Seed, fibre and steppe: cannabis in prehistoric Eurasia

Long before anyone wrote about cannabis, people were handling it. They picked its small, hard fruits, twisted its stems into cord, and in some places, at some point, began to sow it on purpose. None of this left a written record. What survived is scattered and physical: a handful of fruits in a Japanese archaeological deposit, the impression of a seed pressed into the clay of a pot, a fibre preserved by chance, grains of pollen settled in the mud of a lake. This article follows that evidence across prehistoric Eurasia, from the coast of Japan at the start of the Holocene to the herders of the Pontic steppe in the Bronze Age, and it spends as much time on how we know as on what we know. Prehistory is where the history of cannabis is most often overstated, and the best defence against that is to understand what each kind of evidence can and cannot prove.

The story here sits between two others in this series. The plant's botanical origins and the genomics of its domestication are covered in the first article, origins; the first secure chemical evidence for psychoactive use, which appears much later, belongs to the third, antiquity. In between lies a long stretch of time in which cannabis mattered to people chiefly as seed and fibre.

Why prehistory is the hard part

For later periods, historians can at least argue over texts. For prehistory there are no texts, only material traces, and cannabis is a frustrating plant to trace. Its fibres rot. Its seeds are small and look superficially like those of other plants. Its pollen is very hard to distinguish from that of its close relative, the hop. And it grows readily as a wild or weedy plant, so finding traces of it near a human settlement does not automatically mean people were using it, let alone farming it.

The review paper on which this series is based makes a distinction that runs through everything that follows. It separates three broad categories of evidence:

• Pollen evidence, which shows that the plant was present in a landscape.
• Macrofossil and fibre evidence, which can show that people used the plant.
• Chemical residue evidence, which can show psychoactive use.

Confusing these categories, the paper notes, is one of the most common sources of error in popular histories (Long et al., 2017; McPartland, Hegman and Long, 2019). For almost the whole period covered in this article, the evidence is of the first two kinds. Chemical evidence of psychoactive use does not appear until the first millennium BCE, thousands of years after the earliest finds discussed here.

Reading the evidence: a short field guide

Before visiting the sites, it is worth understanding the tools, each with its own strengths and blind spots.

Pollen and palynology

Palynology is the study of pollen and spores. Cannabis is wind-pollinated, and wind-pollinated plants tend to release large quantities of pollen, much of which drifts and settles on lakes, bogs and wetlands. There, buried in successive layers of sediment, it can survive for tens of thousands of years. Researchers drill cores from these sediments, date the layers, and count the pollen grains of different plants under the microscope. The result is a picture of how the vegetation of a region changed over time.

For cannabis, there are two big limitations. The first is identification. The pollen of Cannabis and Humulus (hops), both members of the family Cannabaceae, is hard to tell apart, and many pollen studies report it only under collective labels such as "Cannabis/Humulus" or "Cannabaceae". To get round this, McPartland and colleagues developed probabilistic ecological criteria for assigning such pollen to one genus or the other, based on the kind of vegetation and environment in which it occurs (McPartland, Hegman and Long, 2019). This is a reasoned estimate, not a direct identification.

The second limitation is more fundamental: pollen shows presence, not use. Wild cannabis is an abundant member of the steppe and montane vegetation of Central Eurasia (Ren et al., 2019). A rise in cannabis-type pollen might reflect people growing the plant nearby, or it might reflect forest clearance, climatic change, or the spread of disturbed ground in which weedy plants thrive. Pollen is excellent for asking where the plant lived; it is weak on its own for asking what people did with it.

Achenes, the fruits we call seeds

What people call a hemp "seed" is, botanically, a fruit: an achene, a small, dry, single-seeded fruit that does not split open when ripe. Achenes have a hard outer wall, which helps them survive. When they are charred in a fire, or preserved in waterlogged or very dry conditions, they can last for millennia and be recovered by archaeologists, often by flotation, in which soil samples are washed so that light plant remains float to the surface.

Achenes are called macrofossils or macro-remains because they are visible to the naked eye, unlike pollen. Their great advantage is that they can be identified more confidently than pollen, and when they are found in a settlement, a storage pit or a grave they are physically associated with human activity. They can also be radiocarbon-dated directly. Their limitation is the one already mentioned: a fruit tells you that a plant was there and probably that someone handled it, but it does not reveal whether that plant was sown, tended, or simply gathered from a wild stand.

Achene imprints

Sometimes the fruit itself has gone, but its shape remains. Potters working with damp clay occasionally picked up seeds and other plant material that were lying around, and when the pot was fired the plant material burned away, leaving a cavity or impression. These achene imprints can be studied, and sometimes cast, to identify the plant. Long et al. (2017) included them in their review alongside actual achenes. Imprints are useful because they show the plant was present in the immediate surroundings of a household or workshop, but like whole fruits they cannot by themselves distinguish crop from weed.

Fibres, textiles and cordage

Hemp is one of the great bast-fibre plants: its useful fibres run along the outside of the stem. Fibre is direct evidence of one of the plant's main uses, and textiles, cordage and loose fibres all appear in the reviews discussed below (Long et al., 2017; McPartland and Hegman, 2018). The trouble is preservation. Plant fibres decay quickly in most soils, and survive only in unusual conditions such as waterlogging, extreme dryness, freezing, or contact with metal. Even when fibres survive, distinguishing hemp from other bast fibres such as flax or nettle requires careful microscopic work, and is not always possible.

Pseudomorphs and phytoliths

Two more specialised categories appear in the critical review of European evidence by McPartland and Hegman (2018).

A pseudomorph is, literally, a "false form": a structure that preserves the shape of an object after the original material has been replaced or has disappeared. In archaeology, a common example is textile or fibre that has been replaced by mineral or corrosion products, often where it lay against a metal object; the fibre itself has gone, but its form survives. Identifying the original plant from a pseudomorph is difficult.

Phytoliths are microscopic particles of silica that form inside the tissues of many plants. Because they are mineral, they survive long after the plant has decayed, and some plants produce distinctive shapes. Like pollen, they require careful identification.

Radiocarbon dates and "cal BP"

Most of the dates in this article are radiocarbon dates. Living things take up carbon from the atmosphere, including a small proportion of the radioactive isotope carbon-14. When an organism dies it stops taking up carbon, and the carbon-14 it contains decays at a known rate, halving roughly every 5,700 years. Measuring how much remains in a sample gives an estimate of how long ago the organism died.

The raw measurement is not a calendar date, because the amount of carbon-14 in the atmosphere has varied over time. Scientists therefore calibrate radiocarbon results against records of known age, such as tree rings, to convert them into calendar years. Calibrated dates are written "cal BP" (calibrated years before present, where "present" is conventionally fixed at AD 1950) or "cal BCE" (calibrated years before the Common Era). That is why the same Japanese find can be described as about 10,000 cal BP or about 8000 cal BCE: the two figures express the same calibrated age on different scales. Calibrated dates always carry a margin of uncertainty, so a single date should be read as a range, not a point.

A crucial advantage of radiocarbon is that it can be applied directly to the plant remains themselves. A "directly dated" fruit has had its own carbon measured. That is much stronger evidence than a fruit dated only by association with the layer it was found in, because small objects can move between layers through burrowing animals, roots, or later digging.

Systematic and critical reviews

Finally, much of what follows rests not on single excavations but on reviews that pull together dozens or hundreds of records. A systematic review gathers all the published records that meet stated criteria and analyses them together, so that patterns emerge which no individual site could reveal (Long et al., 2017). A critical review goes further in sifting: it assesses each record's reliability and discards those that do not stand up (McPartland and Hegman, 2018). Both approaches depend on the quality of the underlying reports, and both are only as good as their criteria. But they are far more reliable than the traditional habit of stringing together striking individual finds into a narrative.

The evidence at a glance

Okinoshima: hemp fruits at the dawn of the Jōmon

Who were the Jōmon?

The Jōmon period is the long prehistoric era of the Japanese archipelago that preceded the arrival of wet-rice farming in the first millennium BCE. Its name means "cord-marked", after the impressions of twisted cord that decorate much of its pottery. Jōmon communities are best known as hunters, gatherers and fishers who nevertheless made pottery very early, lived in settled or semi-settled villages, and exploited a wide range of plant and marine foods. Archaeologists divide the period into successive phases, and the deposits discussed below are described as belonging to its earliest part.

They were not grain farmers, but they knew their plants intimately and had good reason to value one that offered both oily seeds and strong fibre.

The find and its date

The earliest directly dated cannabis macrofossils associated with humans come from Japan. At the Okinoshima site on the Bōsō Peninsula, in Chiba Prefecture, east of Tokyo Bay, Cannabis sativa fruits were recovered from earliest-Jōmon deposits. They were radiocarbon-dated to approximately 10,000 cal BP, or around 8000 cal BCE (Kudo et al., 2009; Okazaki et al., 2011).

Three features make this record important.

First, the fruits were dated directly. As explained above, that removes the worry that small seeds might have slipped down from younger layers. The age belongs to the fruits themselves.

Second, they were recovered from an archaeological site, not from a natural sediment core. They are therefore associated with a place where people lived and worked, which is what allows them to be counted as evidence of human use and not merely of the plant's presence.

Third, the date is early. Around 8000 cal BCE is the opening stretch of the Holocene, the warm geological epoch that followed the last glaciation and in which we still live. It is early enough that, if the inference of human use holds, people in Japan were handling cannabis at a time when farming was still in its earliest stages even where it first arose. On the strength of such finds, the review paper concludes that seeds from Jōmon-period Japan show human use by about 10,000 years ago.

What the find does not show is equally important. It does not show that the fruits came from cultivated plants. It does not show what the people of Okinoshima did with them: the fruits might have been eaten, pressed for oil, or simply brought in attached to stems collected for fibre. And it says nothing whatever about psychoactive use. The paper is careful on this point, and so should we be.

A changing coast

Okazaki et al. (2011) set the Okinoshima finds in the context of Early Holocene coastal change. This matters for understanding the site. As the great ice sheets of the last glaciation melted, global sea levels rose substantially, flooding low-lying land and redrawing coastlines around the world, including those of the Japanese archipelago. A coastal site occupied in the Early Holocene therefore sat in a landscape that was changing within the span of human memory: shorelines advancing, inlets forming, wetlands shifting. Reconstructing that environment helps archaeologists understand how a deposit formed and how its layers relate to one another, which is part of the evidence for a find whose significance depends so heavily on its date.

Why an island matters

There is a further reason why Okinoshima attracts attention. The Japanese archipelago lies outside the plant's probable natural centre of origin, which the pollen evidence places on the north-eastern Tibetan Plateau (McPartland, Hegman and Long, 2019). If cannabis did not grow wild in Japan at the end of the last Ice Age, then fruits in an early Jōmon deposit would suggest that people had carried the plant there, or were managing it in some way.

That is why the Okinoshima finds are often taken to suggest early human transport or management of the plant. But as Clarke and Merlin (2013) point out, the inference depends on assumptions about the plant's wild range in the terminal Pleistocene, the closing phase of the last glacial period. We do not have a detailed map of where wild cannabis grew at that time. If its natural range extended further east than is usually assumed, or if it had reached the archipelago by natural means, the Okinoshima fruits could reflect people gathering a locally available wild plant. The find is strong evidence that people and cannabis were together in Japan around 10,000 years ago. Whether they had brought it there is a more open question.

One origin of use, or many?

Long et al.'s systematic review

Single sites, however well dated, can only tell part of the story. To see the bigger picture, Long et al. (2017) carried out a systematic review of Eurasian archaeological and palaeoenvironmental records of cannabis. They brought together four kinds of record: fibres, pollen, achenes and achene imprints. Their aim was to ask where and when human use of cannabis began, and how it spread.

Their answer challenged the tidy idea of a single point of origin. Because cannabis records appear at more or less the same time in Europe and East Asia at the end of the last glaciation, they proposed a multiregional origin for human use (Long et al., 2017). On this view, people at opposite ends of Eurasia, with no knowledge of one another, independently began to make use of a plant that grew in their surroundings.

The end of the last glaciation

The timing is significant. The end of the last glaciation was a period of rapid environmental change across Eurasia. As climates warmed, vegetation zones shifted, forests expanded, and open, disturbed habitats appeared and disappeared. Cannabis is a plant that thrives in open and disturbed ground, the kind of ground that people themselves create around camps and settlements. It is easy to imagine how, in many different places, a useful weedy plant growing near human dwellings could become familiar, and then valued.

In that setting the multiregional hypothesis makes intuitive sense: if wild cannabis was widespread and useful, there is no particular reason why only one group of people should have discovered it.

That said, the hypothesis rests on the near-simultaneous appearance of records at opposite ends of a continent, and records of this age are few and unevenly distributed. "More or less the same time" at this distance in the past can conceal differences of centuries or more, and some of the underlying records are of the weaker kinds (pollen, for instance) discussed above. The multiregional model is a reasoned interpretation of the available data, not a settled fact.

The surge of 5000–4000 cal BP

The second major finding of Long et al. (2017) concerns a much later period. They identified a marked increase in cannabis achene records in East Asia between about 5000 and 4000 cal BP, that is, roughly the third millennium BCE. They suggested that this increase might be associated with a trans-Eurasian exchange and migration network through the steppe zone.

The third millennium BCE was a time of spreading metalworking, widening exchange and increasingly mobile pastoral societies on the Eurasian grasslands, and long-distance movement of crops and technologies across the continent is a major theme of the archaeology of this and the following centuries. The subtitle of Long et al.'s paper speaks of "Bronze Age trans-continental connections".

Two cautions apply. The first is that an increase in the number of records does not necessarily mean an increase in use: it may also reflect where archaeologists have dug, how intensively sites have been sampled, and how many sites of a given period survive. Systematic reviews try to account for this, but cannot eliminate it. The second is that Long et al. offered the steppe connection as a possible association ("might be"), not a demonstrated cause. The paper preserves that hedge, and so should any retelling.

Reconciling the archaeology with the genome

At first sight, the archaeological picture seems to sit awkwardly with the genetic one. The most detailed genomic study to date, which resequenced 110 whole genomes, concluded that C. sativa "was first domesticated in early Neolithic times in East Asia" (Ren et al., 2021). That sounds like a single origin, while Long et al. (2017) argue for many. The genomic study itself, and the debate over its "East Asian" label, are discussed in the first article of this series; here the question is simply how the two fit together.

The answer is that they are describing different things. The genomic study is about domestication: the emergence of a lineage of plants genetically shaped by human selection, from which all modern hemp and drug cultivars descend. The archaeological review is about use: people gathering and handling the plant, whether or not they were changing it genetically. As the review paper puts it, the two positions are not necessarily contradictory: wild plants may have been gathered in several regions before the lineage ancestral to modern cultivars was domesticated in East Asia (Ren et al., 2021).

An analogy helps. People gathered wild grasses in many places long before particular populations of those grasses were turned into domestic cereals. The gathering was widespread; the genetic transformation happened in specific places, and the domesticated forms then spread and often replaced their local wild relatives. Something similar may be true of cannabis. Many groups across Eurasia may have used local wild cannabis, while only one lineage, in East Asia, became the ancestor of the plants grown today.

The genomic timeline also leaves room for this picture. Ren et al. (2021) estimate that early domesticated ancestors of hemp and drug types diverged from basal cannabis about 12,000 years BP, but with a 95% confidence interval stretching from 6,458 to 15,728 years BP. The review paper stresses that this is a model-based estimate, not a precise date. A window that wide overlaps comfortably with the Okinoshima finds, the end of the last glaciation, and much of the early Neolithic. Genomic dating and radiocarbon dating are not yet precise enough to say which came first, and because modern domestic and feral populations hybridise extensively, the genomic picture remains under revision.

What would help most, as the paper notes in its discussion of gaps in the evidence, is ancient DNA from archaeobotanical remains. If DNA could be recovered from dated prehistoric fruits, it would become possible to ask directly whether a given find belonged to a wild population or to the domesticated lineage. Until then, archaeology and genomics are two partial views of the same landscape.

Europe: a native plant, a late crop

Indigenous before people

A traditional view of cannabis in Europe held that it was an Asian plant carried westwards by people. Pollen evidence has complicated that picture considerably. A synthesis of European pollen records argued that cannabis was indigenous to Europe long before people arrived, and that cultivation there began in the Copper or Bronze Age (McPartland and Hegman, 2018). If that is right, the history of cannabis in Europe is not the story of a plant arriving, but the story of a familiar wild plant being drawn, quite late, into agriculture. As at Okinoshima, whether a find counts as evidence of introduction depends on whether the plant already grew locally.

A critical review

To ask when Europeans began to use cannabis, McPartland and Hegman (2018) critically reviewed the European archaeological evidence: textiles, cordage, fibres, seeds, impressions, pseudomorphs and phytoliths. Their approach was deliberately sceptical. Rather than accepting every published report of "hemp", they weighed the reliability of each, given the identification problems outlined earlier.

The result was sobering for anyone who had imagined a deep European hemp tradition stretching back into the Stone Age. They found one robust Copper Age record in south-eastern Europe. More robust evidence appears only in the Bronze Age, on the Pontic steppe. And the spread of hemp cultivation to peoples further north and west, they argued, came later still, in the Iron Age.

The Copper Age: Gumelnița–Varna

The single robust Copper Age record comes from the Gumelnița–Varna culture (McPartland and Hegman, 2018). The Copper Age, or Chalcolithic, is the period in which communities began working copper while still relying heavily on stone tools, before the adoption of bronze. In south-eastern Europe it was a time of large, long-lived farming villages, often built up into settlement mounds, and of impressive craft production.

The Gumelnița–Varna culture flourished in the lower Danube region and along the western Black Sea coast, in what are now Romania and Bulgaria, in the fifth millennium BCE. It is best known for the cemetery at Varna on the Bulgarian coast, whose graves contained some of the oldest worked gold yet found anywhere. These were complex societies with marked differences in wealth, extensive exchange networks and sophisticated technologies. A single robust cannabis record from such a context shows that the plant was being used there, but one record is a slender basis for any broader generalisation. It certainly does not establish a widespread Copper Age hemp economy.

The Bronze Age steppe: Yamnaya and Catacomb

More robust evidence appears in the Bronze Age Yamnaya and Catacomb cultures of the Pontic steppe (McPartland and Hegman, 2018).

The Pontic steppe is the great belt of grassland north of the Black Sea, extending eastwards towards the Caspian. In the late fourth and early third millennia BCE it was home to the Yamnaya culture, mobile pastoralists who herded cattle, sheep and goats, used wheeled wagons, and buried their dead beneath earthen mounds known as kurgans. The Yamnaya have become famous in recent years because ancient-DNA studies have linked them with large-scale movements of people across Europe, though the details and wider implications of those movements are still debated. The Catacomb culture followed the Yamnaya on the Pontic steppe in the third millennium BCE; it takes its name from its graves, in which the burial chamber was cut as a niche to one side of a vertical shaft.

That cannabis evidence should become more robust among these mobile herding peoples is striking. It fits the observation, made in many areas of archaeology, that the steppe was a corridor along which things and knowledge moved. It also chimes with the suggestion of Long et al. (2017) that a trans-Eurasian network through the steppe zone might help explain the surge in East Asian achene records in the same broad period. The two reviews were looking at different ends of the continent, but both point to the Bronze Age steppe as a place where cannabis and human mobility met.

The Iron Age: Scythians as carriers of cultivation

The last step in McPartland and Hegman's (2018) European sequence concerns the Iron Age Scythians. The Scythians were a group of horse-riding nomadic peoples who dominated much of the Eurasian steppe in the first millennium BCE, and who are well known to historians because Greek writers, above all Herodotus, described them. (Herodotus' famous account of Scythian hemp vapour baths, and the archaeological finds associated with it, are discussed in the next article, antiquity.)

McPartland and Hegman argued that the Scythians probably introduced hemp cultivation to Celtic, Slavic and Finno-Ugric peoples. Those three labels are broad language families, conventionally used for the peoples of western and central Europe, of eastern Europe, and of parts of north-eastern Europe and beyond respectively. The claim, then, is that across a wide swathe of Iron Age Europe, the practice of deliberately growing hemp was learned from steppe neighbours, even if wild cannabis had long been present. Note the word "probably": this is an interpretation of patchy evidence, not a documented transfer. The picture fits a familiar image of mobile steppe pastoralists as vectors of plant knowledge.

The "camp follower" problem

There is a significant objection, however. Ren et al. (2019) caution that because wild cannabis was already widespread across Central Eurasia, claims that steppe peoples spread the plant as a "camp follower" need to be clarified.

The idea of a camp follower is that a plant spreads by tagging along with people: its seeds travel in fodder, on animals, in bedding or in stores, and it springs up on the disturbed, nutrient-rich ground around camps and settlements. Cannabis is often described this way. But if the plant already grew across much of the steppe and its margins, then finding it in a new place does not by itself prove that people brought it. It may have been there all along.

This is the same logical problem met at Okinoshima and in the European pollen record. Everything turns on the plant's natural range, and that range is imperfectly known. The steppe peoples may well have spread knowledge of how to grow and process hemp; that is what McPartland and Hegman's argument is chiefly about. Whether they also spread the plant itself is a separate question, and Ren et al.'s caution is a reminder not to run the two together.

Three interpretive cautions

The review paper identifies three cautions that run through the whole prehistoric record (Long et al., 2017; McPartland, Hegman and Long, 2019). Each corresponds to a recurring mistake.

Cord-marked pottery is not hemp evidence

"Cord-marked" pottery is often cited as early evidence of hemp cordage. The reasoning runs: the potters pressed cord into the clay; cord in this region was made of hemp; therefore the pottery proves early hemp use. The weak link is the middle step. The fibre species usually cannot be identified from the impressions. A cord impression records the twist and thickness of a cord, not the plant it was made from. Many plants yield fibre suitable for cordage, and in most cases there is no way of telling them apart from the mark left in clay.

There is a certain irony here for East Asia, since the very name of the Jōmon period means "cord-marked". Cord-decorated ceramics are one of the most characteristic artefacts of prehistoric East Asia, and it is tempting to connect them with a plant famous for its fibre. But the connection is an assumption, not an observation. Where hemp fibre has actually been identified, that is real evidence; where only the impression survives, it is not.

Foraged or farmed?

The second caution is that seeds may come from foraged wild plants, not cultivated ones. This is perhaps the most important single point for interpreting the early record.

In many crops, archaeologists can recognise domestication in the seeds themselves. Domesticated cereals, for example, often develop larger grains and lose the ability to shed their seeds naturally, so that they depend on people to harvest and sow them. The early cannabis finds discussed here have not been shown to carry such markers, so a fruit from an early site tells us that the plant was present and probably handled, but not whether anyone had planted it. Gathering wild plants is an ancient and sophisticated practice in its own right, and many people across prehistoric Eurasia, including the Jōmon, were highly skilled foragers. To call every early cannabis find "cultivation" is to ignore that possibility entirely.

Pollen is not use

The third caution, already discussed, is that pollen shows the plant was present, not that people used it. As explained earlier, pollen is indispensable for reconstructing the plant's natural history, and it was crucial in establishing its probable centre of origin (McPartland, Hegman and Long, 2019). But it cannot carry the weight of claims about human use without supporting evidence of other kinds.

The "first cultivated crop" myth

Many popular histories state that hemp was "the first cultivated crop". The review paper observes that such claims pass over the distinctions just described. Once cord-marked pottery is set aside as unidentifiable, early fruits are allowed to be foraged, and pollen is read as presence rather than use, the case for hemp as humanity's first crop largely dissolves. What remains is something more interesting and more defensible: that cannabis was among the plants people knew and used very early, in more than one region, long before it became a crop in the full sense, and long before anyone, so far as the evidence shows, used it for its psychoactive properties.

Seed and fibre before smoke

In popular imagination cannabis is first and foremost a drug. The archaeology tells a different story. For thousands of years, what survives of cannabis in human contexts is fruits, fibres and the impressions they left. The review paper's discussion notes that archaeology shows seeds and fibre long before securely attested psychoactive use (Long et al., 2017; Kudo et al., 2009).

Why might early peoples have valued the plant? Hemp fruits are oil-rich, a valuable food. Hemp stems yield long, strong bast fibres, suitable for cords, nets, bags and eventually textiles. A plant that offered both, and that grew readily on disturbed ground near human dwellings, was the kind of plant that people would naturally come to know, protect, and eventually sow.

This is why the review paper frames cannabis as a multi-purpose "companion species", whose fibre, seed, resin and flowering tops have been valued in different combinations by different societies. In prehistory, the combination was seed and fibre. Resin and intoxication enter the record later; when they do, in the first millennium BCE, it is in Central Asia and the Levant, and it is the subject of the next article in this series.

One point of method is worth adding. Seeds and fibres are more likely to survive than any trace of how a plant affected the mind, which leaves nothing unless chemical residues are preserved and analysed. The absence of early evidence for psychoactive use does not prove that no one in prehistory discovered it; it means only that, on present evidence, it cannot be shown.

What we know, and what we don't

What the evidence supports

• People were handling cannabis in Japan around 10,000 cal BP (c. 8000 cal BCE), on the evidence of directly radiocarbon-dated fruits from earliest-Jōmon deposits at Okinoshima on the Bōsō Peninsula (Kudo et al., 2009; Okazaki et al., 2011).
• Across Eurasia, cannabis records appear at more or less the same time in Europe and East Asia at the end of the last glaciation, which supports the proposal of a multiregional origin of human use (Long et al., 2017).
• Cannabis achene records in East Asia increase markedly between about 5000 and 4000 cal BP (Long et al., 2017).
• Pollen synthesis indicates that cannabis was indigenous to Europe before people arrived, with cultivation beginning in the Copper or Bronze Age (McPartland and Hegman, 2018).
• Critically sifted European evidence shows one robust Copper Age record (Gumelnița–Varna) and more robust Bronze Age evidence among the Yamnaya and Catacomb cultures of the Pontic steppe (McPartland and Hegman, 2018).
• Throughout prehistory, the evidence of human use is of seed and fibre. Securely attested psychoactive use comes much later.

What remains uncertain

• Whether the Okinoshima fruits were brought to Japan by people. That inference depends on assumptions about the plant's wild range in the terminal Pleistocene (Clarke and Merlin, 2013).
• Whether early finds represent cultivation or foraging. In most cases the fruits themselves cannot tell us.
• Why East Asian achene records increase around 5000–4000 cal BP. A trans-Eurasian exchange network through the steppe is a suggestion, not a demonstrated cause (Long et al., 2017).
• What exactly steppe peoples spread. The Scythians probably introduced hemp cultivation to Celtic, Slavic and Finno-Ugric peoples, but because wild cannabis was already widespread, "camp follower" claims need clarification (McPartland and Hegman, 2018; Ren et al., 2019).
• How use relates to domestication in time. The genomic divergence estimate of about 12,000 years BP carries a confidence interval of 6,458–15,728 years BP and is model-based (Ren et al., 2021).
• How representative the record is. Fibres rarely survive, pollen is hard to identify, and some regions have been studied far more intensively than others. The review paper also acknowledges that non-English scholarship, particularly in Chinese, Japanese and Russian, is under-represented in its sources, which matters a great deal for a story centred on East Asia and the steppe.

What would change the picture

The most promising developments would be ancient DNA from archaeobotanical remains, which could show whether particular prehistoric fruits belonged to wild or domesticated lineages; more directly dated finds from well-recorded contexts; and better methods for separating cannabis from hop pollen and identifying fibres.

Until then, the honest summary is this. Cannabis has accompanied people since at least the early Holocene. For most of that time it was valued for what it could feed and what it could bind. Its spread across Eurasia was bound up with the movements of steppe peoples in the Bronze and Iron Ages, but how far those peoples carried the plant, and how far they carried the know-how to grow it, is still being worked out.

Sources

Clarke, R.C. and Merlin, M.D. (2013) Cannabis: Evolution and Ethnobotany. Berkeley: University of California Press. doi:10.1525/9780520954571.

Kudo, Y., Kobayashi, M., Momohara, A., Noshiro, S., Nakamura, T., Okitsu, S., Yanagisawa, S. and Okamoto, T. (2009) 'Radiocarbon dating of the fossil hemp fruits in the earliest Jomon period from the Okinoshima site, Chiba, Japan', Japanese Journal of Historical Botany, 17, pp. 27–31.

Long, T., Wagner, M., Demske, D., Leipe, C. and Tarasov, P.E. (2017) 'Cannabis in Eurasia: origin of human use and Bronze Age trans-continental connections', Vegetation History and Archaeobotany, 26, pp. 245–258. doi:10.1007/s00334-016-0579-6.

McPartland, J.M. and Hegman, W. (2018) 'Cannabis utilization and diffusion patterns in prehistoric Europe: a critical analysis of archaeological evidence', Vegetation History and Archaeobotany, 27, pp. 627–634. doi:10.1007/s00334-017-0646-7.

McPartland, J.M., Hegman, W. and Long, T. (2019) 'Cannabis in Asia: its center of origin and early cultivation, based on a synthesis of subfossil pollen and archaeobotanical studies', Vegetation History and Archaeobotany, 28(6), pp. 691–702. doi:10.1007/s00334-019-00731-8.

Okazaki, H., Kobayashi, M., Momohara, A., Eguchi, S., Okamoto, T., Yanagisawa, S., Okubo, S. and Kiyonaga, J. (2011) 'Early Holocene coastal environment change inferred from deposits at Okinoshima archeological site, Boso Peninsula, central Japan', Quaternary International, 230, pp. 87–94. doi:10.1016/j.quaint.2009.11.002.

Ren, G., Zhang, X., Li, Y., Ridout, K., Serrano-Serrano, M.L., Yang, Y., Liu, A., Ravikanth, G., Nawaz, M.A., Mumtaz, A.S., Salamin, N. and Fumagalli, L. (2021) 'Large-scale whole-genome resequencing unravels the domestication history of Cannabis sativa', Science Advances, 7(29), eabg2286. doi:10.1126/sciadv.abg2286.

Ren, M., Tang, Z., Wu, X., Spengler, R., Jiang, H., Yang, Y. and Boivin, N. (2019) 'The origins of cannabis smoking: chemical residue evidence from the first millennium BCE in the Pamirs', Science Advances, 5(6), eaaw1391. doi:10.1126/sciadv.aaw1391.