The southwest Gulf Country and the ‘outsiders’

Prior to the arrival of ‘outsiders’, the southwest Gulf country was home to six language groups—Marra, Yanyuwa, Garrwa, Gudanji, Binbingka and Wilangarra. However, frontier violence led to the decline/erasure of the Binbingka and Wilangarra by the early 1800s (Roberts Reference Roberts2005). The Marra, along with the neighbouring Yanyuwa, identify themselves as Saltwater People, that is, people whose lifeworld is embedded in sea country (e.g. Bradley with Yanyuwa Families Reference Bradley2010; Sharp Reference Sharp2002). Marra Country stretches from the southern bank of the Roper River in a southeasterly direction to the northern regions of Rosie Creek, and includes the sea, islands and coastal areas of the Limmen Bight. Marra Country also extends inland up to 60 km from the coast where there are broad plains interspersed with low, rocky sandstone ranges. Today, much of Marra Country, including the locations of the rock art discussed in this paper, forms part of Limmen National Park.

Middle phase: 1870–1900

Pastoral expansion and violence characterized interactions between Marra and outsiders from the early 1870s to 1900. Pastoralists overlanding cattle from Queensland into the Top End and Kimberley created the ‘Gulf Coast Track’ that crossed the traditional lands of all the region's language groups. Gold prospectors on their way to goldfields in Pine Creek and the Kimberley and pastoralist investors keen to secure pastoral leases across the southwest Gulf country meant a further influx of Europeans (Fig. 2). It was during this time that Marra and neighbouring language groups became exposed to Europeans in much greater numbers along with extraordinary amounts of cattle and horses.

Figure 2. The approximate route of the ‘Gulf Coast Track’. (Map adapted from Roberts Reference Roberts2005.)

Roberts (Reference Roberts2005; Reference Roberts2009) provides the most detailed record of the atrocities pastoralists committed during this time, namely the brutal massacres of Marra and other language groups which ultimately led to the demise of the Binbingka and Wilangarra. This era is referred to by Aboriginal people in the area as ‘Wild Times’. Reports of Aboriginal people killing cattle and horses and carrying out stealth attacks on Europeans reveal their heavy presence in the landscape, intense resistance to the newcomers and their increasing exposure to, and familiarity with, introduced animals (cattle, horses, sheep) and European objects such as guns, pottery, glass, wagons and smoking pipes. With around 150,000–200,000 head of cattle and 10,000 horses travelling on the Gulf Coast Track in the 1870s and ’80s (Roberts Reference Roberts2005, 67; Webber Reference Webber, Leather and Webber2015, 27), the Marra landscape and environment was dramatically transformed, with waterholes fouled and fences constructed.

By the late 1880s/early 1890s, pastoral stations were established across the southwest Gulf country. In 1884, John Costello established his Valley of Springs station near the Limmen River, occupying the majority of Marra Country, stretching from the McArthur River near Borroloola westwards to the Roper River. Dickson (Reference Dickson2015, 56) uses data from Costello's son's (Reference Costello1930) memoir to suggest that ‘at least a hundred—perhaps several hundred—people were living on Marra Country when Valley of Springs was established’. Archival narratives reveal details about the Marra's mobility and economic strategies during the life of the station: in the wet season, Marra hunters would lead cattle to wet, boggy, muddy grounds where the animals became stuck and Europeans were unable to free them, thus leading to easy hunting; and in the dry season, hunters would be stealthy and spear animals close to the rugged ranges where they could butcher and collect the meat before disappearing back into the ranges (Costello Reference Costello1930, 131; Roberts Reference Roberts2005, 166). By 1892–3, Valley of Springs was abandoned due to the Marra's widespread cattle-spearing, along with other factors (bushfire, crocodiles, tick fever) (Roberts Reference Roberts2005, 167). ‘Dick’ was the only Marra man recorded to have been working for Costello. Costello described Dick as living on the station, learning some English, how to ride a horse, and becoming ‘serviceable in stock mustering’ (Costello Reference Costello1930, 135). During a survey of the property, Dick eventually abandoned Costello after going to find water among his ‘own people’, presumably Marra (Costello Reference Costello1930, 148).

Hybrid economies and culture contact

Altman's (Reference Altman, Taylor, Ward, Henderson, Davis and Wallis2005; Reference Altman2006) hybrid economy framework was designed to increase attention to the often overlooked contributions that Indigenous people make to developing sustainable economies. To do this, he developed his framework as an ‘analytical construct for the assessment of the particularities of any one situation and the linkages between the market, the state and the customary components of the economy’ (Altman Reference Altman, Taylor, Ward, Henderson, Davis and Wallis2005, 36). This is essentially the interaction of contemporary Aboriginal communities with state and market economies (see also Keen Reference Keen2010). Three interrelated sectors are identified as integral to a hybrid economy:

• • Market (e.g. retail, arts industry, wildlife harvesting, mining, tourism, pastoralism)

• • State (service provider, law enforcement)

• • Customary Indigenous based on cultural continuity (e.g. hunting, gathering, fishing, habitat management, burning)

It is the linkages and interdependencies between these three sectors that he argues are essential to understanding challenges for developing sustainable hybrid economies in rural and remote Indigenous communities (for a recent critique, see e.g. Curchin Reference Curchin2013).

While Altman's framework was designed to deal with contemporary economic development problems, archaeologists have used it to explore linkages between market, state and Indigenous customary economies in historical or cultural contact settings. Market economies included interaction with Macassan trepangers and European economies such as pastoralism, mining (e.g. gold, opals), buffalo hunting, and pearling. State forces included the state (South Australia) and Commonwealth government that resulted in control and ownership of Indigenous lands, government settlements and the building of the Overland Telegraph Line, while Indigenous customary forces, such as subsistence and other non-monetary, non-market, subsistence and other informal economies, continued during this time.

In using Altman's model, we acknowledge the issues surrounding use of the term ‘hybrid’ in characterizing the complexity of transformations that occur as a result of Indigenous–colonial encounters. ‘Hybrid’ and ‘hybridity’ have often been associated with issues around race and ethnicity and imply a degree of stasis in these relationships (e.g. Kapchan & Strong Reference Kapchan and Strong1999; see also Robinson Reference Robinson2013). However, our use of ‘hybrid’ reflects the nomenclature used by Altman and refers to the intertwined economic systems that were operating in the southwest Gulf of Carpentaria at the time, rather than suggesting Indigenous traditions became hybridized as a result of this interaction.

Key examples of a hybrid economy approach come from western Arnhem Land. For example, Wesley and Litster (Reference Wesley and Litster2015) describe how glass beads recovered in excavations were introduced by Macassans sometime after 1800 ad. Aboriginal people bestowed them with traditional significance—they became highly sought-after exchange items, used to negotiate access for Macassans to trepang beds and in exchange for labour in different forms of colonial economies (Wesley & Litster Reference Wesley and Litster2015, 12). A spinoff from this exchange network was that they were also introduced into existing rock-art traditions as a distinctive design element on anthropomorphs (Wesley & Litster Reference Wesley and Litster2015, 12). Wesley (Reference Wesley2013) also describes how firearm (rifle) motifs reflected changes in Aboriginal society, especially Aboriginal engagement with the buffalo-hunting industry in western Arnhem Land (1870s–1920). During this time, Aboriginal people became involved in a ‘well-developed hybrid economy between white Australian shooters and Indigenous families’ (Wesley Reference Wesley2013, 245) whereby Aboriginal people were in possession of different forms of rifles and built an intimate and detailed knowledge of them that resulted in a significant number of firearms used in this economy being depicted in rock art. However, changes by the state from the 1920s onwards that prohibited Aboriginal people from owning firearms meant their involvement in the buffalo-hunting economy declined altogether, thus explaining the lack of post-1920s firearm motifs in Australia.

Using these studies as a platform, we explore how the hybrid economy model can be used to understand how the Marra interacted with the outside market and state forces while continuing to negotiate their own customary cultural economies and traditions, namely the production of introduced subject matter into extant graphic (rock art) systems.

Introduced subject matter in Marra Country rock art

Prior to 2017, there have been very few recordings of motifs depicting introduced subject matter in the southwest Gulf country. In Yanyuwa Country, senior Yanyuwa woman Dinah Norman recalled seeing a painting of a yirrikirri [donkey] at a rock-shelter on South West Island in the 1940s, but this has since disappeared (see Brady et al. Reference Brady, Bradley and Kearney2016). In 1999–2001, Robin Sim (Reference Sim2002) identified four motifs from Yanyuwa Country she thought depicted introduced subject matter (European and Macassan): three motifs (‘metal shears’, a ‘red ochre boat’, a possible ‘prau’) have not been verified by us, and a fourth motif (another possible ‘prau’) was interpreted by senior Yanyuwa men to John Bradley in 1984 as bark canoes. In Garrwa Country, an elaborate painting of a paddle-steamer, the Young Australia that sank in the Roper River in 1872, is located at Jininyina (Spring Creek), while in Marra Country, introduced subject matter was first encountered and recorded in 2015–16 as part of the first cultural heritage surveys at Limmen National Park, with key motifs photographed including all but one introduced subject matter motif.

Rock-art research in Marra Country began in 2017 as part of a collaborative project involving the authors, Marra Families and the Parks & Wildlife Commission of the Northern Territory (PWCNT). Although still in its infancy, research thus far has revealed, among other things, the world's largest-known assemblage of miniature and small-scale stencil motifs, an increase in the geographic distribution of beeswax motifs and insights into Indigenous self-determined conceptions of rock art in relation to kinship, ontology and epistemology (e.g., Brady et al. Reference Brady, Bradley and Wesley2019; Reference Brady, Bradley, Kearney and Wesley2020; Kearney et al. Reference Kearney, Bradley and Brady2021). To date, 56 sites and over 3000 motifs from across Marra Country have been recorded.

A total of 25 motifs thought to depict European introduced subject matter have been recorded from seven sites inside Limmen National Park (Table 1). An additional motif, a dugout canoe, can also be characterized as introduced, given they were obtained from Macassans at the end of each trepanging season. Prior to this, bark canoes were used for sea- and river-going travel (Kearney & Bradley Reference Kearney and Bradley2015). However, as the dugout canoe cannot be attributed to a European origin it is excluded from further analysis. The majority of motifs are located at Matangula (n = 13), followed by Nalangkulurru 2 (n = 7). Motifs are located in rock-shelters of varying dimensions and placement in the landscape (flat plains, mid- or upper-slope). All motifs are pigment-based (paintings n = 18 and stencils n = 7), have been produced on the rear walls or roof of the rock-shelter and are monochrome (n = 12) or bichrome (n = 13), with three colours documented: red, white, yellow (see Table 1). Motifs are depicted in varying degrees of detail: simple or plain and lacking decoration; hybrid motifs where there is a transitioning of features such as traditional and introduced subject matter; and highly detailed motifs reflecting an intimate knowledge of the object, animal, or person depicted.

Table 1. Classification of introduced subject matter from Marra Country, Limmen National Park.

Motifs were classified in a two-level hierarchical format to explore variation in the assemblage (Table 1). Four Group Motif Forms were identified—Material Culture, Anthropomorph, Zoomorph, Anthropomorph + Zoomorph. Material Culture dominated the assemblage (n = 14), followed by Zoomorphs (n = 7), Anthropomorph + Zoomorph (n = 3) and one Anthropomorph. Classification was further refined according to a Specific Motif Form (e.g. Firearm, Horse) with Metal Axe Heads the most frequently depicted introduced subject matter. In three instances (Smoking Pipe, Bandy-legged Anthropomorph and Horse + Rider at Matangula 5), traditional imagery has been depicted overlying the introduced subject matter, indicating painting activity was continuing after the European-themed motifs were created.

Since these are the first recordings of introduced subject matter in the region, a systematic framework for identifying and categorizing motifs was developed, based on the absence/presence of specific characteristics or attributes reflective of European influence (Table 2). Often the differences between introduced subject matter can be subtle, such as the morphological difference between a horse and donkey—a distinction that can have implications for examining the possible context(s) in which a motif is produced (e.g. Frederick Reference Frederick1999; Taçon et al. Reference Taçon, May, Fallon, Travers, Wesley and Lamilami2010). In our project, attributes are derived from various resources and provide descriptions and illustrations of key features such as the names of horse accoutrements (e.g., bridle, throat latch), all of which can be used to create a highly accurate motif identification. Whereas other presence/absence attribute frameworks have targeted only specific motif types such as watercraft and firearms (e.g. Wesley Reference Wesley2013; Wesley et al. Reference Wesley, McKinnon and Raupp2012), our analysis covers a much wider range of motif types, thus adding additional attributes to studies targeting accurate identification of introduced subject matter.

Table 2. List of specific characteristics or attributes for introduced subject matter from Marra Country, Limmen National Park.

The majority of motifs correspond with the key attributes identified for each object, zoomorph, or anthropomorph + zoomorph, and in several instances allow for a more specific identification. Upon closer examination, knife motifs have been specifically identified as butchering knives, and firearms as common rifles (carbines) found in the pastoral industry.

Some notable features not included in the attributes-based description of the assemblage are:

• • Dot outlines: a characteristic feature of non-introduced subject matter motifs from the southwest Gulf country consists of both white dot outlines (typically found around infill motifs) and motifs created entirely using dots (Brady & Bradley Reference Brady and Bradley2014). However, the former has been recorded on the Bandy-legged anthropomoroph at Matangula 5 (Fig. 3), hybrid Horse (Matangula 7) (Fig. 4) and the Leading Pack Horse Team motif (Fig. 5), indicating continuity in the use of a region-specific design convention into introduced subject matter

• • Writing: at Matangula 5, the Horse + Rider motif features a rare instance of writing on a horse (Fig. 6). The letters ‘JOE’ are vertically written twice—immediately in front of and behind the rider. Two other letters, ‘OE’, are written on the horse's neck. The only reference we have been able to find relating to ‘JOE’ in a horse-and-rider context is from Costello's son's memoir. Costello describes how, upon abandoning Valley of Springs Station in the 1890s, his father travelled to his other property, Lake Nash on the Barkly Tablelands. At Lake Nash, he notes the role of his father's horse: ‘[m]ounted on his wonderful blood horse, “Joe,” the pair were a unit of enthusiasm for their work’ (Costello Reference Costello1930, 197). Whether this reference refers to the painting is difficult to ascertain. Whatever the case, the highly detailed depiction of the horse-and-rider motif indicates the artist was very familiar with horses, probably through riding or other close-up encounters with horses.

• • Hybrids: at Matangula 7, a possible hybrid motif features characteristics of a dingo and a horse (see Figure 4). We have chosen to identify it as a horse because 1) the animal has hooves, not paws like a dingo would; 2) pointed ears rather than rounded; and 3) white dots on its head are in the correct place for a bridle and on the body for a saddle, and the hump on the back could be a saddle or saddle bags. The upward-curved, dingo-like tail is somewhat problematic. It is possible the artist might have painted a dingo first (with a curved tail), and then they (or someone else) returned later and decided to change it to a horse, or the artist drew the horse tail flicking upwards (see e.g. Taçon et al. Reference Taçon, Ross, Paterson, May, McDonald and Veth2012, 424).

Figure 3. Matangula 5: motif depicting a man smoking a pipe and outlined in white dots; two cows are located on the left and right sides of the panel.

Figure 4. Matangula 7: horse motif outlined with white dots.

Figure 5. Matangula 5: panel showing an anthropomorph leading a team of pack horses with the horses outlined with white dots, and a cow in the bottom left of the panel.

Figure 6. Matangula 5: detailed depiction of a man riding a horse; note the words ‘JOE’ painted on the horse's body, and ‘OE’ on the horse's neck.

What is of interest with this assemblage is the quantity and types of motifs that can potentially be linked to the pastoral industry. Based on our analysis, at least half of the assemblage has direct links to the industry. While others could tangentially be linked to pastoralism (axes and pipes), it is difficult to link them because of a lack of a direct relationship with pastoralism (Fig. 7). Therefore, of principal interest are zoomorphs, anthropomorphs, firearms and knives.

Figure 7. Introduced subject matter from Marra Country not directly related to the pastoral industry (clockwise from upper left): horseshoe-shape, firearm, firearm, metal axes, smoking pipe.

Zoomorphs: Three classes of zoomorphs are directly related to the pastoral industry: 1) donkeys were involved in droving activities of pastoralists and were used primarily as pack animals if horses were unavailable. Their depictions are characterized by straight, narrow bodies, short tails, ears (long, large, tapering upwards to a point, vertically positioned in relation to the head) and small box-like shaped hooves (Fig. 8); 2) cows were the core of the pastoral industry and Aboriginal stockworkers would have had extensive involvement with them. Their depictions are characterized by rectangular-shaped bodies, straight backs, short faces, thick necks, long thin tails and short, slightly upward-curving horns (see Figs 3 & 5); and 3) horses, also at the centre of pastoral activities through droving. Stockworkers would either have been riding them or caring for them and thus intimately familiar with their physical appearance. They are characterized by key features such as long slender legs, elongated head, oval-shaped hooves, bridle (nose band, brow band, head stall), body harness (neck or collar strap, trace), slightly rounded body, reins, girth strap and small ears (short in relation to head) and tapering upwards to a point (see Figs 5 & 6).

Figure 8. Two donkey motifs at Matangula 5.

Anthropomorphs: the only anthropomorphs directly related to the pastoral industry are those either seated on horses, or leading cattle or pack-horses. The latter are characterized by Stetson-style hats (see Figs 5, 6 & 9).

Figure 9. Man on horse positioned above a cow motif (Tawallah 2).

Three motifs can be identified as firearms; however, only one can be reliably identified to a specific type. The Matangula 7 firearm motif has the characteristics of the large-calibre Snider Enfield carbine with key identifying characteristics including the hammer and the unique firing pin situated in a nipple on the breech block cylinder (Figs 10, 11). The Snider Enfield carbine and rifle were popular among early settlers, miners, buffalo shooters and pastoralists in the Northern Territory during the late nineteenth century (Wesley Reference Wesley2013). This rifle was known to be distributed to Aboriginal men working in the buffalo shooting industry (May et al. Reference May, Wesley, Goldhahn, Litster and Manera2017; Wesley Reference Wesley2013). The two other firearm motifs only have generic features present (firearm furniture [stock], barrel bands), and not enough specific attributes to identify these reliably to a particular make and model.

Figure 10. Butchering knives and Snider Enfield rifle at Matangula 7. (Left) original; (right) digitally enhanced.

Figure 11. (Above) Advertisement for butchering and skinning knives. (Montgomery Ward & Co. Catalogue and Buyers Guide No. 93, 1920. https://archive.org/details/MontgomeryWardCat1920tools); (below) Snider Enfield rifle. (Australian War Memorial REL/10253.)

The same panel at Matangula 7 with the Snider Enfield carbine motif includes three motifs classified as knives owing to their distinctive shape with a blade and handle (Figs 10, 11). These motifs are interpreted as butchering knives typically used in the late nineteenth and early twentieth centuries (Horowitz Reference Horowitz2006, 35). Two of the motifs are similar to butcher/boning knives while the third has the characteristic shape of a skinning knife (Horowitz Reference Horowitz2006, 35). Specialized roles that Aboriginal stockmen carried out on pastoral stations included butchering livestock (McGrath Reference McGrath and Dodson1997, 12). Metal knives have been found at Aboriginal campsites associated with early twentieth-century pastoral stations (Smith Reference Smith2001, 28).