The Database as Material Culture

Martin Newman

'Material culture is the study through artefacts of the beliefs - values, ideas, attitudes, and assumptions - of a particular community or society at a given time.' (Prown 1982, 1)

1. Introduction

Introduction | The NMR's datasets | An analogy from left of field | The materiality of information | How does the medium influence... | Conclusion

Archaeological recording pre-dates the digital age and the first records created from investigations of our past were physical entities in themselves. These notebooks, card indexes, annotated maps and photographic prints have now been replaced by databases, Geographic Information Systems (GIS) and digital photographs. Whereas the former can be handled, conserved and displayed, the latter do not have a physical presence in the same way, although the information content is comparable. Can the study of physical records of archaeology be seen as a study of material culture? Can they shed light on the archaeological society at the time these were created, and the evolution of archaeology as a discipline similar to the study of finds (which are more readily accepted as material culture)? If so, can the digitised equivalents of these records be examined in the same way even though they have no physical presence? Is the information contained within records an example of material culture in its own right that can be divorced from the medium on which it is carried? Can something intangible be regarded as material culture as if it were a physical object?

This article had its origins in considering how information is separated from the medium on which it is carried, whether in the traditional physical form or digital. So, thinking philosophically, consideration can be given to what a record actually is physically as well as conceptually, how its physical media affects how it is perceived and the relationship to the individual data concepts it contains. This is particularly pertinent to the information about the Historic Environment, where antiquarians, archaeologists and architectural historians have been cataloguing and recording the physical remains of past human societies in inventories since the time of antiquarians such as William Stukeley, who during his Tours in the 1720s travelled the country drawing and recording the ancient sites he encountered (Piggott 1985, 161-64). A large number of records exist on the archaeology and historic built environment of Britain, including the local Historic Environment Records (HERs, formerly Sites and Monuments Records or SMRs). One of the best case studies to illustrate this, because of its longevity and the number of changes it has overseen, is the English National Monuments Record (NMR), now the public archive of English Heritage. The NMR contains over 10 million archive (mostly physical) items, as well as material such as photographs, plans and correspondence. It also curates information on a number of databases and GIS including 500,000 monument records (archaeological sites and historic buildings), 80,000 records of recording events, 35,000 records of listed buildings (the legal designation protecting the built heritage) and records on 20,000 scheduled monuments (the legal designation protecting archaeological sites). Many of the datasets originated in a pre-digital age as card indexes, annotated maps and other records, which were then digitised and have since been migrated from one system to its successor, in some cases several times, and reflect evolving data standards and terminologies as well as changing dissemination methods. It is this change in the physical medium and how this alters the record and its perception that this article will explore.

2. The NMR's datasets

Introduction | The NMR's datasets | An analogy from left of field | The materiality of information | How does the medium influence... | Conclusion

Other than data that deal directly with archive material (for example cataloguing and event data), the NMR manages a number of datasets from which the examples used here are drawn.

2.1 The Monument Record

One of the largest datasets at the NMR is the National Record of the Historic Environment (henceforth the Monument Record). This has its origins with the National Archaeological Record (NAR) and the National Buildings Record (NBR). The basis of the work carried out by the NAR was to record antiquities that were subsequently published on maps by the Ordnance Survey (OS), the national mapping agency for Great Britain and Northern Ireland. These were primarily earthworks, ruins and the location of important sites no longer visible. Archaeological recording was expanded and a system of card indexes containing details of sites was set up within the Archaeology Division of the OS by C.W. Philips (the successor to O.G.S. Crawford), who took up the post of OS Archaeological Officer in 1947 (Philips 1959, 202; Aberg and Leech 1992, 158). The archaeological responsibilities of the OS were transferred to the Royal Commission on the Historical Monuments of England (RCHME) in 1983 and became part of the National Archaeological Record. This was later merged with data from the National Buildings Record (NBR). The NBR was founded following a conference held by RIBA in September 1940 to record historic buildings at risk during World War II and commenced work the following year (Summerson 1991, 2-5). The NBR was officially wound up in 1963 and the archive transferred to the RCHME (Croad 1989, 26-28).

Combining the NAR and NBR created the first holistic record of the historic environment of England. The OS '495 Cards' (see Figure 1) contained details of archaeological sites with references, and were often accompanied by 'illustration cards' (Figure 2) with an 'Antiquity Model' showing how a site should be depicted, a sketch or photograph. The site was also marked by hand on 1:10,000 scale OS quarter sheet maps. If the site met the criteria that the OS set for inclusion on a map, then a separate '210 Card' (Figure 3) was created. Whenever a new book or journal came into the library, the site details and reference were entered onto a 'mini card' (Figure 4). These were so named because they were smaller than the '495', 'Illustrative' or '210' cards and were filed by parish. The 'mini cards' would then be checked and the appropriate '495 Card' updated or a new one created, if applicable. This workflow has now been replaced by a streamlined, paperless process involving editing and creating records on a database and the depiction on a GIS layer, and it is the effect of this change on the information and how it is perceived that is crucial to this case study.

Figure 1 Figure 2
Figure 3 Figure 4

Figure 1: An example of an OS 495 Card showing corrections by hand (© Crown copyright. NMR)
Figure 2: An example of an OS Illustrative Card showing an Antiquity Model (© Crown copyright. NMR, map ©Crown copyright. Ordnance Survey)
Figure 3: An example of an OS '210 Card', showing archaeological features as they would appear on a portion of a map (© Crown copyright. NMR)
Figure 4: An example of an OS 'Mini Card' with the corner clipped to show that the bibliographic record had been looked at (© Crown copyright. NMR)

As can be seen from the examples, these record cards contained a combination of text and depictions of sites with varying detail, with the most detailed being the 'antiquity models'. These are physical objects that can be handled, conserved, curated and, prior to digitisation, manually updated. As can be seen in Figs 1-4, cards contain additions in a number of different hands and inks so that additions and deletions can clearly be seen. The textual element of the information was the first to be computerised, starting with the NAR in 1992, using a mixed approach of scanning, optical character recognition and manual entry. The digitisation of NBR data followed. The digitisation of both the NAR and the NBR used a bespoke Oracle system held on a VAX 11/750 computer, accessed via dumb terminals. The process was completed in 1993, with the information being transferred to a bespoke Oracle database called MONARCH (Monuments and Archives). The name NMR for the combined record and archives was first used in 1995. The MONARCH system was then itself subject to two redevelopments (both also Oracle Databases), first NewHIS (New Heritage Information System) and then the current system AMIE (Archives and Monuments Information, England) in 2001. AMIE has since been modified on numerous occasions. Figure 5 shows the Monuments module in the current version of AMIE. The first thing to note is how structured the information is. Although many of the data concepts are comparable to the information on the original '495 Cards', other units of information have been added as data standards have evolved. Individual records originating from the cards have had information added to them, and entirely new database records have been created that never had a card index predecessor.

Figure 5

Figure 5: A monument record in the NMRs current AMIE system (© English Heritage)

This raises two interesting philosophical points: if an index card is a physical object, then how do we perceive its intangible digital modern equivalent: the database record? If a record has been added to over time, whether obviously through the differing handwriting and inks on the card index, or through additions to a database record, what is the original record? On a card it is clear if something is either crossed out or added, but it still remains the same physical object; on a database this can be less clear. If, for example, a user was to delete a sentence, then substitute something else, there would be no way of telling that changes had been made simply by looking at the screen or printing the document. A good well-designed system such as AMIE will contain an automated audit trail with date and time stamping, but that will not enable the user to see exactly what has been changed. If a new card is added then that is both a physical object in its own right but also part of a larger object, the card index. So how does this compare to a new record being added to the intangible and virtual modern database equivalent? The NMR does not have a cut-off date for recording. Consequently, the digitised version of the monument record now contains records for sites that post-date the most recent card in the physical predecessor.

To draw an analogy with a museum object, the artefact may have been altered during its time in use. It was then subsequently restored and then conserved by the museum once it was viewed as a piece of material culture worthy of preservation. How much of it is original and how much does this matter?

Map depictions of archaeological sites raise the same issues. The sites recorded on the cards were marked by hand onto 1:10,000 scale OS quarter sheets. Figure 6 shows a section of one of these quarter sheets, clearly showing it as a physical object, curling at the corner, and not just a piece of information (something that will be returned to later). These were first digitised onto a mapping system running in a CAD environment to accompany the VAX records mentioned earlier. Sites recorded in the Monument Record are now recorded on layers in a Geographic Information System (GIS) linked to the database records in AMIE. As with the textual database, the GIS has now been transferred to a newer system. Figure 7 shows how part of the same area appears in the GIS. One thing to note is how the number of sites has increased, as well as the differences in the depiction of individual sites. However, with both the spatial and textual information these are conceptually the same information, irrespective of the medium and the post-digitisation alteration, just with improved depiction and accuracy.

Figure 6 Figure 7

Figure 6: A section of the OS 1:10,000 scale annotated map for an area of Dorset (© Crown copyright. NMR, map © Crown copyright. Ordnance Survey)
Figure 7: Screenshot from the webGIS system in use at EH today displaying NMR monument data at 1:10,000 scale for part of the area shown in Figure 6 (© English Heritage and © Crown copyright. Ordnance Survey. All rights reserved. English Heritage 100019088. 2010)

2.2 Statutory data

Since the merger of the RCHME and English Heritage in 1999, the NMR has been responsible for records of sites with statutory protection (Listed Buildings, Scheduled Monuments, Registered Parks and Gardens, Registered Battlefields and Protected Wrecks).

2.2.1 Scheduled Monuments

The first archaeological sites were protected by the Ancient Monuments Protection Act in 1882. Scheduling, as it is known today, dates from the 1979 Ancient Monument and Archaeological Areas Act. In 1882 a mere 26 sites were protected in England (Breeze 2006, 57), whereas the schedule today contains 19,271 (English Heritage 2009, 20). Scheduled Monuments have gone from card index, paper forms, hand-annotated maplets and county map volumes to a series of stand-alone county databases built in a database package called Superfile, and then a national system built in Oracle: the Record of Scheduled Monuments (RSM) with a stand-alone computer mapping system and eventually a linked GIS (see Figure 8).

Figure 8

Figure 8: The changing way Scheduled Monuments are recorded from physical objects: a) index cards (top left), b) maplet folder (top right) and c) county map volume (bottom left), to the digital: d) the RSM (bottom right and e) GIS layer (centre), ( a) © English Heritage, b and c) © English Heritage, map © Crown copyright. Ordnance Survey, d) © English Heritage, e) © Crown Copyright, map © Crown copyright. Ordnance Survey. All rights reserved. English Heritage 100019088. 2010).

2.2.2 Listed buildings

Listed Buildings as we know them go back to the Town and Country Planning Acts of 1944 and 1947 (Harvey 1993, 1) and were the last major dataset to be computerised. There were calls for digitisation from the 1980s, including Sir Bernard Feilden, writing in the Transactions of the Ancient Monuments Society, where he called for a system that could answer questions such as 'when was a structure built'? (Feilden 1986, 67-68). The Listed Building record was eventually computerised by late 1994, in a major project that included two years of indexing (limited to the data included in the list entries and conforming to data standards), as well as scanning, optical character recognition and proofreading. This turned the bound volumes (the 'Greenbacks' and 'Bluebacks') and loose-leaf amendments, together with maps hand-annotated by the listing inspectors, into an Oracle database - the Listed Buildings System (LBS) and accompanying GIS layers, as seen in this example from the City of Bath (Figure 9). Conceptually, these still remain distinct as the List and the Schedule.

Figure 9

Figure 9: Listed Buildings recording from (top left to bottom right) a) volumes and amendments, and b) hand-annotated maps, to c) the Listed Buildings System and d) a GIS layer ( a) © Crown Copyright, b) © English Heritage, map © Crown copyright, c) © English Heritage and d) © Crown Copyright. All rights reserved. English Heritage 100019088. 2010).

3. An analogy from left of field

Introduction | The NMR's datasets | An analogy from left of field | The materiality of information | How does the medium influence... | Conclusion

A useful analogy from the world of physical objects is the case of 'Old Number One'. This famous Bentley was the first 6-speed racing car and owes its name to its start number at the 1929 Le Mans, where it won the first of two back-to-back victories. At Brooklands in 1932 during a 500 mile race 'Old Number One' lost a wheel while overtaking a Bugatti at the top of the track and left the banking at 120 mph. Tragically, the driver, Clive Dunfee, was thrown onto the track and killed and his car was smashed to pieces. However, the car is still in running order today, or is it? In 1990 'Old Number One' was the subject of a court case. The car had been rebuilt after the crash. When it was sold in 1990 for £10 million the authenticity of the car was questioned by the purchasers. The case went to the High Court in London and the arguments hinged on the number of times the car was rebuilt during its racing career as well as after the 1932 accident. To quote from the judgement:

'It was this history and reputation, as well as its metal, which was for sale on 7th April 1990...the expression Old Number justifiably applied to the car which in a succession of forms raced at Le Mans between 1929 and 1932 when it crashed. It is the "authentic" "Old Number One"' (Judgement (as revised) by Mr Justice Otton in the case of Hubbard v Middlebridge Scimitar Limited, Royal Courts of Justice London, 27 July 1990)

The analogy is that the records curated today at the NMR are still in essence the same entity as the original although added to, deleted, amended, and enhanced, even though the record today now contains sites built after the original record was begun. This is conceptually quite a fundamental point, with the same logic applying to the intangible and virtual whereas in the case of 'Old Number One' applies to the physical and, in this case, metallic.

4. The materiality of information or when is the intangible material?

Introduction | The NMR's datasets | An analogy from left of field | The materiality of information | How does the medium influence... | Conclusion

So with these parallels between physical records, such as a card index, and maps and their digital equivalent, if a card index can be viewed as material culture in the same way as a pot from an archaeological context, then can the digital record also be seen as a modern piece of material culture in the same way as the racing car? Important to this is the notion of materiality, 'the quality of being material' (Dark 1995, 204). In this sense, once divorced from the physical the information is no longer material in a physical sense; however, it remains material in the sense of relevance and it is its meaning that is crucial to it as an entity. It is the idea of materiality as 'that which continues the matter or material of something' (Brown 1993, 1714) which is pertinent and it is this that is maintained once the object has been transferred to the virtual and intangible. The semantic context is also worth considering. The form of some objects provide evidence of their meaning, so it is with the record cards. A viewer will instinctively know its purpose is the storage of information, but the same is equally true of a database screen such as the one shown in Figure 5 when it is displayed on a monitor, even though it does not have a physical presence in the same way.

Both the card index and the database are historical documents, repositories of information that have evolved over time as well as representing technological developments. This is true in two senses. The hardware and code of the applications that run on it demonstrate the history of technology in the same way as the furnaces preserved at Ironbridge. But in a deeper sense it is the information itself that is a historical artefact. The records are descriptions of something material on a medium that is itself material, although conceptually it is the description rather than the medium which is the object of study.

Jones discusses the relationship between material culture and language and the 'power of language to consider the material world on the construction of language' (Jones 2009, 95), yet in this instance it is the language that is a fundamental component of the object and, as Deetz suggests, language can be seen as material culture (Deetz 1967, 9).

It can be argued that one thing which both the information as discussed here, and the more traditional examples of material culture (such as items recovered during an excavation) have in common is that they can be the focus of study in their own right and can be used to inform our understanding of the society that created them. In this case, this includes the evolution of archaeology and the historic environment as a discipline and a profession. Having information in digital form enables interesting questions to be asked of it and it has the potential to contribute to the understanding of the study of the past itself. As previously mentioned, monuments being recorded now post-date the record they are contributing to. So whereas barrows have always been regarded as belonging to antiquity, when was interest first shown in them in the more recent past? For example, when was something as modern and mundane as a public convenience first recorded as a monument of historic interest? The earliest example of such recording using NMR data was a listing from 1976. What does this tell us about the notion of what comprises the historic environment? This sort of analysis of the data rather than the sites might enable other research questions to be asked. For example, could a correlation be done between houses with certain names like 'castle' or 'Waterloo' and dates, to see what that might say about a sense of history, place and nationalism at the time they were built or given the name?

This type of analysis is possible through databases and enables a reflexive analysis of recording practices to be undertaken to assess the inherent biases that are created over time as differing people undertake recording. This self-analysis is required for a recursive approach to recording to be adopted (Newman 2010, 6). Another question might be to examine how the information concepts used in different manual and computerised information systems changed over time. Was there, for instance, really a greater emphasis on metrics and quantification as one might anticipate during the New Archaeology of the 1960s and 1970s, compared with a greater use of free text fields as post-modern attitudes prevailed? As Huggett observes:

'systems theory in the New Archaeology...which had its origins in a mechanistic view of the world as represented most clearly in the computers which facilitated the archaeological analysis undertaken and which by definition required a formalised, algorithmic, definition of tasks and data in order to operate.' (Huggett 2004)

Computer systems to facilitate archaeological recording have always had to balance a need for structured indexing to facilitate retrieval with descriptive free-text elements. Such structured fields do not always result in quantification. However, there was certainly greater prescription for elements such as entity names, characteristics and field lengths in the data standards that preceded the current MIDAS Heritage standard (FISH 2007). Modern terminology standards permit a level of granularity in recording that is scaleable through the use of hierarchical terminology control such as the Thesaurus of Monument Types.

5. How does the medium influence what is recorded and how it is perceived?

Introduction | The NMR's datasets | An analogy from left of field | The materiality of information | How does the medium influence... | Conclusion

Tilley observes that 'material culture is a relational and crucial category leading us to reflect on object-subject relations' (Tilley 2001, 258). If the idea that material culture is active rather than passive and that it acts back upon the subject is accepted, and if information is seen as material culture and the subject is the user of the information, then it is understandable that it influences the user in a less than objective way. If this is happening here then it is likely that the medium and the materiality will be influencing the perception of the information by the user. It could be argued that something has been lost in translation as the information changes its carrier. Many of the earlier physical systems, in particular the OS antiquity models, were almost works of art. There was an intangible quality about them that itself suggested something about the nature of scholarship at the time, whereas desk-based recording today can have a disconnection with the original material. Moreover, it is this illustrative detail that adds to the understanding of the site, which is one reason why it is important to retain the previous physical versions as well as the modern systems that replaced them as an index into this original data. As Bowden and Mackay observed:

'Perhaps one of the saddest elements of the OS legacy is that the computerisation of the record cards, but not the antiquity models or illustration cards, has created the impression today that the primary OS product was text. This has led to a loss of connection between survey drawing and text, and an over-reliance on the printed summary text descriptions rather than pictures or survey drawings.' (Bowden and Mackay 1999, 11)

This is true to an extent, but the key word here is 'impression', as the computerised record acts as an index into the original archive. Database recording has, however, placed a greater emphasis on the textual side of recording with the accompanying GIS being a later addition. And the annotation of paper maps and subsequent digitisation into GIS have also changed the nature of the record, removing the artistic subtlety of the antiquity models. Dissemination of the record via the web (i.e. could also be seen as amplifying the impression of the dominance of text mentioned by Bowden and Mackay, since initially this lacked a mapped element. However, the change in emphasis in recording through computerisation has brought advantages in retrieval, analysis and dissemination that could never have been achieved through the use of the previous manual systems.

The NMR has recently carried out an enhancement project to the Monument Record designed to provide an index into original archives using the Monuments Protection Programme Industrial Archaeology Step Three Reports. These reports categorised all known sites according to the industry represented with a view to scheduling. These have greatly increased the component of industrial sites in the national record, with 2,250 being added or enhanced. The process has also enabled these data to be interrogated for the first time by means other than by industry and county.

Some of the issues concerning recording from card indexes into a database are still being considered in the planning of a new project using the archive created by the National Record of Industrial Monuments. This record was conceived in 1959 at a conference organised by the Council for British Archaeology (CBA), but it did not commence until 1963. Recording was carried out by volunteers who were sent record cards by the CBA, with the project being administered by the University of Bath. The project was eventually closed, with the cards being transferred to the NMR (then part of the RCHME) in 1981 (Falconer 2000, 59-60). These cards contained a similar mix of structured and free text and illustration that were seen earlier with the OS cards.

There is a growing interest in preserving old computing hardware, as demonstrated by the National Museum of Computing at Bletchley Park and the Museum of Computing in Swindon. However, is anyone showing a similar interest in preserving bespoke software systems of the type used for recording the historic environment? Should they be? With the Listed Buildings System and Record of Scheduled Monuments being replaced by a single system and AMIE also due for replacement, such archiving needs to be considered. The Record of Scheduled Monuments was preceded by a series of separate county databases and a national overview database, the CNSR (County Number Summary Record). There are no longer copies of any of these predecessor systems, which were all built in a now-obsolete database package called Superfile. Information can be divorced from its system, as demonstrated by the National Archive's National Digital Archive of Datasets project. Data stored in Extensible Markup Language (XML) could arguably be seen as the purest form of this. However, the user then has a sterile experience that does not replicate the experience of using the original system and which itself creates a different perception of the data.

Gordon Childe wrote that 'Material culture is thus largely a response to an environment' (Childe 1942, 20). It can be argued that the digital record of the past (even more so than its physical descendant) is the same. The data are created and amended in order to meet a requirement from the historic environment sector whether to record, manage or designate. Similarly, systems are changed to ensure that they deliver what is required by practitioners to respond to sectoral change and to utilise the latest technological developments. In terms of content of the monument record, for example, the NMR no longer attempts to provide a comprehensive record of the totality of the historic environment in England. Instead the NMR works to a series of criteria, which are set out in a philosophical guide to the Monument Record as part of the internal recording guidelines, in order to avoid duplication of recording with local historic environment records.

6. Conclusion

Introduction | The NMR's datasets | An analogy from left of field | The materiality of information? | How does the medium influence... | Conclusion

Archaeological interpretations of the past often deal with causality, whether for example the introduction of a new technology can be shown to have brought about an observable change. So it is with information technology. In this case study, an archaeological approach to the history of recording shows new technology bringing about change and being used to respond to wider changes to ensure a record continues to be relevant to the society it serves.

This case study has also shown that information (in this instance concerning the historic environment) can be divorced from its medium, whether physical or digital. Additions and alterations to a record do not change the fundamental nature of the record itself. However, they can change the way the record is perceived by its users. The introduction of digital technology has changed the nature of what is recorded but it has not removed its previous aspects although it can be argued that it has altered the perception of the user. These digital records have brought advances in retrieval, analysis and dissemination and provide the potential for the data to be used in new ways for the study of both archaeology itself and society at the time the records were created.

With an eye on the future, the Listed Buildings System and Record of Scheduled Monuments are, at the time of writing, in the process of being replaced by a Unified Designation System. AMIE itself will be due for replacement in the near future and GIS will be integrated into new developments. So the evolution of these recording systems is continuing.

O.G.S. Crawford was the first archaeological officer of the OS in 1920 and went on to become one of the original recorders of the National Buildings Record (Hauser 2008, 57 and 233). Crawford was fascinated by recording the modern world through the eyes of an archaeologist. In many ways the NMR today is continuing his legacy, and maybe he would approve of the approaches being taken to update the record with contemporary sites which will be tomorrow's past.

7. Postscript

Since this was written, a reorganisation has taken place within English Heritage and the National Monuments Record no longer exists as a separate entity. The monuments record, together with information on Listed Buildings, Scheduled Monuments, Registered Parks and Gardens, Registered Battlefields and Protected Wrecks, is now part of the Designations Department. The archives, together with their cataloguing, are now in the same group as English Heritage's historic properties and collections. Far from being the end, however, this is just the beginning of a new phase in the story of the recording of the Historic Environment in England.


I would like to thank my colleagues at the NMR for their assistance, especially Rod Fitzgerald for assisting with his knowledge of the OS Archaeology Division's recording systems and the work of the RCHME pre-digitisation, and Edmund Lee and Dr Gillian Grayson for their comments on draft versions of this contribution. I would also like to thank Hilary Malaws from the RCAHMW for pointing me towards references about the history of the RCHME and Heather Dean from Yale University for the quote on material culture with which I began this article.


© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions | File last updated: Tue Feb 22 2011