The ideal Benedictine Monastery: From the Saint Gall map to ontologies

Abstract

We present an OWL 2 ontology, called SaintGall, representing the Saint Gall plan, one of the most ancient documents arrived intact to us. The Saint Gall plan describes the ideal model of a Benedictine monastic complex that inspired the design of many European monasteries. The structural, functional, and architectural specification of an ideal Benedectine monastery is modeled by the SaintGall ontology, which allows one to analyse and model the Monastery architectural type.

This work started with the purpose of relating Catania’s San Nicolò l’Arena Benedectine Monastery with the abstract notion of Benedectine monastery, in the ambit of an ontological model based on the renovation works carried out by the architect Giancarlo De Carlo and developed by the same authors.

The SaintGall ontology opens a research path aiming at comparing different monastic architectures that can be useful in any intervention of refurbishment or design for a monastery.

Keywords

Ontology OWL 2 Digital Humanities Benedictine Monasteries

1. Introduction

The project of valorisation of Catania’s Benedictine Monastery “San Nicolò l’Arena”, currently carried on by the University of Catania, required the realization of an information system allowing the different agents involved in the process to manipulate information concerning the Sicilian late-baroque building nowadays hosting the Department of Humanities and, in particular to query, integrate, share, understand, and modify data concerning its contemporary renovation (1977–2005). In view of the project requirements, the authors proposed an ontology, called ArchivioMuseoFabbrica, representing the most important features of the Monastery and describing the rich archival heritage arisen directly from De Carlo’s building site and storing pictures, diaries, projects, letters, photos, architectural plastics, and so on Cantale et al. (2017b).1

¹
https://cutt.ly/hhrURyj

Besides information on documents and people, the ontology includes a section devoted to the internal description of the building. The latter has been defined from the “Progetto Guida” – designed by the architect Giancarlo De Carlo for the renovation of the Benedictine Monastery (1988) – in order to associate each area of the Monastery with the related documents stored inside the building site.

The ArchivioMuseoFabbrica ontology was developed with the aim of treating the contemporary aspect and the readaptation of the specific building. In fact, it has been conceived to reconstruct relationships among areas involved in the renovation work and people and companies entrusted with the performance of specific tasks. In light of such characteristics, the ontology is very useful for refurbishment activities of the specific monastery. However, the Benedictine Monastery of Catania, as any other historical building refurbished over the years to host modern activities, must be primarily considered in light of its original characteristics. Indeed, a faithful analysis of its original functions and architectural features is required by scholars in Architecture, in Urban Planning, and in Art, to study the evolution history of the building and to comprehend the social and economical changes underlying the architectural choices of refurbishment and re-adaptation phases. Such considerations open up a new path of research consisting in the construction of an ontology of “the ideal monastery”. To accomplish such goal, it was necessary to find a more generic source of knowledge permitting to describe the building in an “atomic” way, namely to consider the smallest characterizing parts of it and to appropriately combine them in a larger landscape.

The scroll of parchment of the Saint Gall plan, dated back to the 9th century A.D. and stored in the Stiftsbibliothek Sankt Gallen, is a kind of a meta project, namely the program of designing of how a Benedectine monastery should be built; it is a precious document not only from historical and artistic points of view – it presents itself as a conceptual map of “the ideal monastery” – but also because the organization of the information is easily readable by means of its high-resolution digitalization.2

The project of digitalization and metadatation of the Saint Gall plan lasted approximately six years and gave to the research community the possibility to remotely vision the scroll in every detail, from front to back, and to zoom and rotate it. There is also a search tool for the metadata included in the map. The project, available at http://www.stgallplan.org/, also includes the Manuscripts of the Reichenau Monastery.

Thanks to the schematic nature of the Saint Gall plan and since it has been conceived with the purpose of organizing the spaces according to the monastic life-style proclaimed by the San Benedict rule, it is a perfect source for the development of an ontology for “the ideal monastery”.

In this paper, we propose an OWL 2 ontology, called SaintGall, defining the architectural type of Benedictine Monastery identified by the Saint Gall scroll.3

A preliminary version of this work is outlined by Cantale et al. (2017a).

Usefulness of the ontology is three-fold. From one side, it helps in identifying similarities and differences between given real monasteries and the Saint Gall map and to comprehend social, economical, and environmental motivations lying behind such differences. In addition, the SaintGall ontology is also a valid support for studying the architectural design of buildings erected by communities different from the Benedictine ones, such as the Franciscan or Jesuit ones, but with analogous communal life, that were founded few centuries later or outside the European borders.

From another side, the ontology helps in re-discovering the original conception of the spaces of monasteries nowadays refurbished to accommodate the functionality of a modern building, also considering the fact that monasteries, in their turn, are evolution of the nucleus of more ancient architectural artefacts that changed their function over time (Grandinetti, 2017; Giani, 2010).

Finally, the SaintGall ontology turns out to be useful in the design of the restoration and refurbishment processes of buildings belonging to a monastic complex and it helps in capturing significant logical, semantic, geometric, and functional knowledge about elements used when designing a monastery or analogous historical buildings.

The SaintGall ontology has been developed by taking into account structural, architectural, and functional details of the buildings included in the plan, together with the textual descriptions provided by Willis (1848). The major problems we faced when we modeled the Willis’ textual descriptions (1848) are the correct interpretation of the socio-cultural context and the lack of encyclopedic and architectural knowledge of the author. Our interpretation of the text, however, is confirmed by the studies of more recent authors (Gombrich, 1950; De Vecchi and Cerchiari (1995); Kubach, 2001).

The paper is organized as follows. Section 2 presents some related work. Section 3 is devoted to the SaintGall ontology. Specifically, Section 3.1 provides some historical notes concerning Benedictine monasteries; Section 3.2 is devoted to the Saint Gall map; Section 3.3 introduces some upper-level ontologies used in the definition of the SaintGall ontology; Section 3.4 presents the ontology and its modelling choices; and Section 3.5 is devoted to the evaluation methods considered in the developing phase of the ontology. Finally, Section 4 concludes the paper, where we draw also some hints for future work.

2. Related work

Ontologies can be applied as a useful tool to represent architectural artefacts and historical buildings, and to describe their conservation state and design process Acierno et al. (2016). Cantale et al. (2017b) have applied ontologies to formally represent Catania’s Benedectine Monastery and its process of refurbishment and renovation. However, spaces and architectural elements of the monastery according to the architect’s refurbishment project have been considered there, while the analysis of their original structure and function has been neglected.

Literature is rich in research concerning the application of ontologies to the design for architectural heritage. In fact, the theoretical baggage of architecture, involving a great variety of knowledge (historical, technical, artistic, subjective knowledge, and so on), fits well to be explored and modeled via ontological methods Roussey et al. (2011). Moreover, ontologies can be extremely useful to manage, organize, and share such a baggage in an unambiguous way, in particular when agents need to unravel knowledge, data, and information concerning some design tasks. In particular, ontologies are applied to conceptual design in architecture in order to facilitate knowledge sharing of shapes and their properties Emdanat and Vakalo (1998), and to manage cartographic urban themes and the complex multi-dimensional data of architectural heritage Noardo (2016).

Recently, documentation processes concerning the conservation of the architectural heritage and of the ex novo architectural design have been carried out focusing efforts and contributions on knowledge representation and management of existing elements, training memories, and personal remembrances (Stufano Melone, 2011; Stufano Melone and Borgo, 2016; Maietti et al., 2017).

Design in architecture is an expression of critical conscience detached from reality and derives from the special personal universe of the architect and her/his “inventive memory”. The analysis of memories (disaggregated knowledge) can be possible through ontological tools (Hois et al., 2009; Borri et al., 2014). According to some literature, the architectural type is a kind of epistemological foundational knowledge. Specifically, an architectural type is a logical statement that describes a formal structure (i.e., a shape) Martís Arís (1990). Types regulate relational and functional systems, a kind of “mental projects” that can be considered pre-representations of the space preceding the design stage Muratori (1960). An ontological analysis can make explicit, clarify, and manage the complexity of references, constraints, functional goals, and designer’s ambition through a cognitively centered design process. Moreover, it helps in highlighting and coherently separating the key elements of the architectural creative process Melone and Borgo (2016).

Types in architecture have been listed, explained, and discussed over time in the attempt to introduce and specialize notions such as form and function. For example, in Palladio’s (1570) and Durand’s (1799) books we find listed some. They are about architectural objects and urban frames referred to buildings and infrastructures intended to accomplish specific functions (burial monuments and hospitals, city walls, city doors, streets, triumphal arches, bridges, public squares, public buildings, temples/churches, palaces, courthouses, colleges, libraries, museums, observatories, markets, lighthouses, customs trade fairs, and theatres).

There exist studies about how to manage via ontological analysis the intrinsic knowledge of architectural types as a first step to explore all the memories with the aim of supporting design processes in both the restoration and refurbishment of architectural heritage (Acierno et al., 2016; Pauwels et al., 2018). Despite the fact that literature is rich of contributions concerning ontologies for architectural design, as far as we know this paper is the first attempt to formally define the architectural type of a Benedictine monastery and to analyze such kind of historical buildings from the point of view of their original conception and function.

3. The ontology of Saint Gall plan

We devote this section to the description of the SaintGall ontology. With this in mind, in Section 3.1 we preliminarly give a brief historical introduction to Benedectine monasteries, with a focus on the Saint Gall map; the latter is further described in Section 3.2; in Section 3.3 we briefly report on the foundational ontologies used in the SaintGall ontology. Section 3.4 is devoted to the description of the ontology. Finally, in Section 3.5, we introduce the evaluation methods of the SaintGall ontology.

3.1. Historical notes on Benedictine monasteries

Monasteries were conceived by the Benedictine monastic order, founded by Saint Benedict of Nursia, after the collapse of the Western Roman Empire. The monastic shape aims at preserving the European Christianity inside small self-sustaining communities where to lead a life of mystic and religious contemplation and introspection. The main principle is to protect and shield Christian religion and tradition from barbarian invasions. In the tradition related to Saint Benedict, the theme of Goths is recurring. For instance, we recall the famous episode of the meeting between the God’s servant Benedict and the king Totila, commander of the Goths army.

Monasteries differ from convents primarily because of their purpose. Monasteries are inhabited by monks belonging to some monastic order such as the Benedictine one, having an ascetic and solitary lifestyle. Convents, originating later with the mendicant orders such as the Franciscan one, are more dependent by the outside world. The two religious constructions arose in different historical periods, carrying out different functions inside the religious community. Monasteries are often also abbeys that are spaces where the nullius diocesis is effective. Such norm, in the canon law, represents the independence of a church and of the related monastery from the diocese in which the building is located. Therefore, the abbot is a substitute for the bishop inside the Benedictine “village”.

«Monasticism has its root in the interpretation of the Christian faith developed in the theology of the 6th century firstly in Orient. Analogously to theology and architecture, it is subjected to a deep transformation in Occident. [ ... ] The Benedictine Order remains for a long time the principal one. Hundreds of convents and monasteries are spread across the Christian Europe and represent cells of Christian tradition and faith, of science, and of culture» Kubach (1972).

Since the foundation of the Monastery of Montecassino (the first Benedictine monastery) in the first half of the 6th century, the Benedictine Order has been subjected to incessant internal reforms, especially when periods of spiritual and moral weakening took over the strictness of the Order. Reforms often influenced the housing architecture of the monks.

Starting from the 7th century, Western Europe is characterized by a capillary network of monasteries, preserved more than castles, public and laic works, precisely because of their religious function Gombrich (1950). Their shape in Occident remained largely unchanged in its characteristics during the whole Middle Age and in all Christian countries, with differences in the volume of some spaces. For example, in more recent periods, larger libraries are built in order to host bigger amounts of books.

The plan of Saint Gall, illustrated in Fig. 1, is a model of monastery better representing the Benedictine architecture since it is strongly inspired by the rule of Saint Benedict, and can be considered a fixed-type for the Middle Age monasteries De Vecchi and Cerchiari (1995).

Fig. 1.

The plan of Saint Gall.

The plan has been designed in the context of Pre-Romanesque Carolingian art and architecture. «[...] The planimetry of Saint Gall Monastery provides the ideal committing schema adopted until the late-Gothic period: the monks dwellings are located on three sides of a rectangular or quadratic courtyard (cloister), while on the other side (on north or on south) runs one side of the church. The internal court is entirely surrounded by a quadrangular arcade that links the buildings. Besides the oriental wing of the cloister, as a rule, we find two small rooms assigned to the sacristy and armarium (book storage), and one room assigned to the monks reunions where they discuss all the community matters under the supervision of the abbot or the prior.

On the east side of the church, there is the monks dormitory, located on the upper floor for space connectivity reasons and because of tasks to be performed inside the church. Often, in front of the church there is the refectory, the dinner room, and the kitchen of the monastery, while on the occidental wing there are other buildings, services, and the entrance.» Kubach (1972). Thus, rooms of common use such as dormitories turn out to be of great relevance. They are usually located beside a cloister containing a little garden or a green area, and, in the middle, a fountain or a water source for the ablutions. The cloister was devoted to contemplation of beauty and meditation, being an open area in a seclusion space. Inside the fortification, right out the cloister area, there are locations assigned to the farms and the dwelling of animals.

As shown in Fig. 1, the space arrangement is inspired by that of a small village: a large church in the centre with annexed cloister and dormitory; at East one can find the novitiate, the infirmary, the cemetery, and the orchards, at South the factories, at North the guest houses, schools, and the dwelling of the abbot, and, finally, at West the farms. Thanks to its shape, the Monastery accomplishes the “nullius diocesis”, being self-sustained and independent from the outside world, and the “ora et labora” rule professed by San Benedict.

3.2. The Saint Gall map and real Benedectine monasteries

Being one of the most ancient descriptions of primitive Benedictine monastery arrived intact today, the Saint Gall map turns out to be an important structural, architectonic, and functional landmark for the Benedictine monasteries. Indeed, many European monasteries are inspired by the Saint Gall plan even though for practical and technical reasons they deviated from it.

In what follows we display the plans of some Benedectine monasteries. Our aim is not to enter the history of each represented monastery, but to indicate the reader the characteristics that all these artefacts have in common, and the shared program that subsides them.

In Fig. 2 (representing the Beaulieu Abbey, in Hampshire, England, founded in 1203–1204), in Fig. 3 (showing the Kirkstall Abbey, in Kirkstall, England, built in 1152), in Fig. 4 (displaying the Montmajour Abbey, in Arles, France, founded in 949), in Fig. 5 (showing the S. Michele in Vulture Abbey, Monticchio, Italy, constructed in the VIII Century), in Fig. 6 (showing a fictional monastery for the Umberto Eco’s novel “Il nome della Rosa”), in Fig. 7 (depicting the Monastery of La Tourette, in Éveux, France, built by LeCorbusier in 1956), and in Fig. 8 (representing the Monastery of San Nicolò l’Arena in Catania, Italy, founded in 1558), we highlighted the church (in red), the cloister (in green), and the dormitory (in blue), as fixed elements of such representations. Such elements are always strictly interconnected in similar successions of spaces, even though difficult topographical conditions required a kind of contraction of the shapes and of the relations between the different elements, as in the S. Michele in Vulture Abbey (Fig. 5).

⁴
Image credited to www.historyfish.net.

Fig. 2.

Beaulieu Abbey, Hampshire, England.4

⁵

Image credited to www.1902encyclopedia.com.

Fig. 3.

Kirkstall Abbey, Kirkstall, England.5

⁶

Image credited to www.heritage-route.com, Aix-Marseille Université.

Fig. 4.

Montmajour Abbey, Arles, France.6

⁷

Image credited to “I monasteri italogreci e benedettini in Basilicata”, Luigi Bubbico, Luigi Caputo, Attilio Maurano, 1996, per il Ministero dei Beni Culturali e Ambientali, Sopraintenenza per i Beni Culturali e Architettonici della Basilicata.

Fig. 5.

S. Michele in Vulture, Monticchio, Italy.7

⁸

Image credited to www.architecturalpapers.ch.

Fig. 6.

Sketch of the Monastery in the novel “Il nome della Rosa” by Umberto Eco.8

⁹

Image credited to www.GreatBuildings.com.

Fig. 7.

Monastery of La Tourette, Éveux, France.9

Fig. 8.

Monastery of San Nicolò l’Arena, Catania, Italy.10

The strength of setting offered by the Saint Gall plan program, namely the fact that it represents a meta design, what should be there and where, travels across centuries. Two emblematic examples are the imaginary monastery represented in Fig. 7 and La Tourette Monastery depicted in Fig. 8, both conceived in the second half the XX century.

Finally, the Benedictine Monastery of San Nicolò l’Arena in Catania (Sicily) (Leonardi and Cantale (2017); Mannino, 2015; Atripaldi and Costa (2008); DeCarlo, 1988; Librando (1988)) contains most of the elements of the Saint Gall plan with the exception of some locations such as the brewery that, for cultural reasons, has been replaced by a distillery. Moreover, Catania’s Benedictine Monastery is a urban monastery and therefore the structure of the animal farms is also slightly modified. Such questions are treated in some details in Section 3.5.

¹⁰

Image credited to “Un Progetto per Catania”, Giancarlo De Carlo (1988).

3.3. Foundational ontologies used in the SaintGall ontology

Next, we briefly report on the two foundational ontologies used by the SaintGall ontology, namely the “CIDOC Conceptual Reference Model (CRM) (CRM; Doerr, 2003)” and the Europeana Data Model (EDM).11

¹¹
https://pro.europeana.eu

CIDOC-CRM is the international official standard ISO 21127:2006 for the controlled exchange of information in cultural heritage since 2006. It has been used by the SaintGall ontology with the purpose of uniforming its structure to the one peculiar of ontologies for cultural heritage.

CIDOC-CRM serves as guideline for good practices of conceptual modelling in the cultural heritage context, and provides a high-level set of specifications used to construct a semantic web-based information system which improves information sharing. Nowadays, many institutions are adopting CIDOC-CRM: galleries, libraries, museums, archives, and several other cultural environments based on cultural heritage information that publish and share semantic web data. The CIDOC-CRM core covers several general aspects of cultural information, such as material and immaterial entities, events, space, and time. The general concepts provided by CIDOC-CRM can be specialized, contextualized and integrated in order to address practical aspects of cultural heritage issues. CIDOC-CRM also introduces several notions acting through temporal entities such as participation, appellation, part-hood and structure, material and immaterial concepts, location, assessment and identification, purpose, motivation, and so on. In this context, actors operate or deal with material and immaterial stuff within a certain temporal window. Such participation can occur in some real or figurative place. Moreover, everything can be classified in more details by introducing a “type” and can be identified by some appellations such as name, labels, or title. The existence of objects that have a persistent life-span such as birth, death, creation, destruction, and so on, can be limited in specific temporal windows. Activities and events that introduce some changes of state in cultural, social, or physical systems can be modelled in CIDOC-CRM together with their results or products. Finally, objects and events that influenced or motivated activities can be also modelled by means of different forms of influence and their mutual relations.

EDM is an ontology for collecting, connecting, and enriching cultural heritage information provided by content providers. In fact, EDM introduces several entities both new and re-used from other namespaces such as the Simple Knowledge Organization System (SKOS) namespace,12

¹²

http://www.w3.org/2004/02/skos/core

an ontology for supporting the use of knowledge organization systems such as thesauri and classification schemes, the Dublin Core namespaces,13

¹³

http://purl.org/dc/elements/1.1, http://purl.org/dc/terms, http://purl.org/dc/dcmitype/

a set of vocabulary terms that can be used to describe digital and physical resources, the OAI Object Reuse andExchange (ORE),14

¹⁴

http://www.openarchives.org/ore/terms/

an ontology for aggregation of resources, and so on.

EDM has been used by the SaintGall ontology to establish a connection with on-line objects from European museums, libraries, archives, and multimedia applications. As formally described in Section 3.4, we used three classes of EDM, namely PhysicalThing, Place, and Agent, to allow the definition of specific ontologies based on the SaintGall ontology that describe entities belonging to museums, archives, and public or private collections.

3.4. Description of the SaintGall ontology

The SaintGall ontology15

¹⁵
https://goo.gl/XN2hc3

is a representation of the elements of the Saint Gall map, namely buildings and green spaces, that takes into account their cardinal orientation, their position with respect to other entities inside the plan, their shape, and their architectural, structural, and functional features. The need of modeling the spatial arrangement of the buildings in the Saint Gall map and of their cardinal position is motivated by the fact that these are fundamental for the daily activities of a monastery: for example to maximize the exposure to the sunlight of certain buildings such as the library, or to minimize the movements of the monks, for instance from the church to the cloister. Furthermore, the ontology models the role of people living inside the monastery.

The SaintGall ontology exploits the following main classes:

Building, describing a generic building,

Garden, specifing a generic green space, and

Element, describing architectural, natural, ornamental, and votive elements, furniture, and tools of common use that are illustrated in the plan.

These are defined as subclasses of the CIDOC-CRM class E18 Physical Thing and of the EDM class PhysicalThing.

One of the most important partition criteria of buildings in the Saint Gall map concerns their functionality. The map identifies five main groups of buildings: buildings for education, for hospitality, for sick and infirm people, factories and farms, and monastic buildings, the latter group including the church, which receives a special consideration. Consequently, we modelled the functional areas of the monastery, by classifying the buildings represented on the map according to their intended use and position as reported in the Saint Gall map and in (Willis, 1848; De Vecchi and Cerchiari (1995); D’Onofrio, 1994; Gombrich, 1950; Kubach, 2001). Specifically, we introduced as subclasses of Building the pairwise disjoint classes BuildingForEducation, BuildingForHospitality, BuildingForTheSickAndInfirm, FarmBuilding, and PrincipalMonasticBuilding.

The class BuildingForEducation includes, in particular, the subclass School, modeling a building intended for the education of scholars, and the subclass NoviceCloister, representing the novice cloister, dwelling of young people oriented to the monastic life. The class BuildingForHospitality contains among others the class HospitiumDistingueshedGuests, modeling the hospitium for the reception of eminent strangers, and the subclass HospitiumPoorTravelersPilgrims, representing the dwelling of poor travellers and pilgrims. The class BuildingForTheSickAndInfirm contains in particular the subclass InfirmaryCloister, representing the cloister where the sick brethren are lodged, and the subclass DoctorHouse, containing among others a private room for the physician and a room for very ill patients. The class FarmBuilding models the factory, the working house, and other buildings devoted to domestic cattle, poultry, and their keepers. The class PrincipalMonasticBuilding includes in particular the subclasses AbbotHouse, modeling the dwelling of the abbot, TheCloister, describing the cloister where monks live, and TheChurch, describing the abbey. The hierarchy of Building is shown in Fig. 9.

The green spaces inside the monastery are modeled by means of the class Garden, having the disjoint subclasses Cemetery, KitchenGarden, and PhysicGarden.

The classes Building and Garden are related by means of the CIDOC-CRM object-property E156 occupies to the classes BuildingPlace and GardenPlace, whose instances represent places in which buildings and gardens occur. Cardinal orientations of places where buildings and gardens occur are modeled by the classes CardinalDirection and CentralPosition, both defined as subclasses of the CIDOC-CRM class E53 Place, and by the object-property hasPosition, subproperty of the CIDOC-CRM object-property P89 falls within. The class CardinalDirection is a finite enumeration of the values East, North, NorthEast, NorthWest, South, SouthEast, SouthWest, and West. The class CentralPosition contains only the individual Centre. In addition, we introduced the defined classes CentralArea, EastArea, NorthArea, NorthEastArea, NorthWestArea, WestArea, SouthEastArea, SouthWestArea, and SouthArea, all of them subclasses of the EDM class Place and of the CIDOC-CRM class E53 Place, whose subclasses represent places in which buildings and gardens of the monastery occur. Figure 10 shows the description of the class NorthArea, while Fig. 11 illustrates the inferred hierarchy of NorthArea.

Fig. 9.

Subclass hierarchy of Building.

Fig. 10.

Description of NorthArea.

Fig. 11.

Inferred hierarchy of NorthArea.

We define the position of buildings or gardens in the map with respect to other contiguous buildings or gardens by means of the object-properties onEastOf, onNorthEastOf, onNorthOf, onNorthWestOf, onSouthEastOf, onSouthOf, onSouthWestOf, and onWestOf, subproperties of the CIDOC-CRM object-property P89 falls within, where onEastOf is the inverse of onWestOf, onNorthOf of onSouthOf, onNorthWestOf of onSouthEastOf, and onNorthEastOf of onSouthWestOf.

For the sake of modelling the shape, the size, and the internal structure of buildings and gardens, we introduce the following entities: the class Shape (subclass of the CIDOC-CRM class E26 Physical Feature, which models the shape of structures, and whose subclass hierarchy is shown in Fig. 12) and the associated object-property hasShape (subproperty of the CIDOC-CRM object-property P56 bears feature); the class Size (modeling the size of buildings and a subclass of the CIDOC-CRM class E54 Dimension) and the related object-property hasSize (subproperty of the CIDOC-CRM object-property P43 has dimension).

Fig. 12.

Subclass hierarchy of Shape.

In the SaintGall ontology, buildings are described together with their essential architectural elements, tools, and furniture necessary for the activities of the monastery, according to Willis (1848), by means of the subclasses of the class Element, ArchitecturalElement, Tool, and Forniture, respectively.

The class ArchitecturalElement describes general architectural elements inside the map, the class Tool represents tools of common use such as Furnace and Boiler, and the class Furniture models objects used in everyday life such as Bedstand and Desk.

In order to relate buildings with the elements they contain, we provided the object-property contains, subproperty of the CIDOC-CRM object-property P46 is composed of, together with its subproperties consistOf, containsAround, and so on (see Fig. 14). We also introduce the object-property isContainedIn, inverse of the object-property contains, together with its subproperties isPartOf, isContainedAround, and so on (see Fig. 14), in order to easily reconstruct relationships among buildings and elements. The hierarchy of subclasses of Element and of their related properties are illustrated in Figs 13 and 14.

Fig. 13.

Subclass hierarchy of Element.

Fig. 14.

Object-properties related to the class Element.

In Fig. 15 we show the picture of the abbot house as depicted in the Saint Gall map, and we report our OWL 2 model of the abbot house, as represented by the tool Protégé.16

¹⁶

https://protege.stanford.edu/

Fig. 15.

Example of the abbot house.

This building, inhabited by the abbot, is surrounded by a fence. In accordance with the nullius diocesis rule, it is a major building consisting of two stories of which the lower one has an open portico on the east and west sides. The inner space is split into two chambers, namely the abbot sleeping and sitting rooms. The upper story contains some small chambers and one large chamber. Details concerning the furniture of the abbot sleeping and sitting rooms are modeled by the classes AbbotSleepingRoom and AbbotSittingRoom, respectively, which are subclasses of Chamber.

As Fig. 15 shows in the specific case of the abbot house, our OWL 2 representation of the Saint Gall map was developed so as to be as faithful as possible to the textual description by Willis (1848) and to the map.

Another important case is represented by the monk cloister which is reported in Fig. 16 (inside the black circle). Even in this case, our OWL 2 representation tightly reflects the map and the descriptions by Willis (1848) and Kubach (2001), the latter partially reported in Section 3.1.

Fig. 16.

Detail of the monks cloister (inside the black circle).

The cloister court is located on the south wing of the church. It is constituted by a large area bounded on the north by the church and on the other sides by three important buildings of the monastery: the parlour, the dormitory, and the refectory. The parlour serves as a vestibule to the cloister and as a space where servants receive orders by superiors and visitors can meet the monks. The refectory is used as dining and amusement room. The cloister also contains a passage to a cellar with larder for storing food. It is surrounded by a covered walk, called porticus, and contains in the centre an open space with a fountain or water source, probably laid out with grass or shrubs. The cloister is characterized by the presence of a portico, rested on a low perimetric wall and admitting light inside the spaces annexed to the gallery. For each side of the cloister, there is a big central arch that gives access to the garden. The cloister connects the different areas of the cenoby; it allows monks to use the portico as a place where to rest and have shelter from bad weather.

Fig. 17.

Description of TheCloister.

Our OWL 2 model of the monk cloister, as represented by the tool Protégé, is illustrated in Fig. 17.

Persons are classified according to the place in which they live and work for most of the day by means of the object-properties livesIn and worksIn, both subproperties of the CIDOC-CRM object-property P55 has current location. For example, the abbot is defined as a monk who lives in the abbot house, as depicted in Fig. 18.

We introduce a hierarchy having as root the class Person, defined as subclass of the CIDOC-CRM class E21 Person and of the EDM class Agent, to model people living in the monastery (see Fig. 19).

Fig. 18.

Description of the abbot of a monastery.

Fig. 19.

Hierarchy describing persons living and working in a monastery.

Table 1

SaintGall ontology structural metrics

Metric type	SaintGall metric value
DL expressivity	SHOIQ(D)
Axioms count	1955
Logical axioms count	1103
Declaration axioms count	564
Classes count	482
Total classes count	482
Object-properties count	64
Data-properties count	3
Individuals count	13
Subclass axioms count	1161
Equivalent classes axioms count	19
Disjoint classes axioms count	16
Hidden GCI count	18
Subobject-property axioms count	53
Inverse object-properties axioms count	23
Transitive object-properties axioms count	8
Symmetric object-properties axioms count	2
Object-property range axioms count	2
Data-property domain axioms count	3
Data-property range axioms count	3
Class assertion axioms count	13
Same individual axioms count	2
Different individuals axioms count	1
Annotation assertion axioms count	85

3.5. Metrics and evaluation of the SaintGall

This section discusses structural metrics and evaluation criteria for the SaintGall ontology. As shown in Table 1, the SaintGall ontology consists of more than 480 classes, more than 60 object-properties, about 1300 logical axioms, and it exploits OWL 2 constructs such as existential restriction and qualified cardinality restriction. Many object-properties defined in the SaintGall ontology are subproperties of some CIDOC-CRM properties and, therefore, they inherit domains and ranges of the latter. For only two of them, namely hasShape and hasPosition, we felt the need of a further restriction of their range. Specifically, the object-property hasShape, subproperty of the CIDOC property P56 bears feature, has as range the class Shape, subclass of the CIDOC class E26 Physical Feature, the latter being range of the property P56 bears feature. We defined the range of the object-property hasPosition, subproperty of CIDOC-CRM object-property P89 falls within, to be the union of the classes CardinalDirection and CentralPosition, both subclasses of the CIDOC-CRM class E53 Place, the latter being range of the object-property P89 falls within.

The main evaluation criteria for the SaintGall ontology have been computed using the OntoQA approach (Tartir et al. (2005)). OntoQA is a feature-based method for evaluating ontologies applying techniques that do not require data training and that involve users in a minimal way. In OntoQA, metrics are divided into two categories, schema metrics (relationship richness, inheritance richness, tree balance, and attribute richness) addressing the design of the schema of an ontology and instance metrics (class richness, average population, and instance coverage) concerning the way instances are organized within an ontology. We chose OntoQA instead of other similar approaches such as OntoClean (Guarino et al. (2004)) because OntoQA provides a user-friendly application system that, given an ontology, automatically computes the corresponding metrics.

The results are shown in Table 2 and are explained below.

Table 2
OntoQA metric results

Metric type SaintGall metric value

Relationship Richness 52

Inheritance richness 4.91

Tree balance 3.23

Attribute richness 0.05

Class Richness 71.73

Average Population 0.02

Instance Coverage 0.35

Metric type	SaintGall metric value
Relationship Richness	52
Inheritance richness	4.91
Tree balance	3.23
Attribute richness	0.05
Class Richness	71.73
Average Population	0.02
Instance Coverage	0.35

The relationship richness metric reflects the diversity of the relations and the placement of the relations occurring in the ontology. An ontology that contains more property relations other than class/subclass relations is richer than a taxonomy (with only class/subclass relationships). The value of relationship richness is a percentage representing how much relationships between classes are rich with respect to all of the possible connections (inheritance and properties). With a value of $52 %$ , the SaintGall ontology has a good balancing between generic relationships and class hierarchies.

The inheritance richness describes the distribution of information across the different levels of the inheritance tree of the classes. It is an indicator of how well knowledge in the ontology is grouped into distinct categories and subcategories. This measure can help to distinguish a horizontal ontology from a vertical one, or from an ontology with different levels of specialization. A horizontal (or flat) ontology has a small number of inheritance levels, and each class has a relatively large number of subclasses. In contrast, a vertical ontology contains a large number of inheritance levels where classes have a small number of subclasses. An ontology with a low inheritance richness would be of a vertical nature, which might reflect the fact that the ontology represents a very specific and well detailed knowledge. An ontology with a high value of inheritance richness has a horizontal nature, which means that the ontology represents a wide range of general knowledge. The SaintGall ontology reports an inheritance richness of $4.91 %$ and hence has a vertical nature since it represents a specific knowledge described in detail.

The tree balance metric is referred to how much class hierarchies differ in deepness. This may be related to the fact that some hierarchies are very deep whereas others are not. The SaintGall ontology provides a tree balance of $3.23 %$ meaning that some hierarchies have been well described (i.e, Room) whereas others remain very general (i.e., VotiveElement).

The attribute richness calculates the average number of attributes per class, which gives insight into how much knowledge about classes is represented in the model. An ontology with a high value for attribute richness indicates that each class has averagely a high number of attributes, namely that it is specified in detail, whereas a low value might indicate that little information is provided about each class.

The class richness is related to how instances are distributed across classes. An ontology having a very low class richness does not have data exemplifying the knowledge represented in the model. On the other hand, a high value of such metric (close to $100 %$ ) indicates that the data represents most of the knowledge described in the considered ontology.

Finally, the average population is an indicator of the number of instances divided by the number of classes, whereas the instance coverage indicates how much each class is populated.

The SaintGall ontology has an attribute richness of $0.05 %$ , a class richness of $71.73 %$ , an average population of $0.02 %$ , and an instance coverage of $0.35 %$ . These values are typical for foundational ontologies such as the SaintGall ontology is.

As additional non-measurable evaluation criteria for the SaintGall ontology, we considered the ones illustrated in (Gómez-Pérez (2004); Obrst et al., 2007). Among them, the most significant ones are:

consistency: consists in verifying that contradictions cannot be inferred from the ontology;

clarity: consists in checking whether terms defined in the ontology effectively communicate the intended meaning and definitions are given in an objective way;

extendibility: regards the issue of checking whether the ontology offers a conceptual foundation for the considered domain and it can be extended and specialized monotonically (new terms can be introduced without the need of revising existing axioms).

The consistency criterium has been evaluated by checking the classification of the ontology with the Fact++ reasoner which performs well with ontologies admitting axioms containing existential (“Some values from”) and universal (“All values from”) OWL restrictions.17

¹⁷

http://owl.cs.manchester.ac.uk/tools/fact/

In order to guarantee that the clarity criterium is fulfilled, the ontology has been developed as a faithful OWL 2 representation of the Saint Gall chart and of its textual description by Willis (1848) (for example, see Fig. 15).

The extendibility criterium has been evaluated by integrating the ontology ArchivioMuseoFabbrica, representing the Benedectine Monastery of San Nicolò l’Arena in Catania, in the SaintGall ontology. In Table 3, the main relationships between the SaintGall and the ArchivioMuseoFabbrica ontologies are summarized.

Table 3

Comparison among the places of the SaintGall and the ArchivioMuseoFabbrica ontologies

SaintGall	Position	ArchivioMuseoFabbrica	Position
KitchenGarden	SouthEast	Not present
PoultryHouse	SouthEast	Not present
Cemetery	East	Not present
NoviceChapel	East	NoviceChapel	North
PhysicGarden	NorthEast	PhysicGarden	East
BuildingForTheSickandInfirm	East	BuildingForTheSickandInfirm	East
StallDomesticCattle	SouthWest	NotPresent
HouseHorses	SouthWest	HouseHorses	SouthWest
TheCloister	Centre	TheWestCloister	West
TheCloister	Centre	TheEastCloister	East
MonksBath	Centre	MonksBath	Center
CellarLarder	Centre	CellarLarder	Center
RefectoryClothesStore	Centre	RefectoryClothesStore	Center
BakeHouseBrewHouse	South	BakeHouseBrewHouse	North
HospitiumPoorTravellersPilgrims	Centre	HospitiumPoorTravellersPilgrims	East
MortarBuilding	South	Not Present
KilnBuilding	South	Not Present
Barn	South	Not Present
AbbotHouse	North	BakeHouseBrewHouse	North
HouseCountryServants	West	HouseCountryServants	South
HospitiumDistinguishedGuests	North	HospitiumDistinguishedGuests	North
TheChurch	Centre	TheChurch	Centre
SacristyVestry	Centre	SacristyVestry	Centre
Parlour	Centre	Not Present
ScriptoriumLibrary	Centre	ScriptoriumLibrary	Centre
PorterLodging	Centre	Not Present

The monastery of Catania has almost all the characteristics described by the SaintGall ontology such as a church, a garden of officinal plants, a stall for horses, a lodge for pilgrims, a refectory, a bake and brew house, a cellar for storing food, two quadrangular-shaped cloisters. Thus, it turned out that the ontology ArchivioMuseoFabbrica could be easily integrated in the SaintGall ontology without additional modifications of the taxonomy. For example, the ArchivioMuseoFabbrica individuals representing the two cloisters are automatically asserted as instances of the SaintGall class TheCloister.

4. Conclusions

We presented an ontology for the Saint Gall plan that describes the ideal model of a monastic Benedictine building. The plan allows one to study the most significant features of European monastic buildings such as the Monastery of San Nicolò l’Arena in Catania, the Abbey of Santa Giustina in Padua, the Cluny Abbey, and the Montecassino Abbey De Vecchi and Cerchiari (1995). We expect that by means of the SaintGall ontology, scholars and researchers in Human Science can effectively compare several distinct monastic architectures, and from their differences and similarities make inferences not only in architectonic and stylistic ambit, but also in the interpretative and theological areas Gombrich (1950).

In order to provide support to the analyses and to the actions concerning the architectural type of Monastery, it was deemed appropriate to model the ontology of an ideal monastery such as the Saint Gall map, which can be potentially extended/specialized, on a case-by-case basis, by providing the architectonic specification of the considered monastery. Indeed, the ontology introduced in this work can be considered as a general support tool for all the refurbishment, recovery, and change-in-use actions concerning a Benedictine monastery, or, more generally, the architectural type of Monastery.

We are currently considering the integration of the SaintGall ontology with other widespread ontologies such as DOLCE Masolo et al. (2004). Many generic classes from the SaintGall ontology, such as Church and Cloister, can be reused to design novel ontologies describing buildings outside the Benedectine context. As an instance, consider the architectonic structure of closed garden (cloister or court), which can be also found in municipal buildings.

The SaintGall ontology contains OWL constructs such as existential/universal quantification and cardinality restriction. Thus, it cannot be represented in the set-theoretic fragment considered by Cantone and Nicolosi-Asmundo (2013), used in recent work by some of the authors for representing and reasoning on ontologies, e.g., Cantone et al. (2015a) and Cantone et al. (2017), such as ArchivioMuseoFabbrica Cantale et al. (2017b), an ontology of the Monastery of Catania, and Ontoceramic Cantone et al. (2015b), an ontology for the classification of pottery. For this reason, we intend to investigate for decidability purposes a new set-theoretic fragment admitting the composition operator which would allow one to represent and reason about the SaintGall ontology. Results by Cantone et al. (2010) and Cantone et al. (2011) can be a starting point to construct an appropriate decision procedure for such a set-theoretic fragment.

References

Acierno, M., Cursi, S., Simeone, D. & Fiorani, D. (2016). Architectural heritage knowledge modelling: An ontology-based framework for conservation process. In Journal of Cultural Heritage (Vol. 24, pp. 124–133). France: Elsevier Masson. 2017.

Atripaldi, A.M. & Costa, M.E. (eds) (2008). Catania – architecture city landscape, Catania – architettura città paesaggio. Roma: Gruppo Mancosu Editore.

Borri, D., Camarda, D. & Stufano, M.R. (2014). Typological axioms for architecture. In

D’Amato Guerrieri (Ed.), Proceedings of the 1st Workshop on the State of the Art and Challenges of Research Efforts at Politecnico di Bari, Roma (pp. 17–22). ISBN 8849299796, 978884929979.

Cantale, C., Cantone, D., Lupica Rinato, M., Nicolosi-Asmundo, M. & Santamaria, D.F. (2017a). The shape of a Benedictine monastery: The SaintGall ontology. In Proceedings of the Joint Ontology Workshops 2017, Episode 3: The Tyrolean Autumn of Ontology, Bolzano, Italy, September 21–23, 2017. CEUR Workshop Proceedings (Vol. 2050).

Cantale, C., Cantone, D., Nicolosi-Asmundo, M. & Santamaria, D.F. (2017b). Distant reading through ontologies: The case study of Catania’s Benedictines monastery. JLIS.it, Italian Journal of Library, Archives, and Information Science, 8(3), 203–219.

Cantone, D., Longo, C., Nicolosi-Asmundo, M. & Santamaria, D.F. (2015a). Web ontology representation and reasoning via fragments of set theory. In

ten Cate and

Mileo (Eds.), Proceedings of Web Reasoning and Rule Systems: 9th International Conference (RR 2015). August 4–5, 2015, Berlin, Germany, Lecture Notes in Computer Science (Vol. 9209, pp. 61–76). Cham: Springer. doi:10.1007/978-3-319-22002-4_6.

Cantone, D. & Nicolosi-Asmundo, M. (2013). On the satisfiability problem for a 4-level quantified syllogistic and some applications to modal logic. Fundamenta Informaticae, 124(4), 427–448. doi:10.3233/FI-2013-842.

Cantone, D., Nicolosi-Asmundo, M. & Orłowska, E. (2010). Dual tableau-based decision procedures for some relational logics. In Proceedings of the 25th Italian Conference on Computational Logic, CEUR-WS, Rende, Italy (Vol. 598, pp. 7–9).

Cantone, D., Nicolosi-Asmundo, M. & Orłowska, E. (2011). Dual tableau-based decision procedures for relational logics with restricted composition operator. Journal of Applied Non-Classical Logics, 21(2), 177–200. doi:10.3166/jancl.21.177-200.

10.

Cantone, D., Nicolosi-Asmundo, M. & Santamaria, D.F. (2017). A set-theoretic approach to ABox reasoning services. In

Costantini,

Franconi ,

Van Woensel ,

Kontchakov ,

Sadri and

Roman (Eds.), Proceedings of the International Joint Conference, RuleML+RR 2017. Lecture Notes in Computer Science London, UK (Vol. 10364, pp. 12–15). Cham: Springer. 87–102, 2017.

11.

Cantone, D., Nicolosi-Asmundo, M., Santamaria, D.F. & Trapani, F. OntoCeramic: An OWL ontology for ceramics classification. In

Ancona ,

Maratea and

Mascardi (Eds.), Proceedings of the 30th Italian Conference on Computational Logic. CEUR Workshop Proceedings Genova, Italy, July 1–3, 2015, (Vol. 1459, pp. 122–127).

12.

De Carlo, G. (1988). Un progetto per Catania. Il recupero del Monastero di San Nicolò l’Arena per l’Università. Genova. (edited by Daniele Brancolino), Sagep.

13.

De Vecchi, P. & Cerchiari, E. (1995). Il Medioevo in Arte nel Tempo, Vol. I, Tomo II. Milano.

14.

Doerr, M. (2003). The CIDOC CRM. An ontological approach to semantic interoperability of metadata. AI Magazine, 24(3), 75–92.

15.

D’Onofrio, M. (1994). Il panorama dell’architettura religiosa, in I Normanni popolo d’Europa 1030-1200, Roma Palazzo Venezia, Venice, January 28–May 30.

16.

Durand, J.N.L. Précis des Lecons d’Architecture, Paris: Gillé.

17.

Emdanat, S. & Vakalo, E. (1998). An ontology for conceptual design in architecture. CAADRIA, 98.

18.

Giani, E. (2010). Caratteri tipologici e distributivi degli edifici.

19.

Gombrich, E. (1950). La storia dell’Arte. London: Phaidon.

20.

Gómez-Pérez, A. (2004). Ontology evaluation. In

Staab and

Studer (Eds.), Handbook on Ontologies (pp. 251–274). Springer. First Edition, chapter 13. doi:10.1007/978-3-540-24750-0_13.

21.

Grandinetti, P. (2017). Tipi architettonici e regole compositive nell’ architettura e nella città antica: la Grecia.

22.

Guarino, N., Welty, N. & Christofer, A. (2004). An overview of OntoClean. In Handbook on Ontologies (pp. 151–171). ISBN 978-3-540-24750-0. doi:10.1007/978-3-540-24750-0_8.

23.

Hois, J., Bhatt, M. & Kutz, O. (2009). Modular ontologies for architectural design. Formal ontologies meet industry. In Proceedings of the 4th Workshop FOMI 2009, September 2, 2009, Vicenza, Italy, in Association with the 10th European Conference on Knowledge Management. Frontiers in Artificial Intelligence and Applications (Vol. 198, pp. 66–77). IOS Press. ISBN: 978-1-60750-047-6.

24.

Kubach, H.E. (1972). Architettura Romanica in Storia Universale dell’Architettura (consulted ed.). Milano: Mi. 2001.

25.

Leonardi, A. & Cantale, C. (a cura di). (2017). 105 Lettere di Giancarlo De Carlo sul Recupero del Monastero dei Benedettini di San Nicolò L’Arena a Catania, Editoriale Agorà, Catania.

26.

Librando, V. (1988). Notizie storiche sul monastero di San Nicolò L’Arena. In Un progetto per Catania, Sagep, Genova.

27.

Maietti, F., Di Giulio, R., Balzani, M., Piaia, E., Medici, M. & Ferrari, F. (2017). Digital memory and integrated data capturing: Innovations for an inclusive cultural heritage in Europe through 3D semantic modelling. In

Ioannides ,

Magnenat-Thalmann and

Papagiannakis (Eds.), Mixed Reality and Gamification for Cultural Heritage, Cham: Springer.

28.

Mannino, F.

(a cura di) (2015). In Breve storia del Monastero dei Bendettini di Catania, Giuseppe Maimone editore, Catania.

29.

Martís Arís, C. (1990). Le variazioni dell’identità. Venezia, CLUP.

30.

Masolo, C., Vieu, L., Bottazzi, E., Catenacci, C., Ferrario, R., Gangemi, A. & Guarino, N. (2004). Social roles and their descriptions. In KR (pp. 267–277).

31.

Melone, M.R.S. & Borgo, S. (2016). In Towards an Understanding of Shapes and Types in Architecture (Vol. 1616). CEUR-ws.org. ISSN 1613-0073.

32.

Muratori, S. (1960). Saggi di critica e di metodo nello studio dell’architettura (1946).

Marinucci (Ed.). Centro Studi di Storia Urbanistica, Roma, 1980.

33.

Noardo, F. (2016). A spatial ontology for architectural heritage information. In Cédric Grueau, Robert Laurini, Jorge Gustavo Rocha, Geographical Information Systems Theory, Applications and Management Second International Conference, GISTAM 2016, Rome, Italy, April 26–27. Springer.

34.

Obrst, L., Ceusters, W., Mani, I., Ray, S. & Smith, B. (2007). The evaluation of ontologies. In

C.J.O.

Baker and

K.-H.

Cheung (Eds.), Revolutionizing Knowledge Discovery in the Life Sciences, chapter 7 (pp. 139–158). Springer.

35.

Palladio, A. (1570). I quattro libri dell’architettura. Dominico de Franceschi, Venezia.

36.

Pauwels, P., Poveda-Villalón, M., Sicilia, A. & Euzenat, J. (2018). Semantic technologies and interoperability in the built environment. In Semantic Web (Vol. 9, pp. 731–734). IOS Press. 2018.

37.

Roussey, C., Pinet, F., Kang, M.A. & Corcho, O. (2011). An Introduction to Ontologies and Ontology Engineering. Ontologies in Urban Development Projects. Advanced Information and Knowledge Processing (Vol. 1). London: Springer.

38.

Stufano Melone, M.R. (2011). Il ruolo della creatività nel processo progettuale in architettura: una approccio cognitivista. Ph.D. thesis, University of Pisa, Ph.D. in Models and Methods for the European City and Territory.

39.

Tartir, S., Arpinar, I.B., Moore, M., Sheth, A.P. & Aleman-Meza, B. (2005). OntoQA: Metric-based ontology quality analysis. In Proceedings of IEEE Workshop on Knowledge Acquisition from Distributed, Autonomous, Semantically Heterogeneous Data and Knowledge Sources.

40.

Willis, R. (1848). Description of the ancient of the monastery of Saint Gall in the ninth century. The Archaelogical Journal.