Scholia Reviews ns 13 (2004) 8.


Simon P. Ellis, Roman Housing. London: Duckworth, 2002. Pp. viii + 224, including 30 figs., 22 black-and-white plates, one map, glossary, endnotes and two indices. ISBN 0-7156-3196-9. UK£16.99.

Alison B. Griffith
Department of Classics, University of Canterbury, Christchurch, NEW ZEALAND

In 1975 A.G. McKay's Houses, Villas, and Palaces in the Roman World was the most comprehensive introductory work on Roman housing.[[1]] This geographically organized survey focused on housing in Italy and treated provincial evidence separately in two final chapters. McKay was concerned with structure and its origins, and his handbook was typological in nature.[[2]] The study of Roman housing has now been so utterly transformed that a new general treatment is long overdue. In the interim, the houses of elite Romans have been rigorously scrutinized and are now regarded as methodically and deliberately created environments in which a range of activities took place, most important among them the display of their owners' status.[[3]] Whether architectural and decorative innovation should be ascribed to individuals or to societal changes is, Ellis suggests, a fundamental question that has prevented anyone from writing a new general book on the subject of Roman housing until now (p. 4).

The aims of the book astonish: Ellis intends not just to provide 'the first empire-wide, overall introduction to Roman housing, covering all provinces and all social classes, from the origins of Rome to the sixth century AD' to the student and general reader but also 'a wealth of comparative evidence' to specialists (p. l). Adding to these Herculean tasks, he further promises 'The Roman house will be taken apart and pieced back together in a way never attempted before' (p. 4). The seven chapter headings are focused topically: Introduction (pp. 1-21), Houses of Pretension (pp. 22-72), Town and Country (pp. 73-113), Decoration (pp. 114-44), Furniture (pp. 145-65), The House and Family (pp. 167-87, and Conclusions (pp. 188-91). Numerous subdivisions in each chapter guide the reader and assist revisitation of particular topics. Ellis's personal interest in 'reception rooms' and other aspects of elite housing provides thematic coherence and is also used to signal, evaluate and demonstrate the implications of recent research concerning Roman houses.[[4]] What the chapter headings conceal is a much-welcome integration of detailed but succinctly related comparative material -- from Britain, the European and North African provinces and the East -- to the main thread of the discussion, whatever the topic might be. In short, Ellis acquaints his readers with the state of the field as articulated recently by work in Pompeii and Italy, but applies its conclusions more broadly in geographic terms than any other standard works currently do. To his survey he brings expertise in town planning and his wide knowledge of Roman Carthage, North Africa and late- antique housing.[[5]]

Ellis's 'holistic approach' to the Roman house, which assumes the inter-relationship between structure, decor, furnishings, and allocation and use of space, is the culmination of the work of many scholars over several decades. Of central concern to work on Roman housing is the way in which Romans used the space they inhabited and the extent to which archaeological and textual evidence increases understanding of the Roman domestic environment.[[6]] Current discussion also assumes that the home is a deliberate construct, at least for wealthy or aristocratic Romans.[[7]] The roots of this conclusion, arguably, began with a desire to move beyond August Mau's 'four Pompeian styles' and the subsequent search for meaning in Roman wall paintings, which quickly yielded attempts to contextualize them culturally, and to view them with Roman eyes.[[8]] We now talk about 'decorative ensembles' within a room, or 'sight-lines' between rooms and throughout the house. We regard Roman houses as venues for personal expression representing, in part, their owners' perception of status and social standing. We accept that a well- appointed Roman house must be 'read,' its decor 'decoded' and the 'underlying grammar and vocabulary' of its 'language' understood if we are to learn anything about its owner. The linguistic metaphor has been applied to Trimalchio who 'parodies the language of Roman luxury rather than communicating in it.'[[9]] Similarly, says Ellis of both Trimalchio and the Vettii, 'Any shortcoming in their [cultural] knowledge was not through want of reading but more from their inability to restructure what they had learnt' (p. 11). In point of fact, trends of cultural expression were so pervasive that it can be nearly impossible to distinguish the social status of an owner on the strength of archaeological evidence alone. That is partly because we can never know for certain which interested parties -- owner, architect, painter, mosaicist -- exerted the most influence in the overall design and presentation of a Roman house, as Ellis admits (pp. 6-9). Moreover, Ellis rightly points out, Roman houses were 'heavily constrained by conventions of the local community and society at large,' (p. 9) so that it is also difficult to discern whether social behaviors preceded and therefore influenced the design of elite houses, or whether adoption of form inspired a consequential behavioral change.

Most important in the last decade is the complete turnabout in our thinking about how Roman houses were used. Recent research has emphasized repeatedly and variously that Roman homes, elite and otherwise, were a venue for 'business' in the broadest sense. Shops, the smallest formal residential unit, had a bed or attached living quarters (pp. 78-80). Further, Ellis is keen to dispel the notion that a higher degree of artisanal or commercial activity at a habitation site necessarily indicates a poorer dwelling (pp. 88f., 107f.). The contents of houses must also be considered, and studies that assess the actual provenience of artifacts are a potent reminder that Roman usage of space does not always conform to modern expectations.[[10]] Investigation into the location of artifacts in the eleven-room house of M. Epidius Primus at Pompeii (I.8.17), for example, indicated in the atrium a horse's harness, the remains of an iron-wheeled cart, and tools there and throughout the house.[[11]] Further up the social scale elite Romans of the Republic used their homes for the morning salutatio and invited special guests and clients to dine in the triclinium at the close of the day. Augustus' choice to reside in his own Palatine domus effectively redefined the role of the house in an imperial context, such that the business of ruling took place in the Palatine residence of subsequent emperors and influenced the incorporation of domestic elements into palatial architecture (pp. 54, 72).

Interest in the origins of the two major urban types of 'houses of pretension' -- the atrium house during the Republic and the peristyle house during the Empire -- continues, but the focus now includes the houses at such sites as Olynthus, Delos, Megara, Hyblaea, Pergamon and Priene as well as Italian antecedents. The possible influence of Etruscan chamber tombs on the Roman atrium house is now discounted in light of our rudimentary understanding of Etruscan housing and settlement sites generally. By contrast the role of the peristyle in palatial residences of Hellenistic kings and in other structures such as the gymnasium, and its influence on the Roman peristyle house, have been the subject of considerable discussion. Even the use of columns has been re-evaluated. These are not just structural elements; rather, their use or reproduction in stucco and paint represents intentional association with public architecture.[[12]] Ellis rightly cautions that apparent structural similarities in plan tell us nothing about patterns of usage, which for him remain paramount (pp. 24, 35). A loss of function in the atrium / tablinum must necessarily precede its absence from Roman houses (p. 36), although little evidence sustains the notion that changes in the early Principate led to a decline in the importance of the atrium and tablinum in houses of the aristocracy in favor of more general-purpose reception rooms such as the oecus and triclinium (p. 69). So also, Ellis argues, the appearance of specialized audience halls in late antiquity certainly suggests that these filled a new spatial need for a reception area.[[13]]

Ellis traces the diffusion of the atrium house, or local translations of the type, but because its heyday paralleled the growth of the Empire, provincial examples are restricted to Spain and southern France. In his survey of these, Ellis treats the evidence conservatively in light of the absence of a tablinum in most cases. By contrast, clear examples of provincial aristocratic housing including a peristyle abound, particularly in the second to fourth centuries. Ellis defines the 'ideal' (not a 'norm') as a central peristyle completely surrounded by 'ranges of rooms', with the main reception room directly opposite the entrance from the street (p. 41). Local and individual expression of the type varies considerably, as Ellis shows, such that reception rooms can sometimes only be identified after careful consideration. The preference for a peristyle can be seen in villas both in Italy and the provinces where there was 'one major reception room located on the central axis of the house off the inner peristyle or yard' (pp. 68f.). This room, Ellis further suggests, fulfilled all the functions of a triclinium, tablinum and oecus. Ellis concludes that the omnipresence of the peristyle house type, or at least of local interpretations of the form, indicates 'a single, empire-wide aristocratic culture' (pp. 108, and 69 and 97).

Recent excavation and re-investigation in Pompeii indicates that the Roman house did not so much consist of spaces dedicated to particular activities, although this was certainly true to some extent, but to different sorts of spaces devoted primarily to display in a social context. The house 'did not merely reflect but generated status' of the dominus; thus Ellis's houses of pretension are defined by the very rooms whose function, in part, related to ostentatious display by the owner: the tablinum / atrium / alae trio and later the oecus / peristyle.[[14]] This much we conclude from the physical remains, but beyond that the archaeological record compels us to abandon modern notions of space compartmentalized according to sex, status, age or activity. We know of no nurseries or gynaecea. It is difficult to identify slave-quarters securely from the archaeological record, although the array of functions that household slaves performed indicates their presence throughout wealthy Roman houses.[[15]] We are now encouraged to recognize how fluid were the lines that might otherwise seem to divide the house. As Ellis and others caution, houses were for business and pleasure, with a time but not necessarily an independent space for each.[[16]]

The now much-studied use of Roman domestic space and the concomitant notion that it was constructed have received increasingly theoretical treatment. Wallace- Hadrill proposed two intersecting axes of space in a Roman house defined according to access to the male head of the household through a main entrance: the private-public axis and the grand-humble axis. In this model architecture and decor regulate circulation within the house.[[17]] With the application of Hillier and Hanson's[[18]] access analysis theory Grahame introduced an extra dimension of privacy -- that separating members of the same household -- to the basic public-private distinction.[[19]] For Ellis, these issues boil down to 'circulation patterns' (pp. 166-70). He faults Wallace-Hadrill's theory as too reliant on the old model of assigning functions to particular rooms and too little cognizant that the duty of a servant might require sleeping in the bedroom of the master or mistress. Hillier and Hanson's theory is praised for its elucidation of spatial organization but ultimately dismissed as 'mechanistic' and unable to take into account that which does not appear in plan (e.g. decor and furnishings). The Roman house, Ellis suggests, had relatively 'open' circulation in that rooms were often grouped around a central court, so that passage from one room to another required crossing this court, which maximized the potential of encountering another member of the household. Still, he notes, there exist multiple circulation routes to particular parts of Roman houses, for example dining rooms, such that the proper route for a visitor necessarily differed from that of a servant.

Despite his 'holistic' approach to the house Ellis devotes Chapter 4 to decoration (pp. 114-44) and here treats wall painting and mosaic in relative isolation. The discussion on painting provides a useful synopsis of recent work,[[20]] and moves outside Italy and beyond A.D. 79 to the late-antique period. The debt to other scholars is also acknowledged in the sections devoted to mosaics. Ellis examines them contextually as floor coverings, noting carefully their placement within a room, their intended viewpoint(s) and interaction with furniture. Also included is a brief nod to sculptural decoration, water features, and the relative improbability that we can understand pervasive mythological allusion in decor as evidence for domestic cults. The final sections of the chapter reunite various forms of decor in order to evaluate scholarship on its impact, relationship to room function and contribution to ambience. Here Ellis argues that symbolic interpretations should not be over-emphasized, but that meaningful associations can reveal 'Romans' conception of their domestic life' (p. 141). Furniture forms the subject of Chapter 5 (pp. 145-65), and is regarded as integral to the decor of the room in which it was located, despite its mobility, in the sense that a range of furnishings indicates flexibility of room function. A clear distinction is made between items for storage and those for other use, and the term is used broadly and includes partitions, wall coverings and lighting.

Two main premises link diverse material in Chapter 3, 'Town and Country' (pp. 73-113): that Romans of lesser means could imitate Roman (aristocratic) behavior through conscious selection of specific elements pertaining to Roman culture, and that the rural/urban dichotomy in terms of the function and activities in Roman houses is false. Ellis also makes a good case for a vernacular tradition of 'housing [that is] Roman by date but not apparently Roman by culture' (p. 87), arguing that adoption of Roman forms only in part, or not at all, does not necessarily indicate resistance to Roman culture (p. 112). In terms of the depth of research and sheer range of topics this chapter might be the richest, although for reasons of space it serves more as an introduction to the considerable body of evidence now available. The sixteen sub-sections include urban housing, shops and taverns, towns, villages, trade and industry, factories, farms, homesteads and fortified farms in Italy, Britain and in the European, African and eastern provinces where excavation has yielded sufficient data. Using apparently anomalous cases such as the medianum 'houses' of Ostia, the House of the Prince of Naples in Pompeii (VI.15.7-8), Lot 11 at Utica and the House of the Brick Walls at Djemila, Ellis examines urban houses of modest but respectable means not of the 'ideal' peristyle type but containing (arguably) identifiable reception rooms. The central sections of the chapter address the range of provincial responses to habitation in villages, towns and small cities where lesser habitation density sometimes allowed a considerable degree of agricultural, artisanal or industrial activity in a residential context. Included in the discussion are so-called strip housing of the northwest provinces, and variations on apartments and peristyle houses evident in Syrian villages and cities and at Karanis and Alexandria in Egypt. The final sections on individual farms and homesteads of Britain and Gaul and factories and fortified farms of North Africa provide a concise but useful introduction to the excavations of the last twenty years. Here, as in previous sections, Ellis questions the extent to which a clear distinction between residential and agricultural activity in particular buildings can be made, while at the same time identifying the sometimes surprising mixture of non-Roman and Roman traits.

Roman Housing lives up to its promises in almost every respect, but certain aspects of this introductory text may be seen by some as limitations. For example, Ellis concentrates on elite housing (p. viii), and although textual evidence is frequently and appropriately taken into consideration, this survey is largely archaeological. Further, the book is aimed at a wide audience, but those with background in Roman culture and history will undoubtedly benefit more than the general reader. The book is technically well produced and has few errors, but illustration is sometimes sparse for a book purporting to be an introduction to the subject, and the lack of plans is sometimes keenly felt where such would be most welcome. The omission of a bibliography is regrettable. These are minor oversights. Roman Housing, a book twenty years in the making according to its author (p. vii), is an extremely well-researched, readable and easily comprehensible survey that will stand as a point of departure for the next generation of students and specialists.


[[1]] London 1975, repr. Baltimore 1998. See also J. Percival, The Roman Villa (London 1975).

[[2]] In addition to the chapter McKay devotes to origins, an iteration of the argument can be found in A. Boethius and J. B. Ward-Perkins, Etruscan and Roman Architecture (Harmondsworth 1970) and in subsequent editions: A. Boethius, Etruscan and Early Roman Architecture (Harmondsworth 1978 and repr. New Haven 1994) and F. Sear, Roman Architecture (Ithaca 1982).

[[3]] S. Hales, The Roman House and Social Identity (Cambridge 2003) was released just as this manuscript was submitted and has not been read in full by the author. Hales accepts this view of the relationship of house and status, but argues further that decor and architecture, or the 'art of impression', in provincial urban housing also convey the Roman identity (Romanitas) of the owner.

[[4]] The bibliographies of the following significant books in this field refer to numerous articles also worthy of consultation. J. C. Anderson, Roman Architecture and Society (Baltimore 1997); E. K. Gazda, Roman Art in the Private Sphere (Ann Arbor 1991); R. Laurence and A. Wallace-Hadrill (edd.), Domestic Space in the Roman World: Pompeii and Beyond (JRA suppl. v. 22: 1997); R. Ling, The Insula of Menander at Pompeii (Oxford 1997); R. Laurence, Roman Pompeii: Space and Society (London 1994); L. Richardson, Pompeii: An Architectural History (Baltimore 1988).

[[5]] S. Ellis, 'Classical Reception Rooms in Romano- British Houses,' Britannia 26 (1995) 163-78; idem, 'Late-antique Dining: Architecture, Furnishings and Behaviour,' in Laurence and Wallace-Hadrill [4] 41-51; idem, 'Power, Architecture and Decor: How the Late Roman Aristocrat Appeared to his Guests,' in Gazda [4] 117-34; idem, 'The End of the Roman house,' AJA 92 (1988) 565-76; idem, 'The 'Palace of the Dux' at Apollonia and Related Houses,' in G. Barker, J. Lloyd, and J. Reynolds (eds.), Cyrenaica in Antiquity (BAR S236, Oxford 1985) 15-25.

[[6]] See D. Perring, The Roman House in Britain (London and New York 2002), in which use and function are regarded as intrinsic to understanding Romano-British housing. See also J. T. Smith, Roman Villas: A Study in Social Structure (London 1997), which has been criticized as unnecessarily typological and 'one- dimensional' in its failure to assess literature and artifacts (J. Rossiter, 'Interpreting Roman Villas,' JRA 13 [2000] 572-77).

[[7]] Prominent among these are J. R. Clarke, The Houses of Roman Italy, 100 B.C. - A.D. 250: Ritual, Space and Decoration (Berkeley 1991); E. Dwyer, 'The Pompeian Atrium House in Theory and Practice,' in Gazda [4] 25-48; A. Wallace-Hadrill, 'The Social Structure of the Roman House,' PBSR 56 (1988) 43-97; and P. Allison, 'The Relationship between Wall-decoration and Room-type in Pompeian Houses: a Case Study of the Casa della Caccia Antica,' JRA 5 (1992) 235-49.

[[8]] See in particular Clarke [7] and E.W. Leach, The Rhetoric of Space: Literary and Artistic Patronage in Ancient Rome (Princeton 1988).

[[9]] A. Wallace-Hadrill, Houses and Society in Pompeii and Herculaneum (Princeton 1994) 6, 14f.

[[10]] P. Allison, 'Artefact Distribution and Spatial Function in Pompeian Houses,' B. Rawson and P. Weaver (edd.), The Roman Family in Italy: Status, Sentiment, Space (Oxford 1997) 321-54; eadem, 'How Do We Identify the Use of Space in Roman Housing?,' in E. M. Moormann (ed.), Functional and Spatial Analysis of Wall Painting: Proceedings of the Fifth International Congress on Ancient Wall Painting (Leiden 1993); eadem, 'Artefact Assemblages: not "the Pompeii Premise"', in E. Herring, R. Whitehouse, and J. Wilkins (edd.), Papers of the Fourth Conference of Italian Archaeology: New Developments in Italian Archaeology, Part I (London 1992) 49-56.

[[11]] J. Berry, 'Household Artefacts: Towards a Re- interpretation of Roman Domestic Space,' in Laurence and Wallace-Hadrill [4] 183-95.

[[12]] J-A Dickmann, 'The Peristyle and the Transformation of Domestic Space,' in Laurence and Wallace-Hadrill [4] 12-36, summarizes this discussion well. See also Wallace-Hadrill [9] 17-37; 20-23.

[[13]] Ellis in Gazda [4] 117-34.

[[14]] Wallace-Hadrill [9] 59. On pp. 59f. Wallace- Hadrill talks about the 'hierarchy of intimacy' in the layout of rooms: waiting room, reception room, inner sanctum.

[[15]] M. George, 'Servus and domus: The Slave in the Roman House,' in Laurence and Wallace-Hadrill [4] 15-24 for a discussion of previous methodology and its problems. Ellis takes exception to Carrandini's assertion that status demarcations were clear (p. 166f. and n.1).

[[16]] In addition to works already cited, see the now considerable literature on women and the family: S. Dixon, The Roman Mother (London 1988) and The Roman Family (Baltimore 1992); S. Treggiari Roman Marriage: iusti coniuges from the Time of Cicero to the Time of Ulpian (Oxford 1991) and B. Rawson (ed.), The Family in Ancient Rome: New Perspectives (Oxford 1986) amongst others.

[[17]] Wallace-Hadrill [9] 11 and 38-61.

[[18]] B. Hillier and J. Hanson, The Social Logic of Space (Cambridge 1984).

[[19]] M. Grahame, 'Public and Private in the Roman house: the Spatial Order of the Casa del Fauno,' in Laurence and Wallace-Hadrill [4] 137-164.

[[20]] R. Ling, Roman Painting (Cambridge 1991) and Clarke [7].