Introduction

A number of developments have occurred recently regarding the origins of farming, or agriculture. Taken together, these developments underscore the fact that despite more than a century of research and theoretical debate, the reason(s) why farming first began remain unclear (Bellwood, 2017; Carey, 2023; Cohen, 2009; Clutton-Brock, 2013; Gerritsen, 2010; Hodder, 2018; Price & Bar-Yosef, 2011). Basic assumptions have been brought into focus, foremost among which is that farming differs fundamentally from hunting and gathering. Another assumption is that farming was always predicated on the food economy. Both these core assumptions about early farming warrant interrogation and are now open to debate.

Definitions

Farming encompasses plant cultivation and/or the husbandry of animals—in each case, with or without domestication, and without presuming any purpose or product. At the outset, early farming would have entailed the cultivation of wild plant species and/or the husbandry of non-domesticated animal species. In some instances, the plant or animal species involved in farming were never domesticated—palm trees cultivated for fibres, for instance, or tame elephants bred in captivity to be used for transport. The term farming is synonymous with agriculture although, in popular usage, agriculture tends to emphasize plant cultivation over animal husbandry. The linguistic origins of agriculture refer to fields (as in agrarian), whereas farming has origins in the management or exploitation of resources more generally (as in the farming or collection of taxes). Farming is favoured over agriculture here, to avoid any bias towards plant cultivation. The Concise Oxford English Dictionary defines agriculture as farming and, alluding to the broad scope of farming, cites the example of wind farms (Stevenson & Waite, 2011). Similarly, the term foraging can be used interchangeably with hunting and gathering, although both these terms tend to explicitly denote the procurement of food rather than other resources (ibid.; Kipfer, 2022).

Debating the Difference

The first core assumption—that farming represents a departure from foraging lifestyles—has two interrelated facets. First, the dichotomy between foraging and farming may be illusory, fabricated on a dualistic perspective that is not without historical, philosophical, and political issues (Abbo & Gopher, 2022; Bogaard et al., 2021; Gerritsen, 2010; O’Brien, 2019; White, 2011). If the conceptual dichotomy between foraging and farming is illusory, with both being no more than upstream and downstream nodes in a continuously graded series of human activities related to purposeful management of plant and animal resources, then farming may have no discrete or discernible origin(s). As such, any effort to explain the emergence of farming in terms of general hypotheses is misguided and doomed to failure. Rather than seeking a single cause—or, for that matter, identifying a universally applicable set of causes—the developments in each region should be examined quite independently, with different factors probably operating in different places and times (Denham, 2007; Fuller et al., 2022). Second, a key challenge for any attempt at explaining the emergence of farming is the apparent failure of farming to develop in some parts of the world, notably on the Australian continent prior to the colonial era (Gilligan, 2010a). A radical solution to this conundrum is to reinterpret traditional foraging as no different in principle to farming, citing examples of resource management and landscape custodianship that were overlooked or downplayed due to a perceptual bias and Western hegemony (Adeleye et al., 2023; Gammage, 2012; Gerritsen, 2014; Pascoe, 2014). Regardless of whether or not this reappraisal is entirely consistent with available evidence (Bennetts, 2021; Keen, 2021; McDonald, 2021; Porr & Vivian-Williams, 2021; Sutton & Walshe, 2021; Veth, 2021), it is clear that conventional approaches and assumptions need to be reevaluated.

Querying the Food Assumption

The presumed priority of the food economy—and even defining agriculture as food production—has come under critical scrutiny from two perspectives. First is the work of Brian Hayden, who points out the many flaws in assuming that early farming offered hunter-gatherers a more secure or abundant food supply for their everyday subsistence. Instead, Hayden proposes the feasting hypothesis, whereby farming supplies surplus resources for social purposes of ceremony and prestige, and the farming products are not necessarily restricted to food—the products can include textiles to make elaborate clothes, for instance (Hayden, 2009). The second proposal is Ian Gilligan’s hypothesis that textile fibres were the primary resource involved (Gilligan, 2007, 2019), as people shifted from wearing animal hides and furs to using woven cloth for garments in warmer post-glacial environments.

Early Farming Dispersal Hypothesis

Another question is whether farming first developed in a few regions (farming centres or homelands) and spread subsequently into surrounding areas (Diamond & Bellwood, 2003), or whether humans were always in a state of dynamic interdependent flux with plant and animal resources, which would suggest a more variable picture for early farming. Related to this question of discrete versus diffuse origins is how farming spread around the world: did the dispersal of farming happen mainly through the adoption of farming practices by local foragers (either de novo, or by acquiring ideas and innovations from neighbouring farmers), or was farming instead carried into new areas primarily by the demographic expansion of farming populations? Data bearing on this question are not limited to archaeology but include findings from other disciplines, particularly linguistics and population genetics (Bellwood & Renfrew, 2002; Heggarty & Beresford-Jones, 2014; Hudson & Robbeets, 2020). Collectively, the findings from relevant disciplines favour demographic expansion from a limited number of farming homelands, providing support for Peter Bellwood’s hypothesis of early farming dispersals (Bellwood, 2022, 2023).

How Many Homelands?

Independent homelands of early farming need to be distinguished from centres of domestication, as the processes responsible for initial transitions may be quite different to the reasons why farming practices spread into neighbouring regions. Also, new plant and animal species were often domesticated locally as farming practices dispersed but the domestication of new species does not necessarily qualify a region as a farming homeland. Bellwood (2023:42) lists seven regions where a significant dependence on ‘food production’ appears to have developed without external influence, or diffusion, although at least one of the regions—African Sahel and Sudan—may not be entirely independent owing to an earlier arrival of Fertile Crescent domesticated animals (cattle, sheep, and goats), even though major local crops were first domesticated in the region (ibid.:113). Another domestication centre promoted as an independent farming homeland is Amazonia (Lombardo et al., 2020; Piperno, 2011; Watling et al., 2018), although among the earliest Amazonian crops is maize, which arrived from Mexico (Iriate et al., 2020). As Bellwood emphasizes, it is easy to underestimate the extent of prehistoric cultural contacts and foreign influences, or diffusion (leaving aside population expansion and migrations)—influences which may leave little trace in terms of archaeological, genetic, or linguistic evidence (Bellwood, 2023:43). For instance, cotton from the Indian subcontinent has been recovered at a site in the Jordan Valley, Israel, dated to 7,200 years ago (Liu et al., 2022). Similarly, while new crops (like sunflower) were first cultivated in the Eastern Woodlands of North America (Smith, 2006, 2009), the subsequent arrival of maize from Mesoamerica (Smith, 2017) might raise a question as to the claimed status of this region as an independent farming homeland. Likewise for the squash bee, which pollinates domesticated squash (the first crop to be cultivated in the Eastern Woodlands): the squash bee migrated from Mesoamerica (López-Uribe et al., 2016; Pope et al., 2023).

The number of domestication centres (regions where new species were domesticated) is quite large, and growing—at least 15 (Zeder, 2017) and perhaps 24 (Fuller & Denham, 2021)—but there could be as few as five indisputably independent farming homelands. In each of these five early farming homelands, along with food crops grown for human consumption, some major crops fed farm animals or, like flax, cotton, and banana, supplied textile fibres (Figure 1, Table 1).

Figure 1
figure 1

World map showing early farming homelands and fibre resources

Full size image
Table 1 Presence of early to mid-Holocene domesticated fibre resources in the five indisputably independent farming homelands
Full size table

Fundamental Questions

Any theoretical approach to farming origins must address a number of fundamental issues. One is the absence of farming, not merely in certain regions (to the extent that farming was absent in places like Australia) but, more important, the complete absence of discernible farming anywhere in the world before the end of the last glacial cycle 11,700 years ago—a key point in Colin Renfrew’s ‘sapient paradox’ (Renfrew, 2012). Another fundamental issue is the well-attested security, flexibility, and efficiency of traditional hunter-gatherer lifestyles (Lee, 1968; Sahlins, 1968; Unaipon, 2001), versus the equally well-attested drawbacks—in terms of dietary breadth, risks, and workload—of early farming practices (Cohen, 2009; Larsen et al., 2019; Mummert et al., 2011). Furthermore, there remain unresolved issues about the role played by global climate changes in promoting early farming, as well as the choices of early plant and animal domesticates, together with the continued reliance on wild food resources in farming homelands for millennia after the advent of farming and—last but not necessarily least—profound discrepancies between the cultural values associated with foraging and farming, discrepancies that would seem to discourage if not preclude a de novo transition from foraging to farming based solely on comestible subsistence considerations (Barker & Janowski, 2011; Naveh & Bird-David, 2014; Sutton & Walshe, 2021).

The Textile Hypothesis

Attested by a body of archaeological evidence which includes the advent of eyed needles in mid-latitude Eurasia nearly 50,000 years ago (d’Errico et al., 2018), fitted—or ‘tailored’—garments were developed among some human populations for added insulation against cold temperatures and wind chill during the latter part of the last glacial cycle (Gilligan, 2010b). By the beginning of the Holocene epoch—Marine Isotope Stage (MIS) 1—the human body had been covered routinely for millennia, and decoration of the body surface was transferred onto clothes. Social functions for clothing—including a new sense of modesty about exposing the naked body (Gilligan, 2023)—meant that among the descendants of those populations, people were inclined to continue wearing clothes in warmer post-glacial climates.

Wearing clothes in warm conditions presents a physiological challenge due to increased sweating (Chan et al., 2016; Choudhury et al., 2011; Guan et al., 2019; Mandal et al., 2022; Mukhopadhyay & Midha, 2016; Rengasamy, 2011; Tang et al., 2015). The physiological stress of wearing clothes in warm weather is exacerbated with high relative humidity levels (Maughan et al, 2012; Sobolewski et al, 2021). This combination of environmental conditions—higher air temperatures and humidity—pertained generally throughout the early Holocene, with regional variations (Arthur et al., 2023; Cheddadi et al., 2021; deMenocal et al., 2000; Gholamreza et al, 2022; Ilvonen et al., 2022; Katsuta et al., 2017; Morellón et al., 2018; Neugebauer et al, 2022; Palmer et al., 2023; Roberts, 2023; Roy et al., 2014; Shanahan et al., 2015; Thompson et al., 2022; Yahiaoui et al., 2023; Zhao et al., 2017). The solution to this moisture management problem with clothes was a shift in clothing materials, from the relatively non-porous hides and furs that were appropriate in drier Pleistocene climates to the use of woven textile fabrics (Gilligan, 2019).

Fibre Production and Farming

Textile technologies were utilized in the Palaeolithic to manufacture a range of perishable artefacts that likely included strings, ropes, nets, mats, and baskets (Hardy, 2008; Soffer, 2004; Xhauflair et al., 2023). Flax, for instance, has been exploited as a fibre resource since at least 30,000 years ago (Kvavadze et al., 2009). However, demand for natural fibre resources increased dramatically with the transition to textile clothing. Flax and cotton were among the prime fibre resources extracted from plants, while wool and wool-like fibres were obtainable from certain animal species, primarily sheep and goats in Southwest Asia and camelids in South America (Dransart, 2002; Montt et al., 2023), along with silk from silkworms in China (Gong et al., 2016; Shelach-Lavi, 2015). Besides an emphasis on quantity favouring the active management of these natural fibre resources for cloth production, two additional aspects are noteworthy: renewable fibre production from the living animals, and the feeding of these animals.

Animals as Living Fibre Factories

As emphasized by Ingold (1984) and Hayden (2009), the taming, feeding, and herding (or husbandry) of animals differs from hunting in one crucial respect: it implies the animals in question became more valuable to humans as living assets rather than as dead carcases. Whereas accessing meat requires killing animals, certain purposes may favour the protection and care of living animals. In addition to acquiring animals as pets or social assets and for breeding purposes, valuable resources can be produced on a renewable basis while some animals are kept alive—namely wool, milk, and the use of animals for transport and traction. However, milk was unlikely to figure prominently in the early phases of animal domestication, due to lactose intolerance in adult humans (Segurel et al., 2020; Stock & Wells, 2023). Wool, on the other hand, was produced in copious quantities by the wild ancestors of sheep, llamas, alpacas, and goats (mohair and cashmere). In other words, ongoing fibre production rendered these animals more valuable alive than dead.

Wool for textile clothing is recognized as playing a significant role in mid-late Holocene socioeconomic transitions (Renfrew, 1972; Sabatini et al., 2019; Schier & Pollock, 2020). A classic interpretation is the secondary products revolution (Sherratt, 1981), where the main secondary products are milk, wool, and traction. While milk and traction are clearly secondary factors in terms of early animal domestication, the situation with sheep and wool is quite different. The main reason for discounting wool as a primary product during the early neolithic is the mistaken belief that wild sheep ‘were unlikely to be woolly in this period’ (Marciniak, 2011:121), and hence sheep could not have been domesticated for wool:

wild sheep lack a woolly outer coat… Hence sheep could not have been domesticated for their wool (Greenfield, 2010:35);

It is accepted that this wild ancestor was not domesticated for wool, for wild sheep have none (Breniquet, 2014:59).

The wild ancestors of domestic sheep produced wool every year in a similar manner to contemporary primitive breeds such as the Soay, which shed their wool in the annual spring moult (Chessa et al., 2009)—the wool tends to fall off their backs, or the wool can be plucked from tame sheep. The dismissal of wool as a potential primary product of husbanded wild sheep arises from a confusion between seasonal wool growth and the evolution of a permanent fleece—with the latter itself indicative of artificial selection for wool (Ryder, 1969, 2005).

Evidence for Early Textiles

Preservation bias with textiles is likely compounded by a theoretical bias—including a gender bias—that privileges the detection (and reporting) of food over fibre products in early farming contexts (Adovasio et al., 2001, 2007, 2014; Soffer, 2004; Xhauflair et al., 2023). No woven cloth has survived from the Pleistocene, and fabric finds dating to the early Holocene have mainly been recovered in either water-logged or dry contexts. Most archaeological textiles are limited to the mid-late Holocene (Margariti et al., 2023). Among the rare early Holocene examples of textile preservation are from desert sites in Peru and Southwest Asia (Bar-Yosef, 1985; Hodder, 2013; Jolie et al., 2011; Rast-Eicher et al., 2021). The oldest woven garment is the 5,000-year-old Tarkhan dress from Ancient Egypt (Stevenson & Dee, 2016). Preserved cloth fragments dated to 7,500 years ago were recovered from the water-logged Windover Bog site in Florida, woven from palm fronds (Adovasio et al., 2001). At the submerged neolithic site of La Marmotta in Lake Bracciano near Rome (dating to 7,700 years ago), remarkable finds include cloth fragments (identified tentatively as linen), spindles with fibres wrapped around them (Figure 2), spindle whorls, loom weights, weaving swords, awls, and seeds from safflower-like plants that might have been used to dye fabrics (Mineo et al., 2023).

Figure 2
figure 2

Textile finds dating to 7,700 years ago from the submerged neolithic site of La Marmotta, Italy: a cloth fragments, identified tentatively as woven with flax; b spindles with fibres wrapped around them (Mineo et al., 2023:320, 323). Reproduced under Creative Commons BY 4.0 license

Full size image

In the homelands of early farming, fibre products include wool and flax in Southwest Asia (Abbo et al., 2015; Arranz-Otaegui & Roe, 2023; Bruford & Townsend, 2006; Saña & Tornero, 2012; Zohary et al., 2012), hemp, ramie, jute, and silk in China (Buckley, 2017; Gong et al., 2016; Liao & Yang, 2016; Liu & Chen, 2012), maguey—sometimes called sisal hemp—and cotton in Mesoamerica (Coppens d'Eeckenbrugge & Lacape, 2014; Scheffler et al., 2012), wool and cotton in Peru (Clutton-Brock, 2013; Dillehay et al., 2017), possibly palm fibre in Amazonia (Iriate et al., 2020; Pennas et al., 2019), possibly Indian hemp and milkweed in eastern North America (Applegate, 2008; Borders & Lee-Mäder, 2014; Claassen, 2011; Heiser, 2003; Thompson & Simon, 2008), cotton on the Indian subcontinent (Moulherat et al, 2002; Rast-Eicher, 2016; Zohary et al., 2012), flax, wool, cotton, and enset (an Ethiopian variety of banana) in Africa (Kriger, 2009; Horsburgh & Rhines, 2010; Magee, 2014; Magnavita, 2008), and banana in Papua New Guinea (Kennedy, 2009; Denham, 2018).

With regard to Papua New Guinea, ethnographic evidence shows that clothing was used throughout Melanesia (Herdt, 2006; Lewis-Harris, 2010; Mel, 2010). Cold conditions in the New Guinea highlands during the late Pleistocene (Prentice et al., 2005) encouraged the use of clothing, and high humidity levels favoured textiles from the outset. Banana fibre is highly absorbent of moisture (Batra, 2007), and banana was an ideal target for cultivation as a fibre crop in this tropical early farming homeland. Clothing was also widespread throughout the Americas—complex clothes were necessary for humans to occupy northwestern Siberia and enter the New World during the late Pleistocene (Bellwood, 2022; Hoffecker, 2017). Even in tropical Amazonia, minimal clothing was still used in recent times (sometimes merely a penis string), and a modicum of modesty was present (Chagnon, 2013; Kohn, 2013; Lizot, 1985).

Crops for Animal Feed and Fodder

Along with secondary foods such as peas, lentils, and sunflower, most of the world’s major food crops—wheat, rice, barley, oats, rye, millets and maize (corn)—are very useful as feed for animals. Moreover, as Hayden (2011) points out, the wild precursors of cereal crops—like rice—represented a poor food option for humans and yet, in their wild forms, these resources were well-suited for feeding domestic animals like pigs, ducks, and turkeys. Moreover, cultivated plant products could be stored as winter fodder, while an enhancement of seed production facilitated the growing of sufficient quantities to provision domesticated animals—especially grazing herbivores like sheep—on a year-round basis.

Farming and Food Production

The textile hypothesis does not deny the obvious role of food production in early farming. Nonetheless, the textile hypothesis insists that food alone—that is, food for human consumption—would never warrant a de novo transition from foraging to farming. Rather, fibre production for textile clothing tipped the balance in favour of farming. Once commenced, farming activities extended to food production for coevolutionary reasons, and humans became more dependent on food production (Angourakis et al., 2022; Zeder, 2017)—though this reliance on food production happened at varying rates in different contexts. While food production figured quite prominently from the outset in Southwest Asia, wild food resources continued to supply a substantial proportion—often the bulk—of the human diet in most farming homelands (Fuller et al, 2018; Stevens et al., 2022), sometimes for four to five millennia after farming was established—in Mesoamerica and in northwest South America, for example (Kennett, 2020; Quilter, 2022). The situation with rice in China is another case in point (Hayden, 2011; Wang, 2023). Cultivated rice contributed relatively little to the human diet for millennia, and it was not until mid-way through the Holocene that a greater dependence on rice farming was established, in the context of rapidly expanding human populations.

Early Weaning and Population Growth

A key aspect of most independent early farming developments was human population growth, promoting further investment in food production. The cause of this demographic transition has long been debated (Bocquet-Appel, 2011; Bocquet-Appel & Bar-Yosef, 2008; Downey et al., 2014; Porčić et al., 2021), since foraging communities typically maintain stable population levels. Furthermore, early farming experiments were generally accompanied by malnutrition and increased mortality rates, due to recurring famines and epidemic diseases in the settlements (Allaby et al., 2022; Larsen et al., 2019). Without a higher survival rate for infants, an increase in the birth rate—greater fertility—was apparently a prerequisite for population levels to increase.

In foraging communities, birth intervals are typically four to six years, maintained by prolonged and intense breastfeeding of infants which acts as an effective natural contraceptive (Gray et al., 1990; Labbok, 2015; Lawrence & Lawrence, 2021; McNeilly, 2001; Veile & Miller, 2021). For the birth rate to increase, the birth interval was reduced, typically to between two and three years in early farming communities, which is associated with a reduction in the intensity and/or duration of breastfeeding (Fouts et al., 2005; Haydock et al., 2013; Henderson et al., 2022; Maviso et al., 2022; Waters-Rist et al., 2011). Availability of cereal-based weaning foods is often posited as the reason for early weaning (Bellwood, 2022:139), although weaning foods are not always linked to early weaning (McKerracher et al., 2017; Tessone et al., 2015; Tsutaya et al., 2016). Moreover, early use of weaning foods might imply that women were motivated to abandon breastfeeding at the earliest opportunity, which is not consistent with the ethnographic picture among recent foragers (Hirasawa, 2005; Konner, 2005, 2016; Plomley & Piard-Bernier, 1993).

Another factor may be related to clothing, indirectly: the emergence of modesty, which can discourage prolonged breastfeeding. In the historical era, modesty has been the main reason why women were reluctant to breastfeed—especially in public—and, as a consequence, in the contemporary world many women choose to forego breastfeeding altogether (Lawrence & Lawrence, 2021; Moran, 1999; Rodriguez-Vazquez et al., 2020). In this scenario, the exponential population growth witnessed from the mid-Holocene was based on reduced birth spacing, due the early weaning of infants. Population expansion and dispersals from the farming homelands carried agricultural practices into nearby forager territories and further afield, resulting ultimately in the global spread of farming.

Other Prehistoric Trends

In addition to a specific role for textile clothing in promoting a shift to farming practices, a number of other trends in the late Pleistocene may relate less directly to clothing. Foremost is sedentism, which emerged between 20,000 and 15,000 years ago and led to commensal (passive) domestication of many animal species including dogs, house mice, cats, and pigs (Cucchi et al., 2020; Evin et al., 2017; Linseele et al., 2007; Vigne et al., 2004; Weissbrod et al., 2017; Yeshurun et al., 2014). Commensal processes might have precursors among mobile foragers (Baumann, 2023; Brumm, 2023) and some domesticated species were adopted by mobile foragers, notably the dingo in Australia (Brumm et al., 2023; Koungoulos, 2021). While a discussion of possible psychological factors associated with clothing that could promote sedentism (and commensalism) lies beyond the scope of this paper, a reduction in mobility might be related also to thermal considerations (eg., shelter requirements) during the colder climate regimes in Eurasia from 40,000 years ago (Gilligan, 2019).

Other trends that lie beyond the present remit—but which nevertheless warrant a brief mention—include material engagement and ‘entanglement’ that favoured both sedentism and farming (Hodder, 2006, 2012, 2018). This materiality is less evident among routinely unclad foragers—in Australia for instance (Peterson, 2013), and may have its psychological roots in the most intimate of material attachments, namely to clothes (Gilligan, 2019). Related to sedentism is the enclosure of spaces which becomes more formalized in the walls and houses of the earliest villages (Hodder, 2006), materializing a shift from ‘open’ to ‘closed’ perceptions of the world (Ingold, 2008). Architecture serves as a ‘third skin’ (Drake, 2007), an external extension or projection of human enclosure by clothing, the second skin (Gilligan, 2019).

Historical Trends

Engaging in farming to obtain material for clothing may be a novel notion in archaeology, but farming for fibres would not be news to most farmers. Throughout the historical era, production of textile fibres—along with growing crops to feed fibre-producing animals, notably sheep—has figured prominently in many parts of the world (Simmons & Ekarius, 2019; Tauger, 2020). Australia is a classic example: sheep-grazing to produce wool was the main farming activity from early in the colonial era, with much of the food crop—mainly wheat—devoted to feeding animals, especially sheep (Beinart & Hughes, 2007; Black et al., 2023; Cottle, 2010; Dove & McMullen, 2009; Henzel, 2007; Macintyre, 2020; Ville, 2005).

The proportion of farming devoted to producing fibre has diminished over the past century due to the advent of synthetic fibres, starting in 1935 with a petroleum-derived polymer, nylon (Gaines, 2002), followed in 1941 by the first polyester fibre, Terylene. The commercial success of synthetics was driven mainly by their cost advantages but when compared to natural fibres, synthetics have the distinct disadvantage of being less adept at managing moisture, especially perspiration (Hes & Williams, 2011; Özdemir, 2017).

Discussion

Compared to existing ideas about early farming based on the unchallenged food paradigm, the unique advantage of the textile hypothesis is that it logically locates the development of complex clothing in the late Pleistocene (Gilligan, 2010b), followed by transitions to farming for textiles due to climate change in the early Holocene (Gilligan, 2019). These contingencies did not pertain anywhere prior to the last glacial cycle, and neither are these trends witnessed across all human cultures. Hence the long delay from when H. sapiens emerged around 300,000 years ago (Hublin et al., 2017; Schlebusch et al., 2017), resolving the ‘sapient paradox’ (Renfrew, 2012). The textile hypothesis can accommodate anomalies in the ethnographic picture, where mobile foraging remained the default, or preferred, option in the food economy prior to the emergence of sedentism and the mid-Holocene demographic explosion. Hayden’s (2009) feasting hypothesis is also relevant, although the amplifying role of social complexity becomes evident only in post-Pleistocene and, especially, mid-late Holocene contexts (Bowles & Choi, 2013; Roscoe et al., 2021). Nonetheless, the independent emergence of farming in at least five regions during the early mid-Holocene does suggest a common set of causal processes, among which is climate change (Bellwood, 2023; Cohen, 2009; Piperno, 2011).

In comparison to the prevailing food paradigm, if the textile hypothesis (Gilligan, 2019) is combined with the early farming dispersal hypothesis (Bellwood, 2023), the fundamental issues and substantive anomalies are seen to be broadly consistent with this new paradigm for farming origins (Table 2).

Table 2 Comparison of explanatory models for early farming in relation to fundamental issues and substantive anomalies, illustrating the advantage of combining the textile and early farming dispersal hypotheses into a new paradigm
Full size table

A New Paradigm

Theoretical issues and the current evidence about early farming may be reconciled by combining elements of the textile and early farming dispersal hypotheses. The textile hypothesis, despite an unfashionable emphasis on non-comestible resources, circumvents the many discrepancies and contradictions connected with the prevailing food paradigm. It directly addresses fundamental questions concerning the absence of farming (pre-Holocene and ethnographically), it acknowledges the greater reliability of the foraging food economy compared to early farming, and it accommodates the coincidence with climate warming. The clothing factor could also be relevant to another issue, namely, the massive population growth witnessed in farming homelands. An increase in birth rates can be linked to early weaning of infants, favouring a greater reliance on farming for the production of a food surplus. Together, the textile and early farming dispersal hypotheses offer a plausible and relatively parsimonious theoretical model for the emergence of farming (Figure 3).

Figure 3
figure 3

Schematic model showing how the textile and early farming dispersal hypotheses articulate in a new paradigm for farming origins; temperature graph based on Barron et al. (2003), Steffensen et al. (2008), Antoine et al. (2009), Annan et al. (2022), Jones et al. (2023)

Full size image

Conclusion

Despite its common-sense appeal and subjective plausibility, the food paradigm for farming origins no longer accommodates all potential answers. Likewise, the conceptual category of foragers or hunter-gatherers—which defines traditional Indigenous lifestyles narrowly on the basis of their food economy—has likely ‘reached the end of its useful life’ (Kelly, 2013:22). On the other hand, the textile hypothesis, for all of its novelty, makes no unfounded presumptions about the pivotal role of food. A mutualistic relationship between textile and food perspectives may finally resolve the riddle of farming origins.

The failure of any farming to develop before the early Holocene remains a critical anomaly. Early farming homelands may have been ‘lucky’ to possess the wild ancestors of those major domesticates that feed such a large proportion of the population of the modern world (Bellwood 2023:317), but it cannot be presumed that similar opportunities were not present anywhere between 300,000 and 11,700 years ago. Opportunities for farming would likely have arisen with favourable climate regimes during the last interglacial (Shoaee et al., 2023), and perhaps there were opportunities in parts of Africa even during glacial times. Yet, as Renfrew might say, nothing happened. Possible reasons why humans would ignore earlier opportunities—or why coevolutionary relationships leading to domestication (Zeder, 2017) failed to eventuate—are rarely stated explicitly, and notions of gradual cultural evolution favouring farming have the flavour of accommodative, ad hoc hypotheses—‘just-so stories’ (Hubálek, 2021).

Besides the absence of farming for more than 95% of the existence of H. sapiens, there is a problematic presumption that other hominins were cognitively incapable of farming. This raises a question: how much intelligence is needed to become a farmer? Neanderthals, for instance, are increasingly recognized as possessing cognitive and other capacities (eg., artworks and fibre technologies) that were once considered a prerogative of H. sapiens (Hardy et al., 2020; Marquet et al, 2023; Sansalone et al., 2023; Zilhão et al., 2010). If other hominins are not denied a capacity for farming, there existed a myriad of suitable environments where farming could have begun but did not—for instance, in Europe during the prolonged MIS11 interglacial around 400,000 years ago (Ashton et al., 2005; Brandon et al., 2020; Hao et al., 2015; Holmes et al., 2010; Hosfield, 2022; Sassoon et al., 2023).

Clearly, the coincidence between early farming and early Holocene global warming is a vital clue. Even if there was a general trend towards farming among foragers almost everywhere, it would seem that besides the constant food factor, another factor emerged to favour a more definitive transition to farming at that time. The missing factor may be clothing and, especially, the transition to using woven fabrics as a suitable material for clothes in the early Holocene.