This website contains plans, data, analyses and photographs of a sample of thirty houses in Pompeii, as interactive and searchable databases.
These thirty Pompeian houses were excavated variously between 1826 and 1978. They are all so-called atrium houses, and all except one have at least one garden as well as the defining front hall, or atrium. They are located throughout the town of Pompeii. Reasons for the choice of houses are provided in the monograph.
The data comprise information on the contents of these houses and was compiled mainly from the excavation records. These records include published reports, predominantly in the Notizie degli Scavi di Antichità, as well as the reports in original daybooks, the Giornali degli Scavi, and the artifact inventories held in the Pompeii archives. Further data were also collected through physical examination of the extant remains and from more recent publications (for example, studies of wall decoration). These data were compiled, principally, so that the spatial distribution of the contents of Pompeian houses could be investigated for a better understanding of: 1) the functions of the rooms in these houses; and 2) the abandonment processes these houses underwent between AD 62 and AD 79. Specific results from these inquiries are included on the website but more general results are in the accompanying monograph (see also Allison 1992b, 1993, 1994a, 1994b, 1995a, 1995c, 1997a, 1997c, 1997d).
The website includes a summary of each of the thirty houses in the sample. These summaries provide basic information on the form of the houses, references to their excavation reports and quality of recording, and an interpretation of their state of occupancy in AD 79, based on the study of their contents.
The house summaries link to a discussion on each of the rooms in the particular house. The room discussions include, for each room in each house: a brief description of the room and its decoration; any stratigraphical information in the excavation reports; a summary of the room contents; and an assessment of past interpretations of the room's use based predominantly on textural nomenclature, and a new interpretation on its use and its occupancy in AD 79 based on the study of its contents.
The house summaries also link to the databases. These databases comprise three relational tables; a table of houses, a table of rooms and a table of artifacts. The fields in these tables as described in the reference aid: explanation of database fields. Each house summary is also linked to the relevant sections of the table of rooms and the table of artifacts. Further searches can also be carried out on these tables.
The significant part of this website is the artifact database. It includes primarily those artifacts that were decontextualized during excavation. That is, the extant architectural structure of a house and its in-situ decoration are not included. However, artifacts like roof tiles are included if they were building material ready for use and not part of the house structure. Also, fixtures that appear to have been associated with habitual room use (for example, latrines) are generally included. The volcanic deposit from the final eruption is not included. However, breaches cut through the walls, that appear to have been associated with either abandonment processes or post-excavation looting, and therefore the integrity of the room's contents, are included.
The artifact database consists of fields that provide basic identification information and provenances and fields that sort the artifacts into various categories (for example, material or function). The actual artifacts were not analysed for this study, only their description in the excavation reports and inventories. These reports were written by a number of different archivists mainly in the late nineteenth and early twentieth centuries. They often defined the artifacts in their local Italian dialect, in Latin or Greek terms, or in italianizations of the ancient terms (e.g. "urceo"). These artifact labels are sometimes ambiguous or have a number of English translations. At other times, they provide artifacts with inappropriate or very specific functions that are unsubstantiated. Because of the potentially interpretative nature of translating many of these terms into English, the original Italian, Latin, or Greek terms are often preserved in the database. Possible translations based on dictionary definitions and knowledge of the particular artifact are provided in the glossary. In addition, because the reports were handwritten it was not always possible to read each word. Therefore, not every artifact can be identified.
The identifications and descriptions of artifacts in the database are, thus, those of the compilers of the excavation reports and inventories, and are not necessarily substantiated by the compilers of this database (see Allison 1999b). A more detailed study of the actual finds from the Insula del Menandro has subsequently been carried out (Allison n.d.), and will be useful for the user to further identify many of the artifacts in this database.
Three fields in the artifact database are specifically designed to help the user sort the artifacts according to type and function. The "category" fields includes very broad artefact categories. The "artifact type" field for each artifact is important because, in the excavation reports, similar artifacts were often given different names by different reporters (e.g. a 'tegame' and a 'casseruola'). This field attempts to group similar objects or parts of similar objects together. The "artifact function" field group the artifacts according to largely functional criteria. Because each artefact was not analysed, attempts have been made to keep of these classifications fairly general.
These databases were used to present an overview of the patterns of spatial distribution of house contents. The complexity of the data collection procedures and the volume of material involved means that there are undoubtedly errors, important to those wanting a precise catalogue of finds. However, such detail is less significant for the aims of the study, Pompeian Households. Nevertheless, I have attempted to keep errors and oversights to a minimum and apologize for any that may have slipped through.
In 1991, multivariate statistical analyses using MV-Arch's Correspondence Analysis (BIGCOR, Wright 1989) were carried out on these data to assess for concordance between architectural room type and their contents. It showed that formal multivariate analyses of this database of Pompeian house contents could potentially produce informative results. However, particular characteristics of the data, such as the varying standards of excavation and the variation in types of material included in it (for example, cupboards to coins) had a significant effect on the results (Allison 1992c). Any further statistical analyses must be take account of such variations.
This study has in no way exhausted all the possible analyses to which these data can be subjected. Rather, it represents a sampling of the ways in which artifact assemblages can be analysed to produce a deeper understanding of life in Pompeii, and to compare Pompeian lifeways with those at other Roman sites and with perspectives gleaned from documentary evidence in other parts of the Roman world. There are also many other questions that may be answerable through these data. Such questions will, I hope, be taken up by other scholars who can add to this body of Pompeian data in more informed ways, less dictated to by the literary evidence.
Note: We are continuing to work on improving this site. Please advise us of any errors you see.