

#### *3.3. Polygon 2*

Polygon 2 or Upper Juuku (Chak Juuku) is 20 km from the edge of the lake and is a narrow valley that consists of dissected terraces on either side of the eastern branch of the Juuku Gorge, a section where red sandstone formations form the steep gorge (Figure 6). This area was chosen for intensive survey due to its location on a terrace above the eastern branch of the Juuku Stream at an elevation about 300 m higher than the Lower Juuku survey polygon. This survey area consists of the two banks (east and west) of the Eastern Juuku stream and is about 0.5 sq km in area. Approximately 37 loci were identified per 0.5 sq km. The loci range from 2060 m asl to 2100 m asl. The natural vegetation includes semi-arid shrubs and grasslands with pockets of spruce in the higher elevations and willows and aspens and other riparian species along the stream banks.

**Figure 6.** Upper Juuku (Chak Juuku), Google Image, Survey.

In Table 3 we describe the settlements, burial mounds, graves, and artifact finds in the Chak Juuku area. In addition, there are two other settlements of unknown age (Loci 184 and Loci 185) found in the vicinity of the Settlement 1, the Wusun period site. It is possible that these stone foundations could be associated with Settlement 1, although no temporal indications such as ceramic sherds were found at these settlements.


**Table 3.** Sites and Finds from Upper Juuku by time period.

\* Groups are defined as linear clusters of graves or kurgans.

#### 3.3.1. Settlement 1

Settlement 1, situated at 2057 m in elevation on an upper terrace above a ravine: It is a Wusun Period site dating from 22–206 cal AD (see Table 4). The settlement is located on both sides of the erosional gully. On the north side there are no indication of house foundations, only shallow depressions of house pits. The south side of the gully consists of the outline of three or four room blocks. Room 1 is about 5 m (north–south) × 3 m (east–west), and its west wall outlines Room 2, which is 5 m (north–south) by 4 m (east–west). The outline of Room 3 is further east, where there are only traces of a south wall (3 m in length) and a trace of a north wall (1 m in length). On the south side of this room block is a thick stone wall two courses wide. These stone walls are constructed of irregular boulders. The

erosional cut of the exposed profile is about 2.5 m in depth from the surface and runs in an east–west direction from the reaches of a small sandstone canyon that dissect this terrace. The stratigraphic profile is located on a south-facing erosional cut found on a steep terrace bank at 2044 m asl. The exposed profile is 3 m in length and has a depth from the present ground surface 2 m in depth (Figure 7). There are roughly about six different archaeological layers within the house pit fill. Each layer is around 20 to 30 cm thick of midden deposit consisting of chestnut and red-clay soils intermixed with charcoal and disintegrating mudbrick. There were many animal bones fragments and one cattle vertebrate found at 80 cm below the present ground surface. A redware Iron Age ceramic with white slip discovered at 90 cm below the surface. There were at least three discernible layers of thin plaster floors. At about 1.5 m below present ground surface there were small angular stones and small pebble and sand deposits that appear to be subsoil (non-cultural soil). From an exposed profile of charcoal room fill, three archaeobotanical soil samples (14.5 l) were taken in layers 4 and 5, near where the Iron Age ceramic sherd was found. The small charcoal sample was also taken in the same vicinity.

**Table 4.** Results from Radiocarbon Samples from Upper Juuku at Settlement 1 (Locus 183) and Settlement 2 (Locus 165).


Preliminary archaeobotanical remains found at Settlement 1: There were a small number of carbonized seeds identified from four domesticated crops including barley (*Hordeum vulgare*), wheat (*Triticum aestivum*), broomcorn millet (*Panicum millaceum*), and foxtail millet (*Setaria italica*). More than half the assemblage included wild plants such as chenopods, wild legumes, and cleavers [33].

**Figure 7.** Profile of the Iron Age Site at Upper Juuku, two archaeologists working at the profile.

#### 3.3.2. Settlement 2

Settlement 2, situated at an upper terrace at an elevation of 2090 m asl: It is a series of double-walled stone alignments of at least four large room blocks and measures about 17.5 m × 12 m. The stratigraphic profile examined is a west-facing cut found at a deep erosional gully that dissects this upper terrace (Figure 8). When this cut was first discovered, a sheep scapula was found at about 1.3 m in depth from the present ground surface. The cultural levels at this profile are complex since the house fill represents at least three or four different occupation levels. The burnt wood (charcoal) sample and two flotation samples were taken approximately between 50 to 80 cm below the present ground surface in the second to third fill or midden sequence. At about 1.7 to 2.0 m below the surface are a series of thin buff-colored plastered floors or laminations that may represent the different layers of an ancient *sufa* or *kang* (sleeping bench). We did not take charcoal wood samples for dating or the flotation samples from the *sufa* or floor levels because there was no visible carbonized ash, soil, or charcoal at those levels. The site is dated to the Medieval Qarakhanid period and has an approximate radiometric dating of 990–1050 cal AD (see Table 4). This is also confirmed by the Medieval redware ceramics and a granite grinding stone found on the surface of this settlement. The archaeobotanical remains showed a paucity of seeds from the 11.5 L of collected sample, only one barley seed (*Hordeum vulgare*) was identified, the rest were wild seeds [33].

**Figure 8.** Photograph at Profile of Medieval Settlement at Upper Juuku.

#### 3.3.3. Radiometric Dating

At Settlement 1 (Iron Age site) and Settlement 2 (Qarakhanid Phase) Medieval Period charcoal samples were taken from the profiles. AMS dating was obtained from both settlements. The following dates are reported in Table 4.

These are AMS (accelerated mass spectrometry) dates. The accuracy is at 95.4%, and each portion or percentage represents the range (or ranges) with an associated probability of an identifiable timescale [30]. IRMS is the method of measurement used at the Beta Analytic Laboratory. The results have been calibrated using INTCAL20 [43]. This IRMS measurement found in Column 4 (d13C o/oo) can be used by future investigators to recalibrate these results should new calibrations be established.

#### *3.4. Mortuary Complexes*

There were many lines of Iron Age burial mounds found in both Upper and Lower Juuku. The largest number of burial sites were earthen or stone mounds known as *kurgans*. Most kurgans are from the Saka period, represented by linear clusters of stone or earthen kurgans. The clusters range from 3 mounds to 16 mounds in alignment. In some cases, the stone Saka kurgans (ranging in diameter from 2.8 m to 11 m in diameter and from flat to 1.5 m in height) are interspersed or found in proximity to the earthen Saka kurgans (ranging in diameter from 5.5 m to 97 m in diameter and from 0.4 to 2.5 m in height). In Lower Juuku, the largest kurgan cluster (16 kurgans) covers a distance of 220 m and is aligned 4.6 degrees east of North. The three largest earthen kurgans cover a linear distance of 400 m and are aligned at 10 degrees west of north. A third group consists of six kurgans covering a distance of 270 km and follow a direction of 13 degrees west of north. In Figure 9 there are two large Saka period earthen mounds found in the Lower Juuku. These earthen kurgans are identified as Saka period elite kurgans based on their size, earthen construction, and the *krepida* (rock apron) found on the north side of the kurgan [45].

**Figure 9.** Photograph of large earthen Saka kurgans found in the Lower Juuku Valley.

In addition to the Saka kurgans are Wusun kurgans that have an outer rectangular or square stone outline with an inner circular stone ring. The Wusun kurgans range in size from 2.5 m × 2.5 m to 9 m × 9 m and are flat to 1.2 m in height; the average is 0.4 m in height. Of particular interest is the close proximity of the Saka stone and earthen kurgans to the Wusun stone kurgans. There appear to be clear spatial boundaries between the Saka and Wusun graves, something that requires further investigation.

#### *3.5. Stone Corral*

In the Upper Juuku Valley, a stone corral was located. Currently, the chronological period for this stone corral is unknown. The corral is built of boulders and medium-sized stones, in places one to three courses in width. The outline of the corral is almost rectangular with rounded corners. It measures 20 m (north–south) × 13 m (east–west). To the north of the corral is a small rectangular stone structure that measures 8 m × 6 m. This could be a storage room or a dwelling used by herders. This corral is an important feature of agropastoralism in the Upper Juuku Valley (Figure 10).

**Figure 10.** Photograph showing a stone corral in Upper Juuku.

#### *3.6. Artifacts*

Many artifacts were either found as isolated finds or as parts of scatters. The grinding stones were usually made of granite with a concave surface. Figure 11 is a granite grinding stone found as an isolated surface find. The grinding stone measures 30 cm (length) × 18 cm (width) × 5 cm (thickness).

**Figure 11.** A granite grinding stone found on the surface.

Over 100 ceramic sherds were found from Upper and Lower Juuku surveys. The majority of the ceramics were redware coarse wares with sand or crushed rock temper. During the survey we made rough distinctions between the Iron Age handmade ceramics (coil or slab) and the Medieval wheel-thrown ceramics. In addition, we found handle, rim, and base fragments, spindle whorls, and a very few pieces of glazed Qarakhanid ceramics. More detailed analyses of these ceramics will be conducted by a ceramics specialist in the future. Figure 12 shows the interior of redware ceramic sherds, some with fabric impressions and red slip. These ceramic pieces date to the Iron Age and represent hand-made or slow wheel ceramics. A detailed description of fabric-impressed ceramics and actual fabric fragments from Bronze and Iron Age contexts at the site of Begash in the Dzhungari Alatau Mountains of southeastern Kazakhstan was conducted by Paula Doumani Dupuy and her colleagues [46]. Similar redware sherds, sometimes with light slip, have been found at Iron Age sites in the Talgar region of southeastern Kazakhstan [14,47].

**Figure 12.** Redware ceramic sherds; interiors showing fabric impressions and red slip, probably Iron Age period ceramics.

#### **4. Discussion**

Although an old-fashioned, tried and true method, we prefer to walk the ground. Pacing the diameter of a burial mound by foot or measuring how high it rises above the ground by standing at its base and eyeballing its relative height gives us a three-dimensional perspective. We also follow a line of mounds along a ridge, just as one experiences a row of graves in a modern cemetery. Since this land was used to memorialize the dead, where did the people farm crops or herd sheep? New technologies such as UAVs (drones) or visual inspection of digital map images of burial mounds, graves, or houses are the initial way to locate sites from your desktop. Afterwards, the field archaeologist checks these loci on the ground. A digital image usually does not show small objects (ceramic sherds and stone tools) on the surface. By walking across a settlement, we often find ceramic sherds or grinding stones inside enclosures. Diagnostic ceramic sherds may indicate when a site was occupied. We begin to intuit why ancient people selected these terraces, ridges, and valleys. Ancient walls and room depressions take form as habitations. Our field sketches as notes create a memory bank of what we have seen and found. If we find a piece of broken pottery on a freshly ploughed field, it triggers our recollections. Why was this pot dropped here and did the plough move the broken sherds to this spot? Often the sherds are most visible in farmers' furrows. So, beneath our feet may lie storage pits or trash middens. The field surveys help us discover ancient settlements. In the future, we or others can undertake full coverage survey by traversing complete landscapes.

The site density at these two polygons in the Juuku Valley appears to be much higher than site densities on the Talgar alluvial fan on the edge of the northern Tian Shan mountains. Moreover, the Juuku Valley appears to have a larger representation of sites from the Bronze Age through historic periods. Thus, the Juuku Valley is an excellent micro-region for more in-depth investigations of agro-pastoral systems over a long period of time and in different cultural phases. In 2021, we conducted further reconnaissance of Iron Age settlements, often noting that it was more difficult to find and locate early period settlements from the Bronze and Iron Ages than we initially expected. Most likely, earlier Bronze and Iron Age settlements have been buried by more recent Medieval and historic sites or have been destroyed by modern agricultural practices such as ploughing and irrigation. In the Lower Juuku, where the majority of large-scale tractor and irrigation agriculture occurs, surveys may yield better results if conducted in the early spring or early fall before intensive cropping and after harvesting.

Today the Lower Juuku Valley is well-suited for the cultivation of wheat, barley, oats, and fodder crops during the summer months when large tracts of land can be irrigated and cultivated using large machinery. Sheep, goats, cattle, and horses are often pastured on the agricultural stubble after harvest. The Upper Juuku is more suitable for summer pasturelands for cattle, sheep, goats, and horses. Its rich forest and riparian areas attract a variety of wild animals and fish, deeming this area desirable for upland herding, fishing, and foraging. In the past there may have been pockets of arable land in the Upper Juuku for the cultivation of short season crops such as barley and millets.

#### **5. Conclusions**

Archaeological surveys are necessary in this region of north-central Kyrgyzstan for two main reasons: (1) the upland areas far from the perimeter of Lake Issyk-Kul are poorly known by Kirghiz archaeologists; (2) the south side of Lake Issyk-Kul is an under-developed tourist area. As this area continues to develop as a tourist area and as an agricultural region for the cultivation of wheat, barley, and fodder, more archaeological sites will be destroyed. Settlement archaeology, especially for the prehistoric periods, is little known in the Republic of Kyrgyzstan. Recently, research on Epipaleolithic through Neolithic layers at Obishir have been conducted by international teams of archaeologists [48–50]. As archaeologists begin to explore the early beginnings of foraging, pastoral, and agricultural economies in Kyrgyzstan, surveys such as the Juuku Valley and the Kochkor surveys will become more essential for the next generation of archaeologists [20,21,51]. Our work represents a modest first step in establishing systematic archaeological survey methods to reconstruct settlement–subsistence systems in this region of Central Asia. The results of the surveys will also be used to select areas for test excavations and large-scale block excavations of both mortuary and settlement complexes. Future test excavations shall yield archeozoological and archaeobotanical materials along with ceramics, stone and bone tools, and metal indicative of ancient herding and farming adaptations. Finally, it is our hope that these physical, archaeological, and cultural landscapes will be preserved through local and national efforts [52]. An inventory of archaeological loci then becomes the initial path toward preserving such fragile landscapes.

**Author Contributions:** C.C. and P.A.T. provided the conceptualization of the paper. The methodology for the field surveys was designed by S.S.I. and P.A.T.; quantitative analyses and tables were prepared by C.C.; validation of results were undertaken by S.S.I., C.C. and P.A.T.; resources were obtained by all three authors; writing, including review and editing was performed by all three authors; visualization, including photography and digital mapping, and the graphical abstract was designed by P.A.T.; field and laboratory supervision was undertaken by S.S.I.; project administration was conducted by S.S.I.; funding acquisition by C.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** The funding for this fieldwork study was supplied by the National Geographic Society, "The effects of earthquakes on Inner Tian Shan passages: Iron Age and Medieval Landscapes" (NGS-59769R-19). No funds were provided for publication.

**Institutional Review Board Statement:** This project did not require an IRB statement since no human or animal subjects were used in this scientific study.

**Data Availability Statement:** The results of the archaeological surveys are currently archived by C.Chang (USA) and S.S.Ivanov (Kyrgyz National University). These include GPS data points, digital mapping, fieldnotes, and preliminary reports. The radiometric data is archived by Beta Analytic Laboratory in Coral Gables, Fl (USA). Artifact collections (ceramics, metal, and stone) are archived at the Kyrgyz National University in the Faculty of Far Eastern Studies. The archaeobotanical material is archived at the Max Planck Institute of Human History, Archaeology Department under the supervision of Robert N. Spengler, III, Laboratory Director of the Archaeobotany Laboratory.

**Acknowledgments:** The authors of this paper are grateful to the fieldwork support provided in 2019 by Kathryn J. Franklin, Department of History, Classics, and Archaeology at Birkbeck, University of London. The preliminary archaeobotanical studies were generously conducted by Basira Mir-Makhamad and Robert N. Spengler, III at the Max Planck Institute for the Study of Human History, Archaeology Department in 2021. Beta Analytic, Inc. (Coral Gables, Florida) conducted the AMS radiometric dating of three charcoal samples. We are particularly grateful to Ronald Hatfield, President of Beta Analytic, Inc., for his detailed explanations of the variations noted in two of our AMS dates and for providing uniform guidelines for accurate C14 reporting.

**Conflicts of Interest:** The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


**Eduardo Herrera Malatesta**

Department of History and Classical Studies, Aarhus University, 8270 Højbjerg, Denmark; ehmalatesta@yahoo.com

**Abstract:** This paper presents an archaeological reconstruction of indigenous landscape transformations in the first colonized region of the Caribbean. The arrival of Columbus in 1492 in the northern region of the island of *Haytí* (the current Dominican Republic and Haiti) signified a profound change in the lives of the island's communities, transforming their everyday actions and their perceptions of landscape. To address this complex topic, this research tackled a key problem in landscape archaeology: while the "landscape" concept has been extensively debated, there is a growing tendency to use the concept without clear definitions and to obscure important methodological aspects of how scholars bridge the divide between their conceptual definitions and the archaeological record. This paper approaches this problem by applying the concepts of 'sites as tendencies' and 'contested taskscapes'. This theoretical and methodological framework allows for the reconstruction of the indigenous landscape and, more importantly, highlights how the colonization process impacted the everyday tasks and perceptions of Hayti's indigenous people through the profound transformation of their landscape.

**Keywords:** landscape; taskscape; indigenous communities; Spanish colonialism; Caribbean

#### **1. Introduction**

The idea of landscape as used in archaeology today is a complex product of decades of archaeological, anthropological and geographical debates [1–9]. In particular, the ambiguous use of the concept today [10] reflects the challenges faced by both processual and post-processual archaeologists in making sense of past realities within their specific theoretical frameworks [11–14]. This dichotomy between processual and postprocessual trends is still palpable, yet clear discussions on the methodological implications of these two very different approaches are less common. One group of scholars continues to use the concept of landscape in the processual sense, i.e., they understand landscape as the passive environmental context for human behavior. By contrast, other researchers use the term to refer to the resulting cultural meanings that humans create in the process of interacting with their surroundings. The issue does not arise from the use of either of these perspectives on landscape per se, but rather the lack of explicit discussion of their methodological and theoretical implications, particularly when scholars seek to draw on a combination of both perspectives.

This problem becomes salient when, for instance, one tries to combine the advantages of regional survey, geographical information systems, and spatial statistics with a conceptual framework that studies transformations in landscape perception as a result of cultural conflict. For this, an articulation of processual and postprocessual methods and theories is imperative. The proposal presented in this paper is based on Tim Ingold's idea of the Dwelling perspective. Over 30 years ago, Ingold suggested a way to overcome the dual and supposedly contradictory ideas of landscape from the processual and post-processual perspectives in archaeology. Ingold aimed "to move beyond the sterile opposition between the naturalistic view of the landscape as a neutral, external backdrop to human activities,

**Citation:** Herrera Malatesta, E. The Transformation of Indigenous Landscape in the First Colonized Region of the Caribbean. *Land* **2022**, *11*, 509. https://doi.org/10.3390/ land11040509

Academic Editors: Paolo Biagi and Elisabetta Starnini

Received: 4 March 2022 Accepted: 23 March 2022 Published: 31 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

and the culturalistic view that every landscape is a particular cognitive or symbolic ordering of space." [15]. The key to achieving this, lay in the realization that the full power of the landscape concept extends beyond the visual and includes "a world in which we can expect to find formations of the land such as hills and valleys, mountains and plains, interspersed with settlements such as villages and towns and threaded by paths, roads and waterways." [16]. For Ingold, there was no fundamental difference between 'landscape' and 'environment', as both are the result of the intertwining combinations of natural and cultural elements over time. The result of the histories of interactions between people, things and nature are that "landscape is not land looked at but land shaped ... In this medieval sense, landscape already couples the land with tasks of shaping: landscape is taskscape because to shape the land is to work it" [17].

Ingold used the concept of taskscape to articulate the complexities of the intertwinement between people, things, tasks and the environment [4]. The concept of taskscape is a bridge between two realms: for archaeology, it can be a heuristic tool that connects the shape and content of the archaeological record with the idea of landscape. There are several recent examples of the use of taskscape in archaeology, for example see: [18–21]. In the Haytí case study, working through the taskscape perspective allowed for the integration of material evidence (e.g., archaeological sites and artefact distributions at the regional scale) with theoretical notions about human movement, action, and landscape knowledge at different spatial scales. This integration, in turn, proved key to understanding both the ways that indigenous tasks shaped the island's landscape and how these were collectively transformed after 1492.

Yet, for two important reasons, the taskscape concept by itself was not sufficient to properly achieve the desired goal of this research. First, in order to avoid ambiguities in the use of 'landscape' and 'taskscape', clearly defined spatial categories had to be developed. Following an extensive review of the concept of the archaeological site, the notion of 'sites as tendencies' was elaborated as a means of defining of taskscapes in the archaeological record. Second, as the case study focused on contexts of colonial war, enslavement and displacement, the concept of taskscape was reconsidered in terms of how it might relate to the contexts of conflict that the indigenous people suffered. To overcome this second challenge, current debates over landscape, colonization and conflict in archaeology were drawn on to create the idea of 'contested taskscapes'.

A complete review of the arguments for the definition of these two concepts is beyond the scope of this paper [22]. In general terms, the idea of 'sites as tendencies' is based on the understanding that human actions in the world leave traces of their particular intentions and activities, which, in addition to their spatial recurrence, also tend to be recurrent in time. Each particular site contains evidence of one or more tasks that were carried out by human actors in particular environmental contexts and places. The analysis of materials and sites from this perspective allows for a clear definition of taskscapes that articulates directly with the archaeological record. The idea of 'contested taskscapes' came as a second stage in the analytical process and emerged through the combination of Ingold's taskscape [4] and Bender's work on contested landscapes [23–25]. Briefly stated, the contested taskscape concept permitted the identification and classification of the material evidence and patterns left on the terrain by cultural conflicts in the past. By grouping and categorizing these conflicts on the ground, it was possible to identify contested taskscapes and define the resulting landscape transformations.

#### **2. Material and Methods**

The study area where these two concepts were explored was the coast of Montecristi, a province located in the northwestern Dominican Republic (Figure 1). After several field seasons in the region, a total of 102 previously unregistered archaeological sites were recorded and grouped by size, location, ceramic affiliation and function [26].

**Figure 1.** The province of Montecristi (Dominican Republic) in the context of the contemporary Caribbean.

The regional archaeological pattern (Figure 2) in the Montecristi coastal area consists of four groups of archaeological sites. The first of these groups include the large habitation sites (>3 ha, *n* = 8), which are always located higher than 150 masl. All of these sites contained ceramics with Meillacoid (800–1550 AD) attributes as the main component, in association with ceramics from the Chicoid series (900–1700 AD) and a mixed ceramic component that includes stylistic attributes from the Meillacoid and Chicoid series (1000–1700 AD). These sites have a high diversity of lithic artefacts (axes, scrappers, flints, cores, hammerstones, and grinding stones) and a limited variety of shell objects (axes and scrapers). Most sites also showed a great assortment of mollusk shells, with a total of 17 species identified across the region, with 5 to 8 of these species recorded at each of the large habitation sites. The most common species at all sites were *Codakia orbicularis* sp., *Lobatus costatus* sp., and *Lobatus gigas* sp. (Figure 3). These sites are usually located over 1 km from water sources, while fewer are located at a distance less than 500 m.

The second class of sites includes medium-sized habitation sites (between 1 to 3 ha, *n* = 17), which are mostly located below 80 masl and with only a few between 100 to 200 masl. Most of these sites contain exclusively Meillacoid ceramics and only very limited numbers of lithic artefacts, mostly axes and/or hammer stones (usually 1 to 2 per site). The mollusk specimens are not diverse, with most sites containing 3 to 5 species. Unlike the previous category, these sites are usually located close to water sources (<500 m).

The third class of sites–the small habitation sites (<1 ha, *n* = 49)–do not show a specific pattern of elevation. At half of these sites, it was not possible to identify a specific ceramic affiliation as materials tend to be very fragmented and scarce. In the other half, mostly Meillacoid ceramics were registered, and in a few cases, together with Chicoid and mixed Meillacoid-Chicoid components. Stone artefacts were limited at these sites, with only one lithic artefact often recorded (usually an axe or a hammerstone). The same pattern generally held for shell artefacts, with only one artefact observed, in most cases, this was either a hammer or a *gubia* (a shell artefact used in woodworking and agriculture). In most of these sites, 1 to 3 species of mollusks were recorded.

**Figure 2.** Registered archaeological sites in the coast of Montecristi, Dominican Republic.

**Figure 3.** Most common mollusk shells registered on the research area.

The final category included is the resource exploitation sites (<3 ha). These sites are characterized by the absence of ceramics or other lithic and shell artefacts. They are exclusively located along the coastline and are usually very close to each other. This pattern suggested that, rather than defining them as sites, they should be perceived as "exploitation zones". The material culture associated with these sites was very limited, comprising only a few ceramic sherds (either unidentified or Meillacoid) and lithic artefacts (either scrapers or flakes). The few shell artefacts recorded were perforators, gubias or hammers.

The preliminary interpretation of these results suggests that the large sites were permanent living places that formed a focus for daily activities. Based on the diversity of material culture present, such sites may have also served as locations where exchange networks were maintained at the local and possibly regional level. Second, the medium-sized sites seem to have been related to temporary or permanent housing activities and, given that both the material evidence and its diversity is less, they may have had fewer inhabitants or existed for shorter periods of time. As in neighboring areas, these medium sites also could have been related to specific activities such as agriculture or could have served as *rancherías* (small hamlets) for hunting parties or fishermen going to the coast [27]. Third, the small sites might have been related to specific tasks, possibly daily activities which recurred over time but that did not need to be done in exactly the same place (e.g., a location where marine products were processed once or multiple times before being transported to hamlets or the main settlement). The patterning of these sites highlights the dynamics and mobility of people in their landscape and the decisions related to this movement.

#### **3. Discussion: Indigenous and Spanish Taskscapes**

The taskscape concept is used here as a bridge to connect the materiality of the archaeological record with the abstract idea of landscape. In conceptual terms, this is rendered by using the taskscape as a layer within the complexity of the landscape [28]. A recent example of this was provided by Ingold [17] by using one of Brueghel's paintings. This painting shows a sixteenth-century European settlement, with representations of people's day-to-day activities being carried out in specific places within the town. From an archaeological perspective, this painting highlights three main aspects: the spatial relationships, the idea of spatial scale, and how they are used to define a place. Drawing on Ingold's reading of this painting, it is clear that 'place' can refer to both the location where a specific task is being carried out as well as the entire town. For Ingold, a place is defined by the overlapping lines left by people's movements through the world, in a process he defined as the meshwork of entangling lines [29]. If each location in the painting is the result of a person or a group of persons' movement in the town and the world, each location (including the town) constitutes a place. Since Ingold's definition would fits multiple scales, the idea of taskscape can be deemed multiscalar. Scale may not be an issue when working the idea of landscape in art or ethnography, since the observer in these disciplines has access to the whole "image". In archaeology, however, the spatial scale is a challenge as considering different spatial scales for analysis will potentially result in different reconstructions and interpretations of past humans patterns within the landscape.

To minimize potential misinterpretation and biased classification, and following the idea that the taskscape forms a key level in the process of archaeological analysis, three spatial scales of taskscape were defined for the case study presented here, each of which relates to the tasks and movement of indigenous people in the past.

#### *3.1. Small-Scale Taskscapes*

The first scale where a set of tasks was identified was that of the site. Based on all of the previously-described size and functional categories, the primary activities at this scale are related to general domestic activities and the exploitation of land and sea resources, which include tasks of craft production, as well as agricultural, hunting, and fishing activities. This first level of the indigenous taskscape relates directly to daily activities, that is, with the decisions and actions at the level of local communities and their dwelling in the world. The general pattern reconstructed from this evidence is that there tends to be a large place (large-size sites) that can be associated with permanent areas of habitation and the common tasks of a settlement. These main settlements were "surrounded" by other places (medium and small-size sites and exploitation of sea resource sites) that were associated with highly specialized activities such as marine resource processing, lithic workshops and/or agricultural production. This whole combination of tasks constitutes the first layer of taskscape, and refers to the internal taskscape of each settlement or areas close to the main settlement.

#### *3.2. Medium-Scale Taskscapes*

A second scale of the indigenous taskscape occurs at the level of the study area. At this scale, it was possible to identify tasks that interconnect most of the archaeological sites. For example, the presence of specific materials at particular sites, such as mollusk shells and lithic artefacts, indicate tasks related to specialized exploitation of marine resources. The presence of these same mollusks, however, at some of the larger settlements indicates their exchange and/or trade throughout the research area. This taskscape also suggests daily interactions between different communities from distant settlements (possibly between culturally different populations). Figure 4 shows a representation of the connections between particular sea resource sites where a specific mollusk shell-type was identified (in this case *Cittarium pica*, sp.) and the habitation settlements where that same type of shell was recorded. As can be inferred from the image, there seems to be a pattern of distribution and redistribution of the marine resources across larger zones of the study area.

**Figure 4.** Ideal representation of medium-scale taskscapes.

#### *3.3. Large-Scale Taskscapes*

The third scale of the indigenous taskscape was defined when evaluating the regional distributions of the ceramic series present in the region. In the northern sector of the island, two ceramic series dating to the Late Ceramic Age (1100–1500 AD) and the early colonial period (1500–1600 AD) have been classified: Meillacoid and Chicoid [30–33]. The particular distribution pattern of the sites related to the Meillacoid ceramics between the coastal areas of the Montecristi province and the eastern province of Puerto Plata, as well as the characteristic distribution patterns of the Chicoid sites in the region, seem to highlight a political taskscape (Figure 5). This taskscape could have been based on the relationships and interactions between different communities and the rights to access certain areas. Sociopolitical and cultural reasons could be the main causes for the absence of sites with Chicoid ceramics in the area of the coast of Montecristi. The coastal area of Montecristi also has

important topographic (its location at the end of the Cordillera Septentrional) and ecological features (endemic areas of the only mammals on the island, the hutia and solenodon, and easy access to the mangroves), which may have played a role in the interest of Meillacoidusing communities in maintaining control over this area and its resources. Alternatively, conflicts among indigenous groups could also have contributed to this particular pattern.

**Figure 5.** Large-scale taskscapes. Ceramic series distribution pattern: (above) Meillacoid, (below) Chicoid. The data for this image comes from [27,34–36].

#### *3.4. Indigenous Landscape*

The indigenous landscape, that the early Spanish observed, can be reconstructed is composed of a variety of places and artefacts that indicate the diversity of activities carried out by past communities. The patterned distribution of places revealed the past inhabitants' recurrent interests of interacting, exploiting and selecting activity places based on topography. The pattern is defined by "central" places related to habitation, trade and perhaps political centrality, followed by a series of medium to small places related to other non-domestic activities such as the exploitation/processing of marine resources, agriculture and, possibly, other activities such as logging and/or production of lithic artefacts. In addition, the non-domestic sites tend to afford the best visibility of the surroundings and marine environment, as has already been proposed for neighboring areas in the region [35,37,38]. These patterns are also observed in other neighboring research areas, such as the coastal region of the Puerto Plata province and northeastern Haiti, with some local particularities possibly related to the topography and culture of the different indigenous communities. The indigenous landscape before 1492 was dynamic and diverse, where communities seem

to have shared similar cultural features that are evident across multiple scales. However, at the largest analytical scale, there are indications of cultural heterogeneity that might have resulted from ethnic differences, cultural territories and/or conflict relations.

#### *3.5. Spanish Taskscapes*

A good example of the Spanish movements during the early phase of colonization and that formed part of their strategies to reach the inland gold area and control of indigenous population is the so-called *Ruta de Colón* [39,40]. Figure 6 shows a reconstruction of this route as well as some hypothetic lines of the Spanish movements, based on the early chronicles reports. The figure also shows the location of the Spanish towns and forts, from the earliest town of La Isabela (1493) to the fort at Jánico (1494) where the gold mines were located. Following Ingold's [16,29] definition of place, each fort and town formed a 'node' resulting from the repeated passage and selection of certain places as a consequence of the move from the coast to the mainland. In addition, the selection by the Spanish of those specific places for settlement responded to the several indigenous settlements that already existed along this route, as reported by the early chronicles [41–44]. The 'lines' created by the Spanish movement is evidence of their intentions to successfully exploit and transfer gold to the coast, as well as the need to dominate the territory and its people. The forts were built near resistant or important settlements of the indigenous groups. For example, the fort of La Magdalena was built in the Cibao valley, close to Los Hidalgos pass through the northern mountain range, within the area known by the indigenous people as the province of Macorís, and to the north of the supposed area of residence of Cacique Caonabo. Cacique is a word the indigenous people used to refer to their leaders. The early Spanish compare this term to that of king. This fort was destroyed by the indigenous people of the area and, shortly after, the Spanish replaced it with the fort of La Esperanza [43]. On the other hand, some Spanish towns were built near key natural resources (gold veins, rivers, coast) as well as near important indigenous settlements. For example, the town of Concepción de la Vega was founded near the settlement of Cacique Guarionex, with whom the Spanish initially had formed an alliance and later fought several fierce battles.

The initial colonial landscape was shaped by the exploitation of the island's mineral and human resources through the *factoría* system [45]. This was a colonialist system originally developed in Portugal and applied to the Spanish in the conquest of the Canary Islands [46]. The Spanish landscape was made up of two taskscapes. One is based on the exploitation of resources and the other on the military control of the territory. These taskscapes allow an understanding of part of the Spanish logic in conquering the indigenous world, and represent elements of the transformation of the indigenous landscape into the Spanish colonial one during the first years of colonization. The Spanish landscape during these the first years on the island of Haytí was characterized by, in the first place, closely spaced settlements to aid the movement of troops and suppress the constant indigenous uprisings, such as the one that destroyed the fort of La Magdalena or the uprising of the caciques Guarionex and Mayobanex [41]. Second, these settlements, whether villages or forts, served as forms of control and domination of the local population and protected of the movement of resources between the inland and the coast.

**Figure 6.** Spanish settlements and movements in the north of Haytí during the first years of colonization (approximate locations based on Sauer [47]).

#### **4. Conclusion: The Transformation of the Indigenous Landscape**

To fully understand the transformation of indigenous landscapes on northern Haytí, the regional archaeological patterns formed by indigenous taskscapes were compared to the regional patterns of movement and settlement of the early Spanish colonizers in the same study region. This comparison led to the idea that the transformation of the indigenous landscape can be explained on two analytical levels. These levels have been conceptualized as *Day-to-day* and *Imaginary*. The first explains the changes that occurred within the indigenous communities in terms of the tasks, practices and movements that characterized the quotidian indigenous taskscape before the arrival of the Spanish. The second level has to do with the representations developed by the first Spanish invaders of the indigenous world and how these representations themselves contributed to the transformation of the indigenous landscape.

#### *4.1. Day-To-Day Level: Contested Taskscapes*

One way to conceptualize the conquest of the indigenous world and its initial transformation is through the idea of contested taskscapes [22]. From the earliest years of colonization, the Spanish invaders depended on the local population to obtain basic products and resources. For example, beginning during the second voyage, Columbus imposed a tribute in gold and food on the indigenous people [41]. In the case of gold, indigenous populations were forced to carry out the task of mineral extraction at known gold veins and rivers, where, in general, the activity followed pre-invasion patterns. However, the intensity of this task radically changed, as did the reasons for the work, with indigenous people preferring to commit suicide and induce abortions to avoid the horrors of enslaved

labor in the gold mines for themselves and their descendants [41,48]. In this sense, although the task remained the same and it was carried out in the same places, the taskscape, the constitutive acts of dwelling, changed radically and consequently a conflict was created.

In the case of food production, something similar happened. Even though the indigenous communities continued to produce food according to their traditional knowledge and practices in the usual places, there was a change in the motivation behind the execution of these same tasks. The Spanish demanded a quota greater than the one they were used to producing, leading to the occurrence of a conflict in the daily tasks that constituted the indigenous agricultural taskscape. An example of this was recorded in the early chronicles, when the cacique Guarionex tried to increased cassava production to pay the Spanish more ever a larger quota and lessen the aggression of the *conquistadores*. However, shortly after these negotiations, Guarionex would go to war with the Columbus brothers since his people had not been able to produce as much food as expected [41,42]. In addition, Guarionex and his people were affected by the Spanish presence near their towns and traditional areas as well as by the fear and rage instilled in them by witnessing what the Spanish were doing to other indigenous groups.

A third contested taskscape can be identified in the Spanish settlements along the island (Figure 7). When Nicolás de Ovando became governor of La Española (1502–1509), other mechanisms to deal with local populations and resources were implemented. By that time, the Spanish had moved to displacing entire communities to Spanish towns to continue fulfilling the tasks of food production and gold exploitation, as well as general servitude. As the people were "inserted" in the Spanish villages, in a process known as *encomiendas* y *repartimientos* [49], the new places where these tasks were to occur were far from where they traditionally took place. An archaeological example that supports the existence of these contested taskscapes comes from the province of Montecristi [26]. During the survey in this area, there was only one place where material culture related to the second half of the 16th-century was recorded, namely in the area where the original foundations of the villa of Montecristi (*circa* 1535) were located [50]. In the rest of the area, no evidence of early 16th-century European material culture was found. In fact, all Spanish or European materials postdate the 17th-century. This could be evidence that during the beginning of the conquest and colonization process (*circa* 1493 to 1502), the Spanish villages did not have rural areas, since their focus was on territorial access and control. After Columbus's second voyage, the Spanish started their attempts to introduce European animals and vegetables to the island environment [51]. As a result of the food crisis in Spain and Portugal and the efforts of Nicolás de Ovando to improve the quality of the colony since 1502, Spanish efforts in cultivation had not fully started [51]. During the first decade of the colonization, the indigenous communities and their towns were the equivalent of the "countryside" for food production. As mentioned, the indigenous agricultural taskscapes were maintained in their traditional places, serving as rural areas for the Spanish, until the populations were displaced to the Spanish villages. This created a key space of contestation, and after several decades, helped to promote the creation of new creole rural communities.

Following this discussion, the idea of the contested taskscape can be connected with two levels of human experience. First, the physical level, which refers to the conflicts that occurred through the materiality of placemaking. For example, the establishment of Spanish forts and settlements close to important indigenous settlements. Second, the cognitive level, which refers to the perceptions of individuals and communities that live in and move between those places where physical conflict was manifest. For example, individuals and communities suffered a strong psychological impact when they become enslaved and had to continue exploiting resources that were known and familiar to them, in places that possibly carried deep cultural meanings. Bender has explained that studies of conflicts and diasporas generally focus on broad social and political scales "without too much consideration of what this might involve in terms of intimate and personal engagement" [23]. In this sense, when day-to-day tasks (the basis of the cultural skills and knowledge of an individual or of a community) were carried out under conditions of force

and enslavery, this affected both the people's perception of those tasks in themselves and their perception of the landscape where those tasks were carried out. This, in turn, would have generated both a personal conflict with the task in question and with the "constitutive acts of dwelling" [17], i.e., with the taskscape. With the development of the colonial state, the stress on indigenous people increased as they were continuously pressured to increase production, to accept new ideologies and religion, and finally to be integrated into the Spanish system as new subjects of a sovereign they did not choose. As Bender [24] has rightly expressed, "the mutual incomprehension engendered by totally different social, political and economic practices extends to include the inability to recognize or at least tolerate a completely different understanding of place and landscape." The indigenous and Spanish taskscapes, at their different spatial scales, naturally came into direct conflict since their basic notions of nature, territory and ownership were radically different.

**Figure 7.** Distributions of early Spanish settlements.

#### *4.2. Level of the Imaginary: Historical Homogeneities and Archaeological Diversities*

The second level concerns how the first Spanish invaders represented the indigenous people. The image created during the first decades of colonization not only transfigured the landscape and the indigenous world in the past but also did so for the future Caribbean populations that came into being with a distorted idea of the island of Haytí before 1492. This issue can be explored from two perspectives. The first concerns the invisibilization of indigenous communities and their presence in the landscape. The early chroniclers, whether due to the difficulty in understanding indigenous cultures or a lack of interest in deep descriptions, presented a vague image of the indigenous people. Based on their understanding of the world, the first Spanish tried to identify in the indigenous communities hierarchical and spatial patterns similar to those that existed in Europe at that time. In spatial terms, this led to the definition of indigenous "territories" [52] that, based on the available archaeological evidence from the northern region of the island [26,27,34,35,53,54], do not seem to reflect the cultural diversity or the political organization before 1492.

An example of this comes from the 1516 map of Andrés de Morales (Figure 8), which divided the island into five regions. While a subsequent report by Peter Martyr d'Anghiera [48] based on Morales' report and map described these as natural regions, the

20th-century economist Bernardo Vega [52] interpreted these regions as political territories related to the indigenous cacicazgos (chiefdoms). However, the map of Morales does not depict borders between these territories and Martyr d'Anghiera simply states that: "The pilot Morales, brings me a new description that from time immemorial the indigenous people used." [48]. The evidence from both the map and the documentary record does not suggest that this description included the delimitation of chiefdoms, as later proposed by Vega [52]. Indeed, the claim that these natural geographical regions which probably were imbued with cultural and historical meanings represented chiefly territories effaced, rather than explained, indigenous diversity. Nevertheless, the wider idea of indigenous territories connected to hierarchical political structures was present in several colonial chronicles and was finally depicted in Charlevoix's map in 1731 (Figure 9).

**Figure 8.** Map of Haytí during the early colonial period by Andres de Morales in 1516 (taken from Frati [55]).

Without doubting the existence of hierarchical societies on the island, it is important to recognize that the representation of homogeneous indigenous territories (i.e., one *cacique* (king) controlling one large territory (nation)) facilitated the explanation of indigenous cultural complexity for initial colonial observers. By contrast, the archaeological data, in combination with the different indigenous taskscapes defined here, reveal a diverse indigenous landscape shaped by multiple spatial scales of community interaction. This patterning is suggestive of the presence of multi-ethnic and politically centralized groups at small scales, rather than large cacicazgos occupying large and homogeneous territories. The first transformation of the indigenous landscapes of Haytí, therefore, happened at this level of Spanish colonial representation. The resulting image of homogeneous territories controlled by highly centralized political polities obscured the island's multiple scales of interaction, the reality of which has been recovered only recently by archaeological research.

A second transformation of the indigenous landscape occurred with the eradication of the diversity of communities by creating homogeneous ethnicities. This point is detailed here through the example of the ethnonym "Taíno". Several authors have highlighted that the early Spanish merged together a wide range of ethnic groups in the island of Haytí since they were unable to recognize internal cultural differences [56]. The Taíno are the classic example of this since they have been identified as the major indigenous group in the Greater Antilles [57,58]. However, Caribbean researchers have realized that, in order to do justice to the complexity of these ethnic groups without obscuring their cultural dynamics, terms such as "Tainoness" [56,59,60] or symbolic reservoir [61] are more appropriate to characterize the indigenous culture of these communities.

**Figure 9.** Map of Haytí showing the supposed indigenous territories by Pierre-Francois-Xavier de Charlevoix in 1731 (courtesy of the digital collection of John Carter Brown Library at Brown University).

The first reference to the term Taíno was made on Columbus's second voyage when upon arriving on the beach of an island he was greeted by people shouting "Taíno, Taíno". The meaning of this word has been identified as "good" or "noble" [60–62]. In the 19th-century, the term Taíno began to be applied to define a particular ethnic group and to their language. The first reference to Taíno as the equivalent of an ethnic group that inhabited the Greater Antilles has been attributed to Rafinesque in 1836 [60], although the term was also used a few years later by Martinus in 1867 [61]. In 1871, Brinton used the term Taíno to describe the linguistic classification of the Arawak language spoken in those islands [62]. The popularization of the term Taíno was a consequence of the historical reconstructions that took place throughout the 19th-century, which condensed the previous generalizations and homogenizations of the ethnic and linguistic diversities of the indigenous groups of the northern Caribbean. For example, in his report about his coexistence with the indigenous people from the northern region of Haytí, Fray Ramón Pané mentioned that he was first sent by Columbus to live in the province of Macorís, where the fort of La Magdalena fort was located, and then to move and live with the cacique Guarionex, because this cacique and his people spoke a language that was understood throughout the island [43]. Still, Pané did not indicate whether Macorís was an ethnic group or that Guarionex and his language were Taíno. The idea that Guarionex was a Taíno cacique first

appeared in the 19th-century historical reconstructions and was further consolidated in archaeological and historical scholarship in the 20th-century. The meaning of Macorís, as a language or ethnic unit, is still strongly debated among specialists.

The criticism raised by recent scholars in their quest to clarify the indigenous patterns before the arrival of Europeans parallels the foregoing critique of colonial cartographic depictions. The representations of indigenous communities made by the early colonizers and missionaries were deeply marked by a European understanding and classification of the world. This "environmental orientalism" [63] was a perspective that originated in the Middle Ages, and shaped how indigenous people were seen. The radically different ways of classifying and understanding the world by indigenous and Spanish groups is perhaps best exemplified in the simplicity with which the Spanish represented the indigenous world and its landscapes. In summary, through their process of interpreting it, the Spanish and other European colonists would transform the diverse, multi-ethnic and pluri-linguistic indigenous landscape into a homogenous "indigenous" landscape characterized by a limited set of homogenous ethnic groups and their territories. This early distorted vision which helped transform the indigenous landscape in the early colonial period, would be reaffirmed by 18th- and 19th-century naturalists, and more recently, further reinforced by archaeological models in the 20th century.

**Funding:** This research has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP7/2007–2013)/ERC grant agreement n◦ 319209, under the direction of Corinne L. Hofman (Leiden University, The Netherlands).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The Montecristi data discussed in this paper is openly available on the EASY DANS repository at https://doi.org/10.17026/dans-xyn-cu72 accessed on 3 March 2022. Detailed information on the archaeological methodologies for data collection and processing is also available on open access [23].

**Acknowledgments:** Many thanks to the communities of the Montecristi province for their collaboration in this research. Thanks in particular to Rami Ramírez, Richard Peña, Eliecer Tati, Natividad Olivo, Abel González, Elvio González, Carlitos Peña, and Jaime Torres. The author also acknowledges the institutions that contributed with permissions, resources and funding: the Museo del Hombre Dominicano, Ministerio de Cultura de la Republica Dominicana, Ministerio de Medio Ambiente y de los Recursos Naturales, Ministerio de Ingeniería y Minas, Museo de Historia de Montecristi, Instituto Montecristeño de Antropología e Historia, European Research Council, and Leiden University. Finally, to Noa Corcoran-Tadd for his valuable comments on the manuscript.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**

