by Mary Harrsch © 2026
In 2008 I photographed a small copper-alloy oil flask in the
British Museum (Registration Number 1824,K/Unknown.0.c), cast in the form of a
young child, curled forward asleep over his knees, a box or lantern tucked
beneath him. A small lid, now lost, would once have completed the figure's
head; the remains of a looped suspension chain are still visible at the
shoulder. The object belongs to the Richard Payne Knight bequest of 1824 and
was found within the walls of Isurium Brigantum — modern Aldborough, North
Yorkshire — the Roman civitas capital of the Brigantes, and it is dated broadly
to the 2nd century CE.
 |
| Copper alloy oil flask in form of a young sleeping slave, 100–200 CE. British Museum, bequeathed by Richard Payne Knight, Object Number 1824,K/Unknown.0.c. Photograph by the author, 2008. |
The subject has a far deeper pedigree than its provincial findspot might suggest. The motif of a servant dozing while he waits — sometimes identified as a lantern-bearer, lanternarius, caught asleep at his post — is a well-documented Hellenistic genre type, with one of its earliest surviving examples in a terracotta vessel from the Fayum in Egypt.
 |
| Ptolemaic terracotta ampulla in the form of a sleeping figure, 2nd century BCE, Fayum, Egypt. British Museum, EA36034. Photograph © The Trustees of the British Museum, (CC BY-NC-SA 4.0) |
Roman adaptations proliferated across media: marble versions served as fountain sculpture in villa gardens at Pompeii and in Syria, while bronze and silver variants were worked into luxury household objects, inkwells, pepper-castors, and — as here — oil flasks for the baths.
 |
| Lanternarius addormentato (sleeping lantern-bearer), 1st–2nd century CE. Museo Nazionale Romano, Terme di Diocleziano, Rome. Photograph by the author, 2005. |
 |
| Silver pepper-pot, Roman, 200–270 CE, from Chaourse, Aisne, France. British Museum, 1889,1019.16. © The Trustees of the British Museum. (CC BY-NC-SA 4.0) |
A close parallel survives in Hannover's Museum August Kestner, and the type's most securely published cousin is a marble statuette in the Metropolitan Museum, dated like ours to the 1st or 2nd century CE.
 |
| Marble statuette of a slave boy with a lantern, Roman, 1st or 2nd century CE. Marble. The Metropolitan Museum of Art, New York. Object Number 23.160.82, Rogers Fund, 1923. Public Domain. |
The Aldborough flask, then, is not a local invention but a provincial instance of a decorative convention that had already circulated for over three centuries by the time it was cast.
It is tempting to read the material as a marker of elite
ownership, but the Roman hierarchy of materials cautions against treating
bronze itself as evidence of elite status. Gold and silver carried a particular
prestige as precious metals, yet elaborate bronze objects were also possessed
and displayed by a much broader range of prosperous Roman households. What does
argue for a comfortable owner here is not the base metal but the workmanship: a
finely modeled, articulated figural vessel with a cast suspension chain
represents considerably more elaborate workmanship than a plain utilitarian
bronze container. Isurium itself supports a civic rather than necessarily
military reading of that owner. Though the settlement grew up near the
short-lived fort at Roecliffe in the conquest period, by the 2nd century it had
developed into a planned, self-governing civitas capital — administered by the
Romanized Brigantian aristocracy through its own ordo, rather than
simply by army officers — with a forum, basilica, and amphitheater, sitting on
Dere Street between York and Hadrian's Wall. If the flask belonged to a
resident household, a prosperous townsman or decurion would therefore be a
plausible owner, even though military traffic continued to pass through the
settlement.
Our flask is not an isolated witness to slavery in Roman Yorkshire. A Samian ware bowl (terra sigillata) in the York Museums Trust collection, recently brought to scholarly attention as part of the Trust's "Old Collections, New Questions" project, depicts eleven captives chained at neck and ankle.
 |
| Roman Samian ware bowl depicting eleven chained captives, c. 170–240 CE, York Museums Trust, YORYM:H32. Photograph by Emily Tilley, York Museums Trust. Image cropped and digitally enhanced by the author. |
Dated to approximately 170–240 CE, it is one of very few representations of Roman captives or slaves in the Trust's collections. It is worth being precise about what such an image does and does not tell us: it reflects how slavery and Roman conquest were represented by Roman citizens, not how bondage was experienced by the enslaved, whose own perspective the object cannot recover. The bowl presents its captives as defeated outsiders and celebrates Roman conquest; it therefore provides evidence for the normalization of slavery within Roman visual culture rather than a first-person account of enslavement.
There is, however, an important distinction to make before
treating this evidence as specifically Brigantian. Cartimandua's surrender of
Caratacus in 51 CE did not constitute the conquest of the Brigantes themselves,
nor is there evidence that large numbers of Brigantians were enslaved as a
consequence. Caratacus was a leader of resistance elsewhere in Britain who had
sought refuge in Brigantian territory; Cartimandua, then a Roman client-queen,
handed him over to Rome while retaining her own political authority. The
eventual Roman conquest of Brigantia came roughly two decades later, after the
collapse of Cartimandua's position and the revolt of Venutius, when Petillius
Cerialis campaigned against the Brigantes in 71–74 CE.
How many Brigantians were enslaved in that conquest is much
harder to establish. The literary sources provide no figure, and archaeology
rarely allows us to distinguish enslaved people from other members of the Roman
population, much less determine their geographic or ethnic origins. What can be
established is that slavery was an entrenched institution in Roman Britain and
that enslaved people were present in northern Britain, including Yorkshire.
Some may have been indigenous Britons; others were likely transported within
the wider Roman slave economy. The distinction matters because the Aldborough
flask demonstrates the presence of an established Roman visual vocabulary of
slavery at Isurium, but it cannot tell us whether the people represented — or
the person who owned it — were Brigantian in origin.
Taken together with the documented presence of enslaved
labor in the households and farms of wealthy Yorkshire villas, and with the
recognition that Roman slave status was not necessarily a lifelong condition —
freedmen and their children went on to play genuine commercial and social roles
in the province — the Samian bowl and the Aldborough flask place slavery within
the established material and social world of Roman Yorkshire. They do not,
however, allow us to reconstruct the size or ethnic composition of the enslaved
population. That distinction is important: evidence for slavery in Roman
Yorkshire is considerably stronger than evidence for the proportion of that
population who were themselves indigenous Brigantes.
That raises a question about meaning rather than ownership
alone: was this genre popular because it depicted slave loyalty or usefulness?
Roman culture certainly possessed an ideological vocabulary surrounding the
behavior expected within master-slave relationships, while Seneca's Letter
47 offers an unusually explicit discussion of Roman attitudes toward
slaves. Writing in the 1st century CE, Seneca urges his correspondent to treat
slaves with familiarity rather than cruelty, while simultaneously describing
the elaborate domestic services slaves performed — including serving wine,
attending banquets, and remaining awake through the night. His argument is
revealing precisely because it takes the institution of slavery for granted
even while criticizing the behavior of individual masters.
But recent scholarship on cognate genre statuary—notably current work on the Oplontis-Borghese “boy with duck” type—argues that such figures are better understood in the context of the Roman trade in pueri delicati: child slaves bought, trained, and displayed as ornamental household possessions.
 |
| Child Strangling a Goose, Roman, 1st century CE, marble copy after a Greek original attributed to Boethos of Chalcedon, 3rd–2nd century BCE. Found at the Villa dei Quintili, Via Appia. Museo Pio Clementino, Vatican Museums, Galleria dei Candelabri, Inv. 2655. Photograph by the author, 2005. |
Their aesthetic appeal, in this interpretation, was rooted in the commodification of enslaved children rather than necessarily in any celebration of fidelity. Here, by “commodification,” I mean the process by which an enslaved or servile person could be represented and treated as an object of aesthetic value and possession, separating the decorative image from the lived experience of the person it represented. The comparison does not prove that every sleeping-servant image carried the same meaning, but it does provide a useful warning against assuming that the apparent charm of such figures represents sympathy toward the enslaved.
The genre's deepest roots push this reading further back
still, into the Hellenistic world. Scholarship on Hellenistic slavery is
explicit that the conquests of Alexander did not soften attitudes toward
bondage as an institution. What they did change was the scale and geography of
the slave trade: conquest generated large numbers of captives and helped spread
Greek-style chattel slavery into regions of the former Persian world where
other forms of dependence had previously predominated. Kelly Wrenhaven's study
of slave representation in Greek art and literature argues that such imagery
did not simply document lived reality but participated in the cultural
representation and normalization of slavery, reflecting how slaveholders wished
to perceive enslaved people rather than how enslaved people experienced
bondage.
Read against that argument, the popularity of the
sleeping-servant type looks less like evidence of sympathy and more like
evidence of a culture deeply accustomed to enslaved labor — a culture in which
a child's exhaustion could be transformed into domestic charm and decorative
art. A Yorkshire bowl showing captives in chains, found within the same broad
provincial world as this flask, suggests that such visual conventions had
travelled with remarkable success to Rome's northern frontier.
What makes the Aldborough flask worth pausing over, then, is
less its individual craftsmanship than its itinerary: an image that emerged in
the Hellenistic world amid the enormous expansion of slave markets associated
with conquest travelled, over three centuries and the length of an empire, into
the material culture of a Romanized Brigantian town at Isurium. It is evidence
of how thoroughly a particular visual vocabulary associated with slavery could
become embedded in provincial life — as far from the Nile as Rome's dominion
reached.
AI disclosure: This article was prepared with the assistance of ChatGPT
and Claude. AI tools were used for research support, source comparison, and
editorial assistance; interpretations and final content were independently
reviewed and determined by the author.
