AnTard, 19, 2011, p .

171- 190

4 - LE DOMAINE PRIVÉ

CHRISTIAN IMAGES IN THE HOME

KIM BOWES

Images chrétiennes dans l’habitat privé

Cet article traite de la fonction des thèmes de l’imagerie chrétienne en contexte domestique, en se concentrant sur
les images permanentes : mosaïques, peintures murales et sculptures architecturales. Après identification, trois caté-
gories principales peuvent être distinguées dans ces images : celles associées au rituel, celles utilisées pour protéger,
ou encore celles pour indiquer le statut et l’identité du propriétaire. On constate que, dans l’habitat antique tardif, les
images chrétiennes mises en œuvre faisaient appel à un choix très individualisé, mais ne s’en adressaient pas moins à
un large public (visiteurs). La relation complexe du domaine privé au spectateur public se traduit d’ailleurs souvent
par des images qui résistent aux catégories interprétatives modernes, parfois trop faciles. [Trad. de la Rédaction.]

John Chrysostom marveled that the faithful so loved
the bishop Meletios that they set up images of him in
their homes, and impressed that image on rings, seals and
bowls1. Theodoret reported that images of Simeon Stylites
the Elder were hung on Roman shop-fronts as protection
against evil2. And even Sidonius Apollinaris could laconi-
cally report that in the sumptuous villa of Pontius Leontius
of Bordeaux, between the weaving chamber and a cozy
living room, lay an apartment decorated with scenes from
the Old Testament3.
This paper addresses this phenomenon of Christian
domestic images. It is a phenomenon whose ubiquity seems
assured by texts like those quoted above, but for which
we have very little clear-cut physical evidence. The best
evidence is, in fact, textual, and that textual evidence often
seems to describe portable objects – panel paintings, rings,
pilgrimage tokens and textiles. These objects are rarely
found in situ and thus practically all the extant physical
exempla are disjecta membra, sat in museum cases, whose
domestic context can be only guessed at. Thus the discus-
sions of Christian images in the home have largely been
1. John Chrysostom, Homilia encomiastica in Meletium, PG 50, col. 515-
516.
2. Theodoret, Hist. Rel. 26.11
3. Sidonius Apollinaris, Carm. 22.200-203.
footnotes to discussions of the objects qua objects, that is,
to discussions of icons, pilgrimage tokens, phylacteries, and
textiles, or to the objects’ iconographic content – images of
the Virgin or so-called magical symbols4. The “domestic”
aspect of Christian domestic images has thus been unders-
tandably somewhat sidestepped.
This paper departs from a different corpus of evidence,
namely Christian images actually known to have derived
4. See respectively, J. Herrin, Women and faith in icons in early Christianity,
in R. Samuel and G. S. Jones (dir.), Culture, Ideology and Politics.
Essays for Eric Hobsbawn, London, 1982, pp. 56-83; R. Cormack,
Women and icons and women in icons, in L. James (dir.), Women,
Men and Eunuchs. Gender in Byzantium, London / New York, 1997,
pp. 24-51; G. Vikan, Byzantine Pilgrimage Art, Washington DC, 1982;
B. Pitarakis, Female piety in context: Understanding developments in
private devotional contexts, in M. Vassilaki (dir.), Images of the Mother
of God: Perceptions of the Theotokos in Byzantium, Ashgate, 2005,
pp. 153-166; H. Maguire, Garments pleasing to God: The signi¿cance
of domestic textile designs in the early Byzantine period, in DOP, 44,
1990, pp. 215-224; H. Maguire, The cult of the Mother of God in
private, in M. Vassilaki (dir.), Mother Of God: Representations of
the Virgin in Byzantine Art, Athens, 2000, pp. 279-289; H. Maguire,
Byzantine domestic art as evidence for the early cult of the Virgin, in
M. Vassilaki (dir.), Images of the Mother of God: Perceptions of the
Theotokos in Byzantium, Ashgate, 2005, pp. 183-194; E. Maguire,
H. Maguire and M. Duncan-Flowers, Art and Holy Powers in the Early
Christian House, Urbana and Chicago, 1989; H. Maguire, Magic and
geometry in early Christian textiles, in Jahrbuch der österreichischen
Byzantinistik, 44, 1994, pp. 265-274.
172 KIM BOWES AnTard, 1 9, 2011

from houses and thus from images found in situ – wall

paintings, mosaic Àoors and architectural sculpture. That
is, it focuses on Christian images permanently inserted
into the home. In doing so, it seeks to accomplish two
goals: to better understand how Christian images operated
within speci¿cally domestic contexts, and to expand our
understanding of the many functional roles such images
served. Beneath the “phenomenon” of Christian images
in the home lurks a multitude of functional phenomena,
in which Christian images were not only used as ritual
objects and agents of protection, but also as statements of
Christian af¿liation and vehicles for the construction of
personal identity.

I. DEFINING “CHRISTIAN” AND “HOUSE”

The history of Christian archaeology is littered with

announcements of the discovery of Christian domestic
images, announcements later generations of scholars have
often dismissed. The grounds for dismissal are typically one
of two reasons: the image in question cannot be shown to Fig. 1 – Aquileia, “Oratorio del Buon Pastore”, 4th c.
a/ House plan; b/ Drawing of mosaic (M. Mirabella
be discernably “Christian,” or the space in question cannot Roberti, Edilizia privata in Aquileia, Vita Sociale Artistica
be proven to be a house. e Commerciale di Aquileia Romana, Aquileia, 1987, tav. 2;
The perils of the scholarly quest for Christian domestic L. Bertacchi, Architettura e mosaico. Gli oratori privati, in Da
Aquileia a Venezia, Milan, 1980, tav. 23, respectively).
images are most evident in the excavations of Aquileia.
Here no less than ¿ve domestic “oratories” were identi¿ed,
three on the basis of the mosaic imagery5. The “Oratorio
con la Scena di Pesca” featured a ¿shing scene, thought to
represent the Christian faithful, while a carpet of medallions
containing animals were interpreted as the earthly paradise.
The “Oratorio del Buon Pastore” featured a central image
of a shepherd, Àanked by sheep and goats and holding a
syrinx, and was assumed to represent Christ as the Good
Shepherd (¿J); while the so-called “Oratorio del Buon
Pastore dall’Abito Singolare,” depicted the only likely
Christian shepherd, here clad pants and tunic, a hand raised
in blessing (¿J ). As later scholarship showed, all of
these rooms were probably reception or dining rooms, not
oratories, and only the Buon Pastore dall’Abito Singolare
has probable Christian imagery6.
Why these mosaics should ever have been considered
Christian is due principally to their local context: Aquileia

5. See G. Brusin and E. Zovatto, Monumenti paleocristiani di Aquileia e

di Grado, Udine, 1957, pp. 191-209; 212-230; G. Brusin, Due nuovi
sacelli cristiani di Aquileia, Aquileia, 1961; P. L. Zovatto, Mosaici
paleocristiani delle Venezie, Udine, 1963, pp. 111-123; G. Bovini,
Le antichità cristiane di Aquileia, Bologna, 1972, pp. 383-439;
L. Bertacchi, Architettura e mosaico. Gli oratori privati, in Da Aquileia
a Venezia, Milan, 1980, pp. 265-269.
6. N. Duval, Quelques remarques sur les ‘églises-halles’, in Aquileia nel
IV Secolo, 2, 1982, pp. 399-412; A. Holden, The cultivation of upper-
class otium: Two Aquileian ‘oratory’ pavements reconsidered, in
Studies in Iconography, 23, 2002, pp. 29-54.
AnTard, 1 9 , 2 0 11
Fig. 2 – Aquileia, “Oratorio del Buon Pastore dall’Abito
Singolare”, late 4th-5th c. Detail of mosaic (L. Bertacchi,
Architettura e mosaico. Gli oratori privati, in Da Aquileia
a Venezia, Milan, 1980, tav. 21).
contains one of the earliest (1st decades of the 4th century)
Christian basilicas in the West, whose rich mosaic pavement
includes a wide-range of seemingly “secular” imagery,
from ¿shing to bird and animal life, as well as explicitly
Christian scenes7. The Christian content of the so-called
oratory pavements was thus established by comparison
with the basilica pavements, while the presence of such
oratories in wealthy homes was presumed to represent the
continuity of house-based communities alongside or even
preceding the public basilica. Indeed, the dating of these
mosaics is largely based on a presumed pre-Nicene/early
Constantinian religious context of domestic worship. At
Aquileia and elsewhere, Christian domestic images were
assumed to be markers of pre-Nicene Christian commu-
nities, part of a pervasive public/private binary in which
7. See among a large bibliography G. C. Menis, I mosaici cristiani di
Aquileia, Udine, 1965; L. Bertacchi, Architettura e mosaico, cit. (n. 5),
pp. 97-336; Id., Basilica, museo e scavi - Aquileia, Rome, 1994.
CHRISTIAN IMAGES IN THE HOME 173
the home was the sole preserve of pre-Nicene Christianity,
while post-Nicene activity was located in public basilicas8.
Similar historiographic temptations have long been at
work in Ostia, with its close ties to Rome, its rich corpus
of domestic architecture, and its presumed early and
numerous Christian communities9. It also provides several
useful lessons for distinguishing a house from other sorts of
buildings. The “Edi¿cio con Opus Sectile,” located outside
the Porta Marina, encapsulates the problem (¿J)10. The
building varies from a typical peristyle domus: set at the
terminus of the decumanus, which literally opens into one
of the house’s reception rooms, the large L-shaped complex
consisted of a raised portico Àanked by rooms on three
sides. Probable western and southern wings were seemingly
destroyed – had they existed the plan would have seemed
more house-like. The extant northern wing contained
the large hall with opus sectile decoration, including a
nimbate ¿gure with long beard and hair, his hand raised
in the two-¿ngered sign of either speech or blessing. The
¿gure seems an astonishing pre¿guration of the bearded
Christ maestas domini of the 5th and 6th centuries and was
identi¿ed as such by its excavator, Giovanni Becatti, who
interpreted the complex as a Christian collegium. Alfred
Frazer countered that Ostia’s collegia must be have been
defunct by the later 4th century and claimed the building
was a house, as has Federico Guidobaldi, while Paul Zanker
contested the Christ identi¿cation, noting similarities with
contemporary images of philosophers, and returned to the
notion of a public philosophical school11.
The domestic identi¿cation is just as thorny as the
identi¿cation of the putative Christ, and hinges around
the absence of bedrooms and other “private quarters”,
the very public access to the building, and the fact that as
much as a third of the complex may be missing. On the
other hand, new work has suggested a far more vibrant
commercial and social Ostia than heretofore suspected,
and thus the notion of a 4th century collegium cannot be
dismissed out of hand12. The controversies surrounding
the “Edi¿cio con Opus Sectile” throw into sharp relief
the dif¿culties of parsing a clear separation between
“domestic” and “public” or “collective” buildings in all
Roman periods, and the high stakes involved in such
8. C.f. K. Bowes, Private Worship, Public Values and Religious Change in
Late Antiquity, Cambridge, 2008.
9. See an overview in F. Bauer, Stadtbild und Heiligenlegende. Die
Christianisierung Ostias in der spätantiken Gedankenwelt, in G. Brands
and H. G. Severin (dir.), Die spätantike Stadt und ihre Christianisierung,
Wiesbaden, 2003, pp. 43-62.
10. The excavation report is G. Becatti, Scavi di Ostia 6. Edi¿cio con opus
sectile fuori Porta Marina, Rome, 1969.
11. A. Frazer, A Critical Review, in AJA, 75, 1971, pp. 319-324; F. Guidobaldi,
La lussuosa aula presso Porta Marina a Ostia. La decorazione in opus
sectile dell’aula, in S. Ensoli and E. La Rocca (dir.), Aurea Roma. Dalla
città pagana alla città cristiana, 2001, pp. 251-262; P. Zanker, The Mask of
Socrates: The Image of the Intellectual in Antiquity, Berkeley, 1995, p. 316.
12. D. Boin, The Late Antique Landscape of Roman Ostia, forthcoming.
174 KIM BOWES
Fig. 3 – Ostia, “Edificio con Opus Sectile”, late 4th c. a/ Plan; b/ Detail of so-called “Christ” panel (G. Becatti, Scavi di Ostia.
Edificio con opus sectile fuori Porta Marina, Rome, 1969, pianta 1; tav. col. 1, respectively).
attributions – Christian club, Christian mansion, pagan-
revivalist strong-hold – when putative Christian images
are involved.
While listing the troubles that confront a study of
domestic Christian images, we should not neglect its equally
troubled textual corpus. The textual sources mentioning
such images have typically been plumbed as though they
were mere descriptions. Rather, the most oft-cited sources
– Theodoret’s description of Roman shopkeepers with
images of Simeon Stylites, John Chrysostom’s mention
of Christian images and crosses proliferating in Christian
homes, the Vita Symeonis’ miracles before domestic images
of the saint – form a subtle but important trope13. The
ubiquity of Christian triumph is described through a public/
private binary – the public church or holy man, surrounded
by a constellation of Christian homes and families14. The
tableau of the saint’s image or the Christian cross in the
house, attended upon by the Christian family or individual
makes concrete the physical permeation of Christian belief,
or the fame of a particular holy man, into society’s smallest
unit – the home. These carefully constructed images of
ubiquity should not be taken to describe real ubiquity.
These dif¿culties in parsing our subject matter need not
only point in negative directions. The dif¿culties distin-
guishing the ‘Christian’ from the ‘pagan’, the house from
13. For Theodoret see above; see also John Chrysostom, Mathaeum
Homilia, 54.4; Id., Contra Judaeos et Gentiles, 9; Vita Symeonis
Stylitae, 188 (ed. van den Ven).
14. On this issue speci¿cally in John Chrysostom’s Antioch writings, see
P. Brown, The Body and Society. Men, Women and Sexual Renunciation
in Early Christianity, New York, 1988, p. 307ff.
AnTard, 1 9, 2011
other spaces suggests a similarly ambiguous ancient world, in
which “Christian” and ‘domestic’were not the most important,
or at least not the most ready, categories in which people
consumed images. Many of the houses containing Christian
images discussed here also displayed ‘pagan’ imagery – the
Rape of Europa and Bellerophon at the Lullingstone villa with
its chapel, or even erotic mythological love scenes in Villa
Fortunatus with its blatantly Christian dominus15. Sometimes
the pagan images were older, retained in the more recent
‘Christian’ phases, other times they were contemporary but
in each case, pagan and Christian images were both visible.
Scholars typically explain such pairings as evidence for the
‘secularization’ of pagan imagery and its consumption by
Christians merely aesthetic or antiquarian, or as evidence for
syncretic beliefs16. But as Christian authors, particularly of
the 4th and 5th centuries make clear, there were a host of ways
of reading ‘pagan’ texts and images as a Christian – as useful
negative exempla, as a muse to inspire Christian hermeneutic,
as aesthetically beautiful objects17. The pairing of ‘Christian’
15. On Lullingstone, see below; on Villa Fortunatus, see below and
M. Guardia Pons, Los mosaicos de la antigüedad tardía en Hispania,
Barcelona, 1992, pp. 91-96.
16. Among a large bibliography, see J. Huskinson, Some pagan mythological
¿gures and their signi¿cance in Early Christian art, in PBSR, 42, 1974,
pp. 68-86; J. Engemann, Deutung und Bedeutung frühchristlicher
Bildwerke, Darmstadt, 1997; Id., Nichtchristliche und christliche
Ikonographie, in Konstantin der Große. Katalog Handbuch. Begleitband
zur gleichnamigen Ausstellung in Trier, Mainz, 2007, pp. 281-294; and
now A. Cameron, The Last Pagans of Rome, Oxford, 2011.
17. C.f. L. James, “Pray not to fall into temptation and be on your guard”:
Pagan statues in Christian Constantinople, in Gesta, 35.1, 1996, pp. 12-
20; S. Bassett, The Urban Image of Late Antique Constantinople,
Cambridge, 2004; T. M. Kristensen, Embodied images: Christian
AnTard, 1 9 , 2 0 11 CHRISTIAN IMAGES IN THE HOME 175

and ‘pagan’ images in the domestic sphere again questions concomitant rituals in those spaces22. Conversely, the
an easy de¿nition of ‘Christian’ domestic images. discovery of Christian images in a house often results in
Lurking behind many of these problems is the question the suggestion that the space in question functioned as a
of how images were viewed in the domestic sphere. As has church or oratory. Of late, archaeologists have been more
been long noted, all religious groups in the late ancient cautious in their assumptions, but the tendency remains
world shared a common image vocabulary.18 The distinction and reveals yet another important scholarly assumption:
between a ‘pagan’ Good Shepherd and a ‘Christian’ Good Christian images don’t only indicate religious af¿liation
Shepherd, long discussed by art historians19, was most they also indicate ritual practice. Belief and practice are
often made by the viewers themselves and the herme- thus often assumed to be inextricably linked, far more so
neutic baggage each viewer carried with them, that is, their than in our interpretation of Greco-Roman paganism.
intent to view subjects in certain ways20. This intentionality To my knowledge, only two sites preserve Christian
would have been particularly present in homes – homes images in a domestic space probably used for Christian
that were themselves the product of individual families and ritual – the villa of Lullingstone, Britain and a townhouse
their tastes. At the same time, homes were public spaces, in Alexandria, Egypt. At Lullingstone, the much-
open to a much wider set of audiences, and thus viewers, discussed room is a small church built into the villa’s
than the modern term ‘private’ conveys21. Images in the
late ancient house were produced by highly individua-
lized choices, yet were also consumed by a wide range of
viewers on whom the householders might be expected to
push their own intentions and interpretations. This complex
of viewers would have been particularly acute in late
antiquity as the religious perspectives of both family and
visitors multiplied. As will be discussed throughout this
essay, the need to cater to individual and general viewers
often produced imagery that resists modern interpretative
categories and easy interpretation.

II. IMAGES, RITUAL AND THE HOME

The presence of religious imagery in traditional

Greco-Roman houses – from the sacro-idyllic landscapes Fig. 4 – Lullingstone villa, plan, with chapel indicated in grey,
of Boscoreale to the Bacchic themes in the villa of late 4th c. (D. Neal, Lullingstone Roman Villa, London, 1991).
Piazza Armerina – is rarely interpreted as evidence for

Fig. 5 – Lullingstone villa, reconstruction of vestibule

and west end of chapel (G. W. Meates,
Lullingstone Roman Villa, London, 1955, fig. 3b)
response and destruction in late antique Egypt, in Journal of Late
Antiquity, 2.2, 2009, pp. 224-250.
18. A. Grabar, Early Christian Art, A Study of its Origins, Princeton, 1968,
remains fundamental.
19. E.g. T. Klauser, Studien zur Entstehungsgeschichte der
christlichen Kunst I, in JAChr, 1, 1958, pp. 20-51; Id., Studien zur
Entstehungsgeschichte der christlichen Kunst III. Schafträger und
Orans als Vergegenwärtigung einer populären Zweitugendethik auf
Sarkophagen der Kaiserzeit, in JAChr, 3, 1960, pp. 112-133; Id., Studien
zur Entstehungsgeschichte der christlichen Kunst VIII. Vorbemerkungen
zu abschließenden Untersuchungen über das Schafträger-Motiv, in
JAChr, 8-9, 1965-1966, pp. 126-170; cf. H. Brandenburg, Das Ende der
antiken Sarkophagkunst in Rom. Pagane und christliche Sarkophage im
4. Jahrhundert, in G. Koch (dir.), Akten des Symposiums “frühchristliche
Sarkophage” (Marburg 1999), Mainz am Rhein, 2002, pp. 19-39.
20. C.f. B. Ewald, Rollenbilder und Geschlechterverhältnis in der
römischen Grabkunst, in N. Sojc (dir.), Archäologische’ Anmerkungen
zur Geschichte der Sexualität, Berlin, 2005, pp. 71-72.
21. C.f. A. Wallace-Hadrill, Houses and Society in Pompei and Herculaneum,
Princeton, 1994; A. Riggsby, ‘Public’ and ‘private’ in Roman culture: the 22. The exception seems to be late Roman Romano-British mosaics, for
case of the cubiculum, in JRA, 10, 1997, pp. 36-56. which see below.
176 KIM BOWES
Fig. 6 – Lullingstone villa, chapel west wall frescoes (© Trustees of the British Museum).
north side in the later years of the 4th century23 (¿JV).
Entering from the outside (another, secondary entrance
was probably through the villa itself) the visitor would
pass through two vestibules, in the second of which s/
he would be immediately faced with a large chi-rho. The
symbol was repeated on the south wall of the church itself,
visible from the vestibule and thus guiding the viewer into
the complex. The main church room was covered with
paintings, all but the western wall of which were heavily
damaged. The western wall paintings, like the chi-rhos,
respond to the viewer’s own experience: depicting a
group of well-dressed worshippers, their arms raised in
prayer, the painted congregation imitated the actions of
the real one (¿J ). The church’s side walls contained
¿gural scenes, but the fragmentary remains could not be
made to yield any of the stock Old and New Testament
imagery of the period, both because of their poor state of
preservation and because it is very likely that such scenes
were not depicted. Rather the room seems to have been
covered with a more eclectic choice of Christian imagery,
seemingly of the villa-owner’s own making.
23. The main publications include G. W. Meates, Lullingstone Roman Villa,
London, 1955; Id. (dir.), The Roman Villa at Lullingstone, Kent, 1. The
Site, Maidstone, Kent, 1979; G. W. Meates et al. (dir.), The Roman Villa
at Lullingstone, Kent, 2. The Wall Paintings and Finds, Maidstone, Kent,
1987; D. Neal, Lullingstone Roman Villa, London, 1991.
AnTard, 1 9, 2011
Jumping forward to the 6th century and east to
Alexandria, a combination workshop/residence located
in the city center was organized around a long, narrow
courtyard (16 x 3 m) (¿J )24. Along the south wall of
the courtyard at its mid-point was painted a large (1.5 m
preserved, possibly 2 m total width) image of the Virgin
enthroned holding an infant Christ, Àanked by an angel
and donor or saint on the left, a tableau which may have
been repeated on the right (¿J ). On the other side of
the courtyard were found two iron brackets, presumably
for lamps, whose placement exactly framed the painting
opposite. The excavators thus concluded that the Virgin
and Child image were the focus of veneration by the
occupants of the house.
These two instance of permanent images used in
ritual contexts have little in common. Lullingstone is an
eccentric, very personal monument: one of a handful of
Christian buildings from late antique Britain and tucked
away in the countryside of Kent, it bespeaks the personal
religiosity of a devout individual or family. In Britain of
the later 4th century, the chi-rho seems to have been parti-
cularly potent sign of not only Christian af¿liation but
24. M. Rodziewicz, Les habitations romaines tardives d’Alexandrie à
la lumière des fouilles polonaises à Kôm el-Dikka (Alexandrie, 3),
Warsaw, 1984, esp. pp. 67ff, 194-208.
AnTard, 1 9 , 2 0 11 CHRISTIAN IMAGES IN THE HOME 177

also of a trendy Romanitas25. Its prominent, repeated use One the one hand, there seems little reason to doubt
at Lullingstone to guide a visitor into the villa church, that devout Christians prayed before images in the home
combined with the undecipherable narrative paintings, and that those images took many forms. The Alexandria
loudly proclaimed the individuality and piety of its fresco seems a culmination of a practice ultimately rooted
patron. The over-large painting of the virgin and child in in pagan domestic shrines, statuettes and panel ‘icons’, and
Alexandria, on the other hand, references the world of whose Christian expression can now only be sensed from
public Christianity. So imposing is the image and icono- out-of-situ portable objects like pilgrimage ampullae bearing
graphy for its small space that the excavators posited, images of Minas or Symeon, or portable reliquaries carrying
probably rightly, that it copied the apse decoration of a images like the Sancta Sanctorum casket. On the other
local Alexandrian church26. By the 6th century, the images hand, some caution might be in order in assuming that such
that adorned the great public basilicas had come to form a veneration necessarily moved from house to basilica. For
canon of Christian images. As is indicated by the imagery
from pilgrimage ampullae, the apse decoration and other
imagery from famous public basilicas might be copied and
widely disseminated for use in more personal religious
rites27. The images from public basilicas, in other words,
assumed a holy power and were thus subject to appro-
priation in the private sphere.
That the reverse trend might also be true is suggested by
the history of icon worship. Textual sources, like the above-
mentioned Chrysostom encomium to bishop Meletios of
Antioch or the Life of Symeon Stylites the Younger, seem
to place the earliest documented veneration of holy images
in the domestic sphere. Some scholars have taken this to
mean that icon worship began as a private domestic rite,
perhaps as a development of pagan practice, and from there
moved to the public basilicas28. Judith Herrin has even
credited women with the innovation, noting that many of Fig. 7 – Alexandria, Kôm el-Dikka, House D, plan, 6th c.
the protagonists in these stories are women, although her (M. Rodziewicz, Remarks on the domestic and monastic architec-
conclusions have been hotly contested29. The only in-situ ture in Alexandria and surroundings, in The Archaeology of the
Nile Delta. Problems and Priorities [Proceedings of the Seminar
archaeological evidence for domestic image worship is the held in Cairo 19-22, October 1986], Amsterdam, 1988, fig. 1).
Alexandria fresco, for portable wooden images, possibly
the most common medium for such objects, do not readily
survive in archaeological contexts. We are thus reliant on
the textual evidence, the problems with which as factual
Fig. 8 – Alexandria, Kôm el-Dikka, House D, reconstruction
witnesses have been outlined above. drawing of fresco of Virgin and Child from courtyard, south
wall (M. Rodziewicz, Les habitations romaines tardives
d’Alexandrie à la lumière des fouilles polonaises à Kôm
el-Dikka, Warsaw, 1984, fig. 236).

25. D. Mattingly, An Imperial Possession: Britain in the Roman Empire,

London, 2006, p. 465.
26. Ibid., p. 203
27. K. Weitzmann, Loca sancta and the representational arts of Palestine,
in DOP, 28, 1974, pp. 31-55; c.f. G. Vikan, Early Byzantine pilgrimage
devotionalia, in E. Dassman and J. Engemann (dir.), Akten des
XII. internationalen Kongresses für Christliche Archäologie, Bonn,
1995, pp. 377-388; C. Hahn, Loca sancta souvenirs: Sealing the pilgrim’s
experience, in R. Ousterhout (dir.), The Blessings of Pilgrimage, Urbana /
Chicago, 1990, pp. 85-96.
28. J. Herrin, Women and faith, cit. (n. 4); T. Mathews, Clash of the Gods:
A Reinterpretation of Early Christian Art, Princeton, 2nd ed., 1999,
pp. 177-190; T. Mathews and N. Muller, Isis and Mary in early Icons,
in M. Vassilaki (dir.), Images of the Mother of God: Perceptions of the
Theotokos in Byzantium, Ashgate, 2005, pp. 3-12.
29. J. Herrin, Women and faith, cit. (n. 4); c.f. R. Cormack, Woman
and icons, cit. (n. 4); Pitarakis, Female piety in context, cit. (n. 4);
L. Brubaker, Image, audience and place: Interaction and reproduction,
in R. Ousterhout and L. Brubaker (dir.), Sacred Image East and West,
Urbana / Chicago, 1995, pp. 204-220.
178 KIM BOWES
one, the scholarly discussion has focused on portable images,
thus reÀecting a continued tendency to restrict conceptions
of the ‘icon’ to wooden panel paintings and other portable
objects. Late ancient Christians recognized no such distinc-
tions and understood fresco and mosaic portraits in churches,
panel paintings, and other kinds of images as equivalent, all
reproductions of holy persons that might serve as agents for
epiphany and connective tissue with the divine30. Worship
and miracles before this broader class of images appears in
our sources in both church and home more or less at the same
time, that is, during the later ¿fth and sixth centuries in the
eastern empire. We should probably thus understand the use
of devotional images in the home as part of the proliferation
of Christian imagery of all kinds during this period31.
III. PROTECTING THE HOME
Domestic rituals took many forms; Eucharistic rites and
prayers occur to us ¿rst because they have a collective, public
church counterpart, but other rituals were doubtless more
common, and for the occupants of the house, more critical.
Rites of protection – to guard against evil spirits, the evil eye,
and other divine bringers of catastrophes big and small –
were as critical to the fabric of the house as were its walls and
doors. Scholars have often classed these rites as ‘magical’,
a term that, for a variety of reasons is less than helpful32.
‘Magic’ in the ancient world, and even in the eyes of early
Christian writers, was not so much a particular type of ritual
as a moral container in which to put suspect ritual33. Legally,
it was restricted to rituals meant to inÀict harm or provide
its recipient with an unfair advantage – in love, in court, in
the ¿elds – while rhetorically it was used to lambast private
rituals, nocturnal rituals, and rituals performed by women
30. G. Ladner, The concept of the image in the Greek fathers and the
Byzantine iconoclastic controversy, in DOP, 7, 1953, pp. 1-34;
H. Belting, Likeness and Presence: A History of the Image in the Era
Before Art, Chicago / London, 1994; R. Cormack, Writing in Gold:
Byzantine Society and its Icons, London, 1985.
31. C.f. G. Dagron, L’Église et la chrétienté byzantines entre les invasions
et l’iconolasme (VIIe-début VIIIe siècle), in G. Dagron, P. Riché and
A. Vauchez (dir.), Évêques, moines et empereurs (610-1054), Paris,
1993, pp. 9-91, who places the phenomenon as part of a larger
“privatization” of ritual in the later 6th-7th c.
32. E.g. H. Maguire, Magic and geometry, cit. (n. 4); Id., Garments
pleasing to God, cit. (n. 4); E. Maguire, H. Maguire and M. Duncan-
Flowers, Art and Holy Powers, cit. (n. 4).
33. Among a vast bibliography, C. R. Phillips, Nullum crimen sine lege:
Socioreligous sanctions on magic, in C. Faraone and D. Obbink (dir.),
Magika Hiera. Ancient Greek Magic and Religion, Oxford, 1991,
pp. 260-276; H. Kippenberg, Magic in Roman civil discourse: Why
rituals could be illegal, in P. Schäfer and H. Kippenberg (dir.),
Envisioning Magic. A Princeton Seminar and Symposium, Leiden,
1997, pp. 137-163; D. Grodzynski, ‘Superstitio’, in RÉA, 76, 1974,
pp. 36-60; P. Brown, Sorcery, demons and the rise of Christianity, in
M. Douglas (dir.), Witchcraft Confessions and Accusations, London,
1970, pp. 17-45.
AnTard, 1 9, 2011
or slaves. Thus, rituals and ritual objects used to protect the
home – the phallus over the door of the Pompeiian house, or
as we shall see, the cross over the door of a Syrian farm –
might be classed as ‘magic’ by some or as pious gestures
by others. To term such practices ‘magical’ also arti¿cially
separates them from other kinds of practices which might
have institutional counterparts, but to which they are functio-
nally and structurally homologous34. To light lamps and pray
before the image of the Virgin and Child in an Alexandrian
home has the same ritual intention as carving an image of
the cross, a Solomon’s knot and a Psalm over one’s doorway.
The in-situ material evidence for the protective use of
images is spotty, but has some interesting characteristics,
particularly in its iconography and domestic topography.
The ‘images’ most readily identi¿able as protective are
actually symbols – the cross, the chi-rho, or other older
symbols such as Solomon’s Knot or the swastika. These
symbols typically appear at window and door lintels or
jambs, or possibly in pavements near doorways. Indeed, it
is in part from these liminal locations that their protective
functions are evident: the doors and windows of buildings
are its ori¿ces, its contact points with the outside world
and thus the weakest point in its spiritual defenses where
prying eyes, disease, and ill-intensions might seep in.
For instance, a set of four stone window grilles from the
so-called Triconch Palace in Butrint, Albania, included
two that contained the chi-rho motif embedded in the
lunette (¿J)35. These windows originally stood high on
the inner wall of the triconch dining room, made invisible
to visitors from the exterior by the covering wall of the
portico below (¿J). While they could have been seen
once one entered the hall, it seems they were principally
oriented towards an invisible viewership, the daemones
against which they protected.
The largest corpus of such protective symbols comes
from Syrian villages of the northern Massif36. To a large
34. N. Janowitz, Icons of Power. Ritual Practices in Late Antiquity,
University Park, 2002; D. Frankfurter, Beyond magic and superstition,
in V. Burrus (dir.), Late Ancient Christianity. A People’s History of
Christianity, 2, Minneapolis, 2005, pp. 255-284. From earlier periods,
cf. F. Graf, Prayer in magic and religious ritual, in C. Faraone and
D. Obbink (dir.), Magika Hiera. Ancient Greek Magic and Religion,
Oxford, 1991, pp. 188-213; Id., How to cope with a dif¿cult life: A view
of ancient magic, in P. Schäfer and H. Kippenberg (dir.), Envisioning
Magic. A Princeton Seminar and Symposium, Leiden, 1997, pp. 93-114.
35. W. Bowden, K. Francis, O. Gilkes and K. Lako, The domus and the
Triconch Palace, in W. Bowden and R. Hodges (dir.), Excavations in
the Triconch Palace, Oxford, 2011, p. 41.
36. For a summary, see E. Maguire, H. Maguire and M. Duncan-Flowers,
Art and Holy Power, cit. (n. 4). On the villages and their houses more
generally, see G. Tchalenko, Villages antiques de la Syrie du Nord,
Paris, 1953-1959; G. Tate, Les campagnes de la Syrie du Nord du IIe au
VIIe siècle : un exemple d’expansion démographique et économique
à la ¿n de l’Antiquité, Paris, 1992; and chapters in C. Castel, M. Al-
Maqdissi and F. Villeneuve (dir.), Les maisons dans la Syrie antique
du IIIe millénaire aux débuts de l’Islam (Bibliothèque Archéologique et
Historique, 150), Beirut, 1997.
AnTard, 1 9 , 2 0 11
degree, this is a result of the particular material and
preservation of these houses: Syrian village houses were
built almost entirely of ashlar masonry, and the protective
images and inscriptions were carved into the stone itself.
Since so many of these villages lie more or less intact,
CHRISTIAN IMAGES IN THE HOME 179
image and/or inscription can be often paired with a certain
or likely domestic context. Unfortunately, very few of
these villages have been subject to modern excavation,
and the evidence is recorded only in the various corpuses
of inscriptions made by foreign missions in the early
Fig. 9 – Butrint, “Triconch Palace”, window
grills, early 5th c. (W. Bowden et al., The domus
and the Triconch Palace, in W. Bowden and
R. Hodges (dir.), Excavations in the Triconch
Palace, Oxford, 2011, fig. 2.55).
Fig. 10 – Butrint, “Triconch Palace”, computer
reconstruction of entrance from peristyle,
showing window grills in situ (W. Bowden et al.,
The domus and the Triconch Palace,
in W. Bowden and R. Hodges (dir.), Excavations
in the Triconch Palace, Oxford, 2011, fig. 2.54).
180 KIM BOWES
Fig. 11 – Anasartha, drawing of lintel inscription, early 7th c.
(H. Maguire, Magic and Geometry in Early Christian Textiles,
in Jahrbuch der österreichischen Byzantinistik, 44, 1994, fig. 6).
20th century37. Often, these record only the text of the
inscription, not its context, although sometimes a brief
assessment of location – house, tomb, tower – appears. An
overview of Prentice’s catalogue of the Greek inscriptions,
which although not the most plentiful corpus nonetheless
provides some helpful contextual information, indicates
the ubiquity of the Syrian phenomenon38.
The images consist typically of crosses, chi-rho’s and
other symbols, which are often used to frame an inscription.
While the texts obviously attracted the attention of the
epigraphers, thus biasing what evidence was recorded,
the unity of image and text in these Syrian houses is a
useful reminder of how words and images might work in
concert, particularly in protective rites. One inscription
from Anasartha states this as fact: “+ When we engrave
your cross, O Christ the God...we escape every form of
wickedness”39 (¿J). The cross, the act of engraving the
cross, and the reformulation of Psalm 56:7 here all perform
the same work, thus reduplicating the protective intent40.
In general, the inscriptions praise God, and call on
Christ’s help for protection from evil for the house’s
denizens. In il-Bârah, for instance, three houses bore the
same inscription from Psalm 121.8: “The Lord shall preserve
thy coming and going, from this time forth for evermore.
Amen.41” Similar to the protectants against the evil eye and
against jealousy is the call for just treatment at the hands
of one’s neighbors. At KƗtûrƗ, inscription on a house dated
37. E.g. L. Jalabert, Inscriptions grecques et latines de la Syrie, Paris,
1929- (IGLS); W. Prentice, Greek and Latin Inscriptions (Part III of the
Publications of an American Archaeological Expedition to Syria, 1899-
1900), New York, 1908; E. Littmann, Semitic Inscriptions (Publications
of an American Archaeological Expedition to Syria in 1899-1900),
New York, 1904.
38. Of c. 300 inscriptions in the 1908 volume, some 30 are listed as coming
from de¿nite or likely domestic contexts.
39. IGLS 298.
40. On the agency of making in the ef¿cacy of ritual, see A. Gell, Art and
Agency: Towards a New Anthropological Theory, Oxford, 1998.
41. W. Prentice, Greek and Latin Inscriptions, cit. (n. 37), nos. 192-194.
AnTard, 1 9, 2011
to 336, reads: “Christ of God, Help (us)! (There is) one
God only. Thalasis built (this). Whatever you say, friend,
(may that be) the same to you twofold. In the year 385.
Enter, O Christ.42” The phrase “whatever you say may that
also be to you” seems a paraphrase of Matthew 7:12 but
the sentiment is echoed in both the Old Testament and by
various pagan writers. In a Syrian domestic context, the
identical phrasing also appears on a pagan house in which
the inhabitants call upon Zeus to guarantee its ef¿cacy43.
While there is no doubt that this use of Christian texts
and images to protect houses extended beyond Syria, the size
and cohesion of the Syrian corpus calls for some speci¿c
comment. Of particular importance is the context of these
protective image/texts – namely, the rural village. Pre-modern
rural villages, as Le Roy Ladurie’s study of Montaillou vividly
describes, were closely-packed small worlds which, although
they might show a uni¿ed front to outsiders, were rife with
internal tensions – jealousy, petty slights, and conÀict over
limited resources44. In late antique Syria, the social tensions
between landowners and farmers, city merchants and
locals, even husbands and wives could boil over into public
quarrels45. In these villages of the northern Massif, houses
are packed cheek to jowl, and while internal courtyards and
occasional precinct walls screened some activities from
view, the houses’ many windows, and multi-story arrange-
ments would have meant constant surveillance from one’s
neighbors46. The vulnerability of the house to prying eyes and
thus to evil intensions, envy and neighborly malice rings out
42. Ibid., no. 116.
43. Ibid., no. 114.
44. E. Le Roy Ladurie, Montaillou, village occitan de 1294 à 1324, Paris,
1975.
45. P. Brown, The Rise and Function of the Holy Man in Late Antiquity, in
JRS, 61, 1971, pp. 80-101.
46. Cf. the excavations at Déhès: J.-P. Sodini et al., Déhès (Syrie du
Nord). Campagnes I-III (1976-1978). Recherches sur l’habitat rural,
in Syria, 57.1, 1980, pp. 1-181; 183-301; 303-304; cf. A. Sitz, Déhès: A
Late Antique Village in Context, Unpublished seminar paper. Thanks to
the author for sharing this prior to publication.
AnTard, 1 9 , 2 0 11
from these inscriptions, placed, as they were, at the very doors
and windows that admitted evil. Christian texts and images
would thus have played a particularly important protective
role in these rural village contexts.
The combined use of both clearly Christian symbols
– the chi-rho and the cross – with other kinds of symbols
– the circle, the rosette, the swastika – in many of the
Syrian house inscriptions raises the question of how
other geometric patterns found in houses should be
read. Domestic mosaic pavements, such as the House of
Eustolios in Kourion (discussed more fully below) often
included so-called Solomon’s Knots, swastikas and other
symbols with possible apotropaic properties. As the
fragmentary remains of Egyptian textiles indicates, the
fabrics that served as doors, window dressings, and wall
decorations would also have been covered with interlaced
circles, simple crosses and other symbols that in other
contexts might have magical content. Henry Maguire
has emphasized the power and ubiquity of this symbolic
language in the domestic sphere, and suggested that the
very multiplication of these signs, rather than diluting their
signi¿cation actually increased their protective powers47.
Yet as he also admits, sometimes a knot is just a knot, that
is, symbols might have a plurality of meanings, or, given
their ubiquity as part of decorative motifs, no meaning at
all. How are we to know the difference, or more impor-
tantly, how did ancient viewers?
The question raises again the problem of viewing
inside the ancient house and how, precisely, such images
were thought to work. Was it enough that the symbol be
placed in its desired place, that is, was the principal agent
the intention of its patron/maker, or was it necessary that
other viewers understood its meaning48? Did a rosette or
Solomon’s Knot on a doorframe do its work simply by
virtue of an intention-¿lled act of making, or did visitors
also need to recognize it as potent symbol? The evidence
supports both scenarios. The Syrian houses insistently
(although not always) use text – typically quotations from
the Psalms or simple invocations of Christ to help – to
make precise the protective meaning of crosses and other
symbols through a ritual of protection that addressed both
outside viewers as well as the invisible world. Likewise,
the Solomon’s knots that appear in the House of Eustolios
surround an inscription that reads “Enter to your good
fortune and may your coming bless the house49.” The
protective function of otherwise ambiguous images was
made speci¿c to outsiders through textual framing. Yet in
47. H. Maguire, Magic and geometry, cit. (n. 4); Id., Garments pleasing to
God, cit. (n. 4); E. Maguire, H. Maguire and M. Duncan-Flowers, Art
and Holy Powers, cit. (n. 4).
48. Cf. A. Gell, Art and Agency, cit. (n. 40), who tends to assume the ritual
ef¿cacy of making objects lies in its conscious making and knowing
reception.
49. See n. 70 below.
CHRISTIAN IMAGES IN THE HOME 181
most other domestic contexts, the knot, the swastika, the
spiral all appear without comment, heavy or empty with
meaning depending on the viewer’s perspective. In these
cases, their intention is irrecoverably lost to us. Were they
meant to act protectively, it would seem that, in these
instances, it was enough that the homeowner caused them
to made and placed at doors and windows for them to do
their protective work.
IV. CHRISTIAN IMAGES, STATUS AND IDENTITY
One of the few scholarly consensus about early Christian
images is that they are principally about religion, that is,
Christian images reÀect Christian piety, ritual practice,
hopes for the afterlife, or connections to the divine. Christian
images in the home, however, demand to be examined from
other perspectives. The Roman house was a machine for
the production and maintenance of status, crafting and
manipulating a particular image of self and family through
domestic space and imagery50. Late antique houses were
still very much Roman in this sense, and performed the
same self-fashioning on behalf of their owners as those
of Pompeii had done centuries earlier51. Some Christian
images, when examined within their domestic and/or local
contexts, seem principally intended to make claims about
the homeowner’s learnedness and power. Perhaps surpri-
singly, proclaiming the homeowner’s Christian af¿liation
is not always central, but rather these Christian images
assume a variety of other roles, from discursive agents to
symbols of seigniorial power.
Some of the best candidates for such an interpretation
might also seem the most unlikely, for they have been
typically plumbed for complex cultic readings. The mosaic
Àoors unearthed in two Romano-British villas, at Hinton St.
Mary and Frampton, both in Dorset and both dating to the
middle of the 4th century, contain the only Christian imagery
found on domestic Àoor mosaics. Neither villa was ever
fully excavated, and the Frampton mosaics are preserved
only in an 18th century drawing. The Hinton St. Mary
pavement is divided into two unequal portions separated
by two short walls, but seemingly laid as a piece (¿J)52.
50. E.g. S. Hales, The Roman House and Social Identity, Cambridge, 2003.
51. K. Bowes, Houses and Society in the Later Roman Empire, London,
2010.
52. Excavations reports and major articles on Hinton St. Mary: J. Toynbee,
The Christian Roman mosaic, Hinton St. Mary, Dorset, in Proceedings
of the Dorset Natural History and Archaeological Society, 85, 1963,
pp. 116-121; Id., A new roman mosaic pavement found in Dorset, in
JRS, 54, 1964, pp. 7-14; K. S. Painter, Excavation of the roman villa at
Hinton St. Mary, 1964, in Proceedings of the Dorset Natural History
and Archaeological Society, 86, 1964, pp. 150-154; Id., Excavation of
the roman villa at Hinton St. Mary, Dorset, 1965, in ibid., 87, 1965,
pp. 102-103; A. Payne, Hinton St. Mary, Dorset. Report on Geophysical
182 KIM BOWES
Fig. 12 – Hinton St. Mary villa, mosaic pavement, mid-4th c.
(© Trustees of the British Museum).
The smaller section depicts scenes of stags Àeeing dogs
and Bellerophon killing the chimaera. The larger portion is
organized around a central roundel depicting a young man
clad in a toga, his head framed by a chi-rho and Àanked
by two pomegranates (¿J ). Surrounding this central
images are semi-circular frames containing trees, more
stages hunted by dogs, while in the corners are four bust
portraits of men. The central image is generally, although
not universally, thought to be Christ, while the corner
images are variously identi¿ed as the personi¿cation of
the winds or the four evangelists. Further excavation and
geophysics has still left its domestic context uncertain, but
the fact that the covering ¿gural imagery left no undeco-
rated space left for couches, and that the imagery itself has
two different planed viewing points has argued against its
use as a dining room53.
The Frampton mosaic’s context is slightly better
understood: the villa’s main entrance was through a
main fronting corridor, whose two ends were capped by
two wings, or “end rooms”, a common arrangement in
Surveys, July 1996 (Unpublished report); J. Toynbee, Pagan motifs and
practices in Christian art and ritual in Roman Britain, in M. W. Barley
and R. P. C. Hanson (dir.), Christianity in Britain, 300-700, Leicester,
1967, pp. 177-192; T. Moorhead, The Hinton St Mary head of Christ
and a coin of Magnentius, in L. Webster and D. Williams (dir.), Image,
Craft and the Classical World, Montagnac, 2005, pp. 209-212.
53. P. Witts, Mosaics and room function: The evidence from some fourth-
century Romano-British villas, in Britannia, 31, 2000, p. 309; S. Cosh,
Seasonal dining-rooms in Romano-British houses, in Britannia, 32,
2001, p. 229.
AnTard, 1 9, 2011
Fig. 13 – Hinton St. Mary villa, mosaic pavement, detail
of “Christ” (© Trustees of the British Museum).
Romano-British villas (¿J )54. The north “end-room”
contained a mosaic depiction of Neptune, while the south
included a vestibule depicting Dionysius and his panther,
Àanked by images of men hunting stags, and separated
from the main apsed room by a boarder of peltae – a motif
that resembled an axe and shield and which has thus been
read as having apotropaic qualities (¿J). Once inside
the room the visitor pass over another mosaic boarder
containing an image of cupid and half of an inscription
(the other half is lost) reading, “…and no service unless
you will it, Cupid.55” Both cupid and the inscription are
oriented towards a viewer already in the main room. The
central roundel of the main room showed Bellerophon
atop Pegasus killing the chimaera, and was Àanked by
four scenes of mythological loves, three of which survive
– Venus and Adonis (or Selene and Endymion), Attis and
Sagaritis (or Paris and Oenone) and possibly Venus and
Cupid. Three bands of mosaic marked the cord of the apse:
the ¿rst contained an inscription, “the head of Neptune,
whose dark blue ¿gure is Àanked by two dolphins, is
54. Excavation reports and major articles: S. Lysons, Reliquiae Britannico-
Romanae III, London, 1817; R. A. H. Farrar, Archaeological ¿eldwork in
Dorset in 1956. The ‘Frampton Villa’, Maiden Newton, in Proceedings of
the Dorset Natural History and Archaeological Society, 78, 1956, pp. 81-
83; M. Henig, James Engleheart’s drawing of a mosaic at Frampton,
in ibid., 106, 1984, pp. 146; D. Perring, ‘Gnosticism’ in fourth-century
Britain: The Frampton mosaics reconsidered, in Britannia, 34, 2003,
pp. 97-127.
55. [Nec mu]nus per¿cis ullum / [si di]gnare cupido.
AnTard, 1 9 , 2 0 11
Fig. 14 – Frampton villa, plan, mid-4th c.
(S. Lysons, Reliquiae Britannico-Romanae III,
London, 1817).
allotted the domain stirred by the winds,56” the second
band contained the aforementioned mask of Neptune
with a band of dolphins emerging from his mouth, and
the third, on the chord, a row of leafy spirals framed the
chi-rho symbol. The apse itself was paved with geometric
motifs, including Solomon’s Knots, and a central image
of a kantharos, or two-handled cup.
These pavements have been subject to decades of
analysis, typically by “de-coding” the pavements’ individual
iconographic components using contemporary texts and
sarcophagi57. The Bellerophon image has been read as the
triumph over evil, the hunting of stags the struggle of the
soul, Dionysius as an image of immortality, the mytho-
logical lovers the relationship between humans and the
divine, all of which are overseen by the ultimate Christian
56. Neptuni vertex reg[i]men isortiti mobile ventis I scul[p]tum cui
c[a]erulea [est] / del¿nis cincta duob[us].
57. E.g. J. Toynbee, Pagan motifs and practices, cit. (n. 52); K. S. Painter,
Villas and Christianity in Roman Britain, in British Museum
Quarterly, 35, 1971, pp. 156-175; J. Huskinson, Some pagan
mythological ¿gures, cit. (n. 16); E. W. Black, The Roman Villas of
South-East England, Oxford, 1987; M. Henig, “Ita intellexit numine
inductus tuo”: some personal interpretations of deity in Roman
religion, in M. Henig and A. King (dir.), Pagan Gods and Shrines of the
Roman Empire, Oxford, 1986, pp. 159-170.
CHRISTIAN IMAGES IN THE HOME 183
Fig. 15 – Frampton villa, drawing of mosaic pavement
(S. Lysons, Reliquiae Britannico-Romanae III,
London, 1817).
triumph, represented by the Christ or the chi-rho, placed in a
central position. Scholars have thus read them as statements
on temptation and salvation, as representing “Gnostic”
theologies, or, on the contrary, as evidence for the Romano-
British elite’s “lukewarm” Christian commitments58.
Before endorsing any one of these interpretations, some
very basic facts might repay consideration. The two sets
of pavements have important commonalities that, I would
argue, encouraged comparison, not only by modern
scholars but also by their original patrons. Stylistic simila-
rities suggest they were laid by the same group of mosaic
craftsman, the so-called Durnovarian School, employed
by various villa owners around the region’s major town of
Dorchester59. The similarities are so great that some have
even suggested the same craftsman was at work60. They
contain the only two instances of a chi-rho in a domestic
mosaic pavement in Britain, and one of only three extant
58. E.g. J. Huskinson, Some pagan mythological ¿gures, cit. (n. 16);
D. Perring, ‘Gnosticism’ in fourth-century Britain: The Frampton mosaics
reconsidered, in Britannia, 34, 2003, pp. 97-127; W. H. C. Frend, Pagans,
Christians, and ‘the Barbarian Conspiracy’ of A. D. 367 in Roman
Britain, in Britannia, 23, 1992, pp. 121-131, respectively.
59. P. Johnson, Romano-British Mosaics, Buckingham, 1982, ch. 6.
60. Ibid.; D. Perring, ‘Gnosticism’, cit. (n. 58), p. 101.
184 KIM BOWES
in the Roman world. Finally, the two villas lie less than
30km apart. In short, it seems very likely that these two
very unusual projects were commissioned in dialogue with
one another, possibly in the context of competitive learned
exchange that lay at the heart of the creation of elite status.
In short, it seems likely that the viewing context for these
mosaics was one of comment, response, and dialogue.
If the modern analysis of these images has gone astray
it has been in assuming the presence of a single meaning
that only wants decoding to deduce its substance. Rather,
I would suggest that Frampton and Hinton St. Mary were
deliberately composed to resist straightforward analysis
and provoke debate between the dominus and his visitor-
peers61. A number of elements in their composition suggest
this: both sets of mosaics are composed around multiple
viewpoints a fact that has puzzled modern scholars who have
tried, unsuccessfully, to extract room function from viewing
perspective62. The mosaics’ numerous orientations (two at
Hinton St. Mary, three at Frampton) require the viewer to
move around to properly view the whole assemblage; they
are interactive by design. Second, while the inscriptions on
Syrian houses restrict meaning, the inscriptions at Frampton
61. Cf. S. Scott, Art and Society in Fourth-Century Britain: Villa Mosaics
in Context, Oxford, 2000; D. Mattingly, An Imperial Possession, cit.
(n. 25); R. Ling, Mosaics in Roman Britain: Discoveries and research
since 1945, in Britannia, 28, 1997, p. 278, all of whom have read the
mosaics as attempts to display erudition and learnedness.
62. P. Witts, Mosaics and room function, cit. (n. 53); S. Cosh, Seasonal
dining-rooms, cit. (n. 53).
AnTard, 1 9, 2011
Fig. 16 – “Villa Fortunatus”, plan, late 4th c.
with location of Fortunatus mosaic (M)
indicated (M. Guardia Pons, Los Mosaicos
de la Antigüedad Tardía en Hispania,
Barcelona, 1992, fig. 7).
are teases. The fragmentary couplet on Cupid might be read
as lighthearted commentary on the power of love, even over
the servants, while the Neptune inscription might represent
an ekphrastic musing over the mosaic itself, but above all
they do not tell us what the mosaics mean. Rather, they
model the acts of looking and interpreting and encourage
viewers to do the same Finally, there is the very plurality
of images themselves: at Hinton St. Mary is combined
Bellerophon, Christ (?), Seasons/Winds and hunts; while
at Frampton the viewer was presented with Dionysius,
hunts, Cupid, Bellerophon again, mythological lovers,
Neptune, chi-rho. In selecting so many scenes and symbols
with so many possible signi¿ers, the patrons produced a
veritable tangle of meanings. Those who argue for syncre-
tistic or “Gnostic” meaning claim these multiple references
reÀect these groups’ tendency towards allegorical readings
and arcane connections. But the allegorical reading of a
Plotinus or Porphyry is teleological, as quotations from
Homer or Plato are insistently forced towards a single inter-
pretive end – the illumination of the Good, the ¿nal ascent
of the soul. Very much unlike the texts that have been used
to interpret them, these images, embedded in symmetrical
compositions around central motifs, fail to push inter-
pretation in one direction. Rather, their very composition
encourages open-ended connections around their central
images of Christ (?) and Bellerophon, respectively.
Promoting discourse was one road to status, but there
were more straightforward ways, ways in which Christian
af¿liation might play a simpler role. Two villas in Hispania
suggest variants on a process. At Villa Fortunatus, near
AnTard, 1 9 , 2 0 11
Fig. 17 – “Villa Fortunatus”, mosaic from peristyle exedra
(phot. courtesy of the Museo de Zaragoza).
Lleida, a large peristyle villa was remodeled during the
later 4th century to include an entryway audience hall, new
baths, and centered on one side of the peristyle, a raised
exedra, carved out from earlier cubicula (¿J )63. The
rectangular room seems to have acted as a kind of viewing
pavilion, possibly for dining, from which to look out
over the garden, but the mosaic was oriented to be seen
from that same garden. The pavement’s center is covered
with images of bounteous nature – fruit, legumes, ducks,
rabbits, vases sprouting vines, frolicking putti, while at the
head of an acanthus boarder appears the dominus’ name,
“FORTUNATUS,” bisected by a chi-rho with alpha and
omega (¿J)64. The chi-rho may be a sign that Fortunatus
is a Christian, but just as importantly, in the context of
seignorially-derived abundance, also a symbol of power
that enhances and speci¿es the landowner’s power to grant
prosperity. If the images of fruit baskets and small animals
are additionally interpreted as xenia, or host-gifts provided
to guests65, the chi-rho reinforces Fortunatus’ claims to
bene¿cence. Thus, the Christian symbol not only marks
religious af¿liation, but surrounded by images of plenty and
placed, literally, at the head of the villa, it also reinforces
particular aspects of seigniorial power66.
63. On the latest phases, see M. Guardia Pons, Los mosaicos, cit. (n. 15),
pp. 83-100; R. Navarro Sáez, Villa Fortunatus, in Del Romà al Romànic,
Barcelona, 1999, pp. 146-150;
64. On the mosaic and its context, see M. Guardia Pons, Los mosaicos,
cit. (n. 15), pp. 96-98; D. Fernández-Galiano, Mosaicos romanos del
Convento Caesaraugustano, Zaragoza, 1987, pp. 86-87.
65. Cf. ibid., p. 87.
66. Cf. I. Morand, Idéologie, culture et spiritualité chez les propriétaires
ruraux de l’Hispanie romaine, Paris, 1994.
CHRISTIAN IMAGES IN THE HOME 185
Somewhat more complex is the much-debated villa of
Centcelles, outside of Tarragona67. The main entrance to
the L-shaped villa is a domed hall, which in turn provided
access to an impressive tetraconch room next door
(¿J ). The dome was covered with mosaics in three
tiers separated by a thick band of decorative scales: the
lowest depicted hunting scenes, the second Old and New
Testament scenes, and the third heavily damaged images
of seated ¿gures Àanked by attendants, all backed in gold
tesserae and separated by personi¿cations of the seasons.
The apex of the dome is almost entirely gone, but was also
laid with gold tesserae and traces of two heads are still
visible (¿J). The domed room and indeed, the entire
later phase of the villa were never completed, and thus the
room lacks paving. Centered beneath the dome is a small
subterranean room, whose chronological relationship
with the building above is hotly debated.
The ¿rst comprehensive study of Centcelles by Helmut
Schlunk and Theodor Hauschild took the use of gold
tesserae, the enthroned ¿gures and the general splendor of
the program to indicate an imperial sponsor, identi¿ed as
the emperor Constans, said to have been murdered in Spain
in 350 A.D. The subterranean chamber was thus identi¿ed
as a tomb and whole thus an imperial mausoleum68. A more
recent conference organized by Javier Arce revisited these
conclusions; a re-survey of the ceramic evidence pointed to
a date in the later 4th or even early 5th century for the ¿nal
67. The excavation monograph is H. Schlunk, Die Mosaikkuppel von
Centcelles, Berlin, 1988.
68. Ibid.; T. Hauschild and A. Arbeiter, La villa romana de Centcelles,
Madrid, 1994.
186 KIM BOWES
Fig. 18 – Centcelles villa, late 4th c. a/ Plan; b/ Section (A. Chavarría, Villas in Hispania during the Fourth and Fifth Centuries, in K. Bowes
and M. Kulikowski (dir.), Hispania in Late Antiquity: Current Perspectives, Leiden, 2005, fig. 4C).
phase of the villa; the function of the subterranean room as
crypt was contested; and the identi¿cation of an imperial
patron subject to harsh criticism on both iconographic and
chronological grounds69. The overall conclusions of the
conference have placed the villa once again in a possible, if
not probable, domestic, elite ambit, and it is in this context
it is considered here. The location of the domed room at
the apex of the villa’s courtyard, and its large doorway
both suggest it was intended as a grand vestibule to the
tetraconch reception room beyond.
The three bands of imagery would seem to function
independently of one another, but there are subtle clues that
connect all three. The hunting scene is the most dynamic
and closest to the viewer and thus sets the tone: it is actually
69. J.-A. Remolà, Centcelles y las villae de Tarraco durante la antiguedad
tardía, pp. 97-109; B. Brenk, Zum problem der Krypta under
spätantiken Rundbauten, pp. 59-81; R. Warland, Die Kuppelmosaiken
von Centcelles als Bildprogramm spätantiker Privatrepräsentation,
pp. 21-35, respectively, among other articles in J. Arce (dir.), Centcelles.
El monumento tardorromano, iconografía y arquitectura, Rome, 2002.
AnTard, 1 9, 2011
two hunts, one for a stag and another for a boar that depart
laterally from a cluster of men who mark the beginning of
the hunt on the north wall and thus face the viewer as s/he
entered. The group includes what appears to be a portrait
of the dominus himself, underlining the personal stake in
what transpires. The hunts run out to the east and west,
each culminating half way around the dome in the death of
the animal before returning in triumph to the villa over the
south window. Thus, the cardinal points of the dome are
each given dramatic emphasis.
Unlike the unitary hunt scenes, the Christian scenes
above are separated by frames and include the “greatest
hits’ common in 4th century funerary art – the Three
Hebrews in the Fiery Furnace, the Good Shepherd, the
Raising of Lazarus, Jonah and the Whale, as well as some
whose identity is unclear. Their internal arrangement
seems to lack any programmatic or narrative coherence70;
70. H. Schlunk, Die Mosaikkuppel, cit. (n. 67), p. 120.
AnTard, 1 9 , 2 0 11
rather, their placement may be partially dictated by an
effort to make connections with the hunt register below
and the enthroned scenes above. The Good Shepherd is
placed directly over the portrait of the dominus, and the
excavators noted a deliberate similarity in their portraits.
To the east, over the death of the stag, is placed the last of
the two Jonah scenes showing Jonah at Rest; at the west
over the death of the boar a two-part narrative of the Three
Hebrews is interrupted by the Raising of Lazarus. The
Christian scene over the southern villa scenes is unfor-
tunately destroyed. The attempt to identify the dominus
with the Good Shepherd is plain, thus subtly associating
him with Christ in his role as caretaker. The images of
Jonah at Rest and the Raising of Lazarus were two of the
most-oft used, and thus seemingly powerful statements on
the Christian promise of resurrection. Tying them compo-
sitionally to the hunting scenes below, the dominus’ virtus,
as displayed in the hunt, is associated with the guarantee
of bodily resurrection, as described in the Old and New
Testament images of Jonah and Lazarus.
The damaged third-tier scenes have been hardest to
identify, but almost certainly represent the dominus and
CHRISTIAN IMAGES IN THE HOME 187
Fig. 19 – Centcelles villa, drawing
of mosaics in domed room (H. Schlunk,
Die Mosaikkuppel von Centcelles,
Berlin, 1988, fig. 9).
his family in poses emphasizing their of¿cial capacities71:
convincing interpretations include the dominus as philo-
sopher, the domina surrounded by servants bearing jewelry
and mirrors, and perhaps, for the northern scene over
the Good Shepherd, the awarding of prizes at the circus
games72. These scenes are also oriented at the four cardinal
71. R. Warland, Die Kuppelmosaiken, cit. (n. 69); cf. Id., Status und
Formular in der Repäsentation der spätantiken Führungsschicht, in Röm.
Mitt., 101, 1994, pp. 175-202. I ¿nd the suggestion these represent an
episcopal audience unconvincing, not least because there is no evidence
the bishopric of Tarragona in this period by anyone wealthy or powerful
enough to commission these mosaics: cf. A. Isla Frez, La epifanía
episcopal en los mosaicos de la villa de Centcelles, in J. Arce (dir.),
Centcelles. El monumento tardorromano, cit. (n. 69), pp. 37-50.
72. The image contains a seated frontally facing ¿gure holding a green
object, perhaps a crown, Àanked by attendants, one of whom holds a
fringed blanket. In the background is what appears to be a herm. Previous
suggestions included a funerary portrait of the deceased with ancestral
imagines (Schlunk) or an image of the domina holding jewelry and gazing
into a mirror (Warland). None have, to my knowledge, noted the parallels
between the fringed blanket and imagery on some Terra Sigillata C plates
in which the blanket seems to be offered as a prize in the hippodrome
(J. W. Solomonson, Spätrömische rote Tonware mit Reliefverzierung auf
Nordafrikan Werkstätten, in Bervordering der Kennis van de Antieke
188 KIM BOWES AnTard, 1 9, 2011

points, thus tying the public and private status of the family
to their Christian faith and rural virtus below.
If the room in question were, in fact, the entrance hall
to the villa, this interweaving of Christian promise and
seigniorial power would have summed up the identity
of the dominus to the visitor upon his/her entrance to the
villa. The elaboration of entrance complexes in late antique
houses has been long marked upon73, as rituals of entrance
and arrival were marked with architectural and visual
Àourishes intended to announce the identity and character
of the inhabitants. At Centcelles, that identity includes
Christian af¿liation as part of the accouterment of virtus
and potestas.
Yet another mode of interweaving status and Christian
identity is found in the so-called Palace of Eustolios at
Kourion on Cyprus74. The large complex was built behind
the theater on the acropolis and its ¿nal phases, including
the Christian inscriptions, are dated to the late 4th or early
5th century Centered on a large peristyle were a suite of
baths on the north, a set of large, possible reception rooms
to the east, unexcavated rooms to the south, and a suite of
interconnected rooms to the west (¿J). Mosaics covered Fig. 20 – Kourion, “Palace of Eustolios” plan,
the entrance vestibule, peristyle and baths and contained late 4th-5th c, with location of mosaic
inscriptions/images (M) noted (D. Christou,
a variety of inscriptions and images. In the entrance Kourion. Its Monuments and Local
vestibule, an inscription (M1) Àanked by Solomon’s Knots Museum, Nicosia, fig. 22).
read “Enter for the good fortune of the house.” Passing into
the peristyle, one confronted a sequence of inscriptions.
The ¿rst, at the entrance to a large reception room centered
on the peristyle (M2), is fragmentary but reads something of Christ”. (¿J ) The latter inscription is surrounded
like, “Eustolios, having seen that the Kourians, though by a meander and four-petaled rosettes, but no clear cross
previously very wealthy, were in abject misery, did not symbol. Finally, a personi¿cation of Ktesis or Foundation
forget the city of his ancestors, but ¿rst having presented holding a measuring rod, appears in the frigidarium of the
the baths to our city, he was then taking care of Kourion baths complex (M5), just before its main apsed pool.
as once did Phoebus [Apollo] and built this cool refuge On the basis of its large baths and accompanying
sheltered from the wind.” Two inscriptions marked the inscription seeming to imply that Eustolios gave baths
corner of the southeastern portico, near the large reception to the city, the complex has been identi¿ed as a public
rooms. At the end of the east portico (M3) one read: “The bath with attached “annex” cum social club75. While this
sisters Reverence, Temperance and Obedience to the laws is certainly possible, there is no reason to suppose that the
tend this exedra and the fragrant hall,” while at the start “public” quality of the baths disquali¿es the complex as
of the south (M4) a second claimed, “This house, in place a house – and indeed the complex is termed a domos and
of its ancient armament of walls and iron and bronze and oikos in the inscriptions. Domestic baths were regularly
adamant, has now girt itself in the much-venerated signs hired out for public use and after the mid-4th century earth-
quakes that damaged the city, it may be that Eustolios’s
complex was one of the only bathing facilities in town.
It is also possible that the inscription does not refer to
Beschaving te’s-Gravenhage, 44, 1969. pp. 4-109). The “herm” image these baths, but another facility built elsewhere. The
in the background would thus be the imperial images displayed in the “public” nature of the inscription is also explainable by
kathisma, or main viewing box in the circus or amphitheater. its location: placed at the entrance to the house’s grandest
73. Cf. I. Baldini Lippolis, La domus tardoantica: forme e rappresentazioni
dello spazio domestico nelle città del Mediterraneo, Bologna, 2002, room it continues a trend in late antique houses of setting
pp. 53ff. up inscriptions in the houses of benefactors to display in
74. Excavation report: J. De Courtney Fales, Kourion - The amusement private their public muni¿cence.
area, in University Museum Bulletin, 14, 1950, pp. 27-37; Inscriptions:
T. B. Mitford, The Inscriptions of Kourion, Philadelphia, 1971;
interpretations: D. W. Rupp, Eustolios complex, in H. Wylde (dir.), An
Archaeological Guide to Ancient Kourion, Nicosia, 1982, pp. 132-139;
D. Soren and J. James, Kourion: The Search for a Lost Roman City, 75. D. Soren (in D. Sorfen and J. James, Kourion, cit. [n. 74], p. 20)
New York, 1988, pp. 20-22; 200-202. thought the “annex” was a private house, while the baths were public.
AnTard, 1 9 , 2 0 11
Fig. 21 – Kourion, “Palace of Eustolios”, mosaic inscription from south peristyle
(D. Christou, Kourion. Its Monuments and Local Museum, Nicosia, fig. 24).
The Christian inscription in the southeast corner has
typically been read as a public statement of faith on the
part of the Kourions76. If the complex is Eustolios’ private
residence, as seems most likely, the inscription has more
individual functions. It is composed in Homeric dactylic
hexameters and claims that Christ, who has taken the place
of regular building materials, supports the very fabric of
the house. Those regular building materials are seemingly
associated with a thrown-off pagan past, but also reference
Eustolios’ role as builder after the earthquake. That is, the
house’s Christianity is subtly linked to Eustolios’ building
projects, proclaimed nearby and in the baths through the
Ktesis image. As Eustolios has taken on the role of Apollo
as city patron, his projects, i.e. his house with its baths,
are built with the help of Christ. Finally, if Reverence,
Temperance and Obedience to the Law are to be read as
speci¿cally Christian personi¿cations (one scholar has
read “Obedience to the Law of God” here), these virtues
are explicitly tied to house’s ¿ne aspect and other charms,
again echoing sentiments expressed in the Eustolios
inscription. Thus, Eustolios locates Christian protection
and virtue not in himself, but in the very fabric of the house,
while he describes himself as the successor to Apollo and
benefactor of the city. Christian af¿liation is thus attributed
76. Ibid., p. 22.
CHRISTIAN IMAGES IN THE HOME 189
to Eustolios’ projects rather than himself, and thus plays a
secondary, more subtle role in the articulation of his status
as homeowner than at Centcelles. Unlike Centcelles, too, is
the deliberate, albeit subtle rejection of a pagan past in the
articulation of that Christian sentiment: Eustolios replaces
Apollo and his house rejects the “ancient” material world
in favor of the signs of Christ.
V. CONCLUSIONS
It is hoped that the above discussion reveals the range
of iconographies and functions of Christian images in late
antique houses. Christian images might be used in ritual
contexts, as objects of veneration or as signposts to guide
ritual activity. They might serve as protective agents to keep
evil forces from penetrating the house and its occupants.
And they might simultaneously be part of the elite house’
status apparatus, deployed in the complex interactions
between aristocratic peers, or woven into statements about
individual identity.
Many of these images and their functions seem puzzling
to us as modern scholars. The majority of extant Christian
images, and thus the basis for the study of early Christian
art, are products of the public church. Christian images in
homes, on the other hand, were produced by individuals and
thus reÀect their personal beliefs and needs as householders.
190 KIM BOWES
Those impulses, and thus the full meaning of these images,
are often dif¿cult to recover with any precision77. The insti-
tutional hand that regularized iconography and meaning in
the public sphere need not apply so insistently in the home,
and thus our own interpretative tools, so heavily shaped by
that public church, often fail us. It is a salutatory failing, for
it recalls to us a more complex late antique reality, in which
so much Christian life took place outside the bounds of the
basilica and the regularizing impulses of its bishops.
Finally, it is worth asking whether there is anything
particularly Christian about the use of Christian images
in the home. Not only are many of the Christian examples
discussed here embedded in a program of pagan imagery,
but they also seemingly ful¿lled many of the same basic
functional roles: the veneration of images, the reproduction
of symbols of power for protection, the use of religious
af¿liation to produce social status – none of these seems
a speci¿cally Christian use of images. I would argue that
if there were something particularly “Christian” about
Christian domestic images, it was not in the functions they
served, but rather in the functions they didn’t.
As Susannah Muth and others have shown, generations
of elite “pagan” householders used mythological scenes to
think through the most basic problems of domestic life78.
By shaping their speci¿c composition and iconography
and carefully placing them in their respective rooms,
Greco-Roman householders used these images to think
about erotic love in marriage, faithfulness, domestic roles
and other issues. Mythological narrative, in short, was
tailored to think with. There is no Christian parallel for this
kind of thinking that we can see either in the permanent
77. Although the sense of hermeneutic vertigo is perhaps particularly
pronounced with Christian images, domestic images of all kinds resist
easy decoding exercises: cf. on myth, S. Muth, Eine Kultur zwischen
Veränderung und Stagnation. Zum Umgang mit den Mythenbildern
im spätantiken Haus, in F. A. Bauer and N. Zimmerman (dir.),
Epochenwandel? Kunst und Kultur zwischen Antike und Mittelalter,
Mainz am Rhein, 2001, p. 100.
78. S. Muth, Erleben von Raum, Leben im Raum. Zur Funktion mythologischer
Mosaikbilder in der römisch-kaiserzeitlichen Wohnarchitektur,
Heidelberg, 1998; Id., Eine Kultur zwischen Veränderung, cit. (n. 77).
AnTard, 1 9, 2011
domestic images or the numerous candidates among
portable objects. Christian narrative scenes are largely
absent from the home, and with the notable exception of
Centcelles, particularly from the admittedly thin corpus
of permanent domestic images. While in the mosaic arts,
mythological scenes continued to be produced through
the 4th and 5th centuries in many regions and continued to
be used to make sense of gender relationships and other
issues79, there was seemingly no impulse to do the same
with Christian Old and New Testament images. That is,
in the domestic sphere, Christian narrative images never
assumed the same role as discursive agents on domestic
life. If Christian images entered into discursive contexts, as
has been suggested here for Hinton St. Mary and Frampton,
they did so using symbols or single images, not stories,
and the discourse engendered seemingly looked past the
household to the issues of salvation and the soul.
This is doubly strange if we consider that churchmen
were, in fact, using biblical narratives precisely for the
purpose myth had served – for ordering, parsing and
making sense of domestic relationships80. That house-
holders did not use Christian images to do the same thing
may reÀect the particular way they viewed such images – as
emblems not musings, as symbols not stories, and above
all, as pointers that gestured beyond domestic concerns to
concepts of power and authority. If there is one commo-
nality that unites permanent Christian domestic images, it
is their strange reluctance to comment upon their human
domestic communities.
Department of Classical Studies,
University of Pennsylvania
79. Ibid.
80. See among a large bibliography, P. Brown, The Body and Society,
cit. (n. 14); A. Jacobs and R. Kraweic, Fathers know best? Christian
Families in the Age of Asceticism, in Journal of Early Christian
Studies, 11, 2003, pp. 257-263; K. Cooper, The Fall of the Roman
Household, Cambridge, 2007.