DHQ: Digital Humanities Quarterly
2021
Volume 15 Number 3
Volume 15 Number 3
Probing Through Iranian Architectural History Within the Framework of an Ontology Development Process
Abstract
This paper presents a prototype ontology developed in the field of Iranian architectural history. The paper’s central arguments offer a response to questions regarding how to create an ontology in Iranian architectural history, what consideration must be addressed here, and how to resolve problematic issues in developing an ontology in a field such as Iranian architectural history, which lacks a formalized knowledge. The paper is organized into two parts. It primarily presents a discussion on the specific domain of architectural history and what it encompasses and moves on to examine why it is deemed complicated. After that, the process of creating Iranian architectural history ontology and the methodology applied to match the intended domain is explained. In the second part of the paper, the content of the developed ontology is discussed, which includes various parts of the ontology and what it implies to illustrate how the structure of ontology helps in logically representing the domain in a machine-readable format.
Introduction
To a great extent, what is known about architectural history stems from its practice,
discussions in its various fields of study, and the accumulation of that information
into knowledge. Such information has partly been expressed in an explicit form shared
between different parties (from people to software agents). This mode of sharing
calls for a mutual understanding of the content in the intended domain to unify the
conception of different groups involved in the subject matter. The intended consensus
also relies on interaction in a formal realm with a unified language. Therefore, in
order to make such interaction possible in the field of architectural history, one
needs to identify the domain and express the relevant knowledge in a clear and
explicit manner. To do so, it seems viable to apply the tools designed specifically
for the purpose.
Over the past few decades, in the field of computer science and artificial
intelligence, a branch called “ontology” has developed, which provides the
mentioned clarity in expressing the domain knowledge both by providing an inventory
of key concepts and treating the relations between the concepts.
With the progress of science and technology, human beings’ ability to create new and
advanced tools accelerated to the point of surpassing mundane affairs. Such tools
were born to replicate human functions. Richard Sennett offers an interesting
categorization of human-made tools, which he places into two groups: the replicant
and the robot. A replicant is a tool that mirrors humans by mimicking their actions.
However, a robot is an ameliorated human; it is more robust, works faster, and never
tires. Still, we make sense of its functions by referring to our own human measures
[Sennett 2008]. The mentioned functions may range from automatically
completing an order, saving things in its memory, finding and retrieving certain
items in bulk, or other tasks. Such robot-like tools are established on a somewhat
simple basis: mirroring the human mind. Unlike the human mind, computers by character
lack the willpower to learn and recall information, and it is up to us to accumulate
the knowledge into its memory and decide how information is recounted (described) for
machines.
In this paper, we aim to acknowledge the necessity of creating an intelligent
robot-like system in the field of Iranian architectural history in the form of a
semantic network; a system capable of functioning faster and more precise than humans
and able to assist in completing certain tasks that are naturally too difficult,
impossible, or subject to errors when done manually.
In order to provide these functions, it is necessary to familiarize the desired
system with the field of inquiry by describing what it encompasses. Henceforth, the
system would ground its additional functions based on clearly defined descriptions.
In other words, human knowledge needs to be presented to the machine.
All forms of knowledge representation, including ontologies, are both mediums of
expression for human beings and ways to communicate with machines in order to tell
them about the world [Brewster et al. 2004]. Therefore, engaging with sets of
information related to the intended field, which needs to be categorized and
described explicitly and logically, is a crucial act while developing an ontology.
Accordingly, this research mainly applies ontology to create a clear, explicit, and
descriptive model of the semantic system in Iranian architectural history, taking
into account various aspects of the field based on information provided by prominent
scholars. In other words, here, the effort is focused on developing an ontology in
the aforementioned field of knowledge, which is both explicit and readable for the
machine, through capturing and representing the relevant entities in this field. What
ensues this approach helps organize the views and provides consensus between
different agents (people with each other or machines with people) on Iranian
architectural history concepts.
Ontology of Iranian architectural history: terminology
This title of “Ontology of Iranian architectural history”
consists of four terms: “ontology,”
“Iranian architecture,” and “history,” each of which offers multifaceted
denotation causing the original term to indicate different meanings. To clarify what
this title covers in the course of this paper, an overview of the specialized
terminology is explained below.
Ontology: In the terminology of Artificial Intelligence, the term Ontology may refer
to both the act of describing intended concepts explicitly and the specific outcome
that is constructed and saved in the form of an XML or UML file, which contains data
on concepts and their descriptions. That is so that we may have ontology files with
different purposes and save them on our computers or share them with others. It
should be noted that description in an ontology means providing formal definitions of
terms in the domain vocabulary and representing each concept’s information in a
relational form.
Concepts in an ontology must be structured in a way that is usually referred to as
linked data; that is, data interlinked with other data to make information readable
for machines and make queries within large bulks of information possible. Such links
represent relationships within an area of study, and the model that contains all
those concepts and relationships is what we call an ontology. Although at the heart
of each ontology there is hierarchical taxonomy, yet since an ontology allows more
complex and non-hierarchal relationships to be defined, it can be used to represent
the fluidity of the real world, where things can be related to each other in more
complex ways than are represented by a hierarchical tree-like structure. In that
sense, an ontology is more like a spider’s web [Blayney 2017].
Iranian architecture: Iran in this paper refers to the Iranian world, which
transcends modern Iran’s current political boundaries. Simultaneously, Iranian
architecture covers all architectural works associated with Iran’s territory
throughout its history.
History: Michael Stanford distinguishes two meanings of the term history. History
“may refer either to the course of events, what actually
happened. (history 1), or what is believed and written about those events (history
2). Sometimes they are distinguished as ‘history-as-event’ and
‘history-as-account.’”
[Stanford 1998].
Iranian architectural history
Based on Stanford’s distinction, Iranian architectural history can imply two
meanings, where architectural history may be considered as series of events,
actually occurred in the world (Iranian architectural history 1), and
architectural history as the act of studying and describing what has occurred in
the world (Iranian architectural history 2). Given that in developing ontologies,
we are appointed to capture how human beings characterize their concepts, rather
than actual events and whether or not they happened in a specific way, we need to
be associated with Iranian architectural history 2, as the scholarship of
architectural history is classified under it, and involves the study of
architectural history-as-event [Qayyoomi Bidhendi 2009].
Therefore, the intended domain comprises a broad spectrum of studies on
architectural evidence and different sorts of human architectural activities. In
other words, we regard Iranian architectural history as a series of reports that
narrate certain historical events related to works of architecture, architects,
the act of building.
process of creating The IArchHist ontology[1]
Architecture and architectural history are among cultural matters related to human
affairs that enfold all the changeability and historicity that such matters enjoy in
any given culture, including the Iranian culture. Additionally, unlike western
tradition, the historiography of Iranian architecture is not a longstanding practice.
Contemporary attempts to write Iranian architectural history have been performed
mainly by two groups of individuals: first, European travelers, archeologists, and
architects that came to Iran during the 20th century and composed reports of their
observations; second, Iranian specialists, such as archeologists, architects, and
authors, who were mostly employed in institutions like Society for National Heritage
(SNH) and were responsible for documenting historical buildings for the purpose of
conservation. Both groups lacked academic training in historiography. In other words,
the existing historiography of Iranian architecture has been developed by
non-historians in its academic sense. Therefore, Iranian architectural history has
yet to become an established academic discipline with formalized terminological and
theoretical peripheries. The inconsistent quality of historiography in the sphere of
Iranian architectural history has given rise to its close association with
architectural works, their chronology, and the vocabulary and methods that were used
to examine them throughout history.
Developing ontologies is drawn on logical relationships, so in branches of the
humanities, such as history, art, and architectural history, which are less concerned
with positive knowledge, creating an ontology that requires data derived from reasons
and logical observation poses some challenges. On top of that, even in the western
scholarly world, architectural history is still highly dominated by the disciplines
of art history and history of architecture and is advancing towards a distinct
discipline. Accordingly, Iranian architectural history, in its current state, lacks a
formalized terminology and is still undergoing its formation process.[2]
Based on the facts explained about the field of Iranian architectural history,
describing the terms through their relationships is also challenging since the terms
have not been used uniformly in the context of writings from different scholars or
even a single source. In other words, the field’s scholars lack a unified language
when referring to concepts.
Since Iranian architectural history has yet to evolve as an academic field, its
existing sources involve subjective reports of buildings that are influenced by
personal feelings, tastes, or opinions of the narrator. This poses a challenge in
machinizing Iranian architectural history knowledge, which requires objective and
straightforward data.
However, the very challenges provoked by the field’s nature simultaneously make it
essential to undertake available tools (like ontologies) to organize its knowledge.
In fields such as Iranian architectural history, developing an ontology contributes
to identifying the field’s ambiguities and immaturities in a precise manner.
Therefore, here, the ontology provides advantages related to overseeing the field in
the course of its formation.
Premise
Usually, the first step for developing an ontology is to consider using the
existing models, or the least, an existing taxonomy. Over the past few decades,
several representational models have become available in the field of art,
architecture, and cultural heritage. These models include database schemata for
integrating heritage information, such as CIDOC CRM (CIDOC Conceptual Reference
Model) metadata for description of museum objects, such as Object ID, controlled
vocabularies of related fields like Getty art & architecture thesaurus (AAT).
Assuming that we were to use concepts represented in AAT, for instance, each
concept must be an exact equivalent in Persian and English, in terms of
definition, implication, and relations to other concepts; otherwise, the goal to
represent the reality of Iranian architectural history identities would not be
fulfilled. This level of correspondence simply does not exist. That is because the
Persian-speaking societies of art and architecture naturally do not match up with
the anglosphere one. Many of the identities discussed here are either only found
in Persian-speaking cultures, or even if found elsewhere, do not encompass the
exact same thing, mainly due to cultural differences that are often neglected in
universal attempts to develop comprehensive formal ontologies. Therefore, in most
cases, when observed carefully, concepts (that are embedded in words), albeit
close in meaning, are not identical in the Persian and anglosphere worlds. Based
on these considerations, the IArchHist ontology was developed as a unilingual
ontology in Persian.
The most basic premise held in developing the IArchHist ontology is that it is
possible to represent the knowledge of architectural history through its
identified concepts. Such concepts must be thoroughly defined in the course of
developing the ontology and be clarified to reach the processing level of the
machines.
Here an issue arises: how to make relatively ambiguous concepts understandable for
machines. At this point, the second premise emerges, stating that concepts are
understood and defined through their association with other concepts. The human
mind identifies each concept’s meaning by placing it in an elaborate network of
its related concepts and recognizing how they are linked. The ontology replicates
the same process by representing existing relational links between concepts. As a
result, each concept is conceptualized within a specific structure, which is
processable for machines.
The scope of the inventory
As a discipline, Iranian architectural history involves certain concepts, which
are indefinite by nature, are not directly expressed, and therefore, cannot be
adequately described. For instance, there is the concept of “principle” in
Iranian historical buildings’ architectural design. Pirnia, a distinguished
Iranian architectural historian, proposes five principles of Iranian architecture
as follows: human scale, introversion, self-sufficiency, precluding futility,
structural soundness, and proportion [Pirnia 2004]. Although he
further describes these concepts in the course of his studies, because of their
propensity to diverse interpretations, it is hard or even impossible to catch a
clear and explicit, and more importantly, a consensual definition; hence, the
possibility of failure in representing these concepts through IArchHist ontology.
Therefore, it is essential to delimit the semantic boundary for the selection of
concepts in the relevant field.
The fundamental criterion is that each IArchHist ontology entity requires a formal
description that is logical and certified, which would explain the concept it
represents. Therefore, all those concepts that are describable in logical phrases
clearly and explicitly are covered in the IArchHist ontology, while evaluative or
rhetoric concepts are avoided.[3] Consequently, the project scope covers Iranian
architectural historians’ works who referred to both the material evidence (the
relics) and the textual evidence of Iranian architecture.
The information used to build the IArchHist ontology was collected from
contemporary (written or published during the 80s and 90s in Iran) domain-specific
works.[4] The appointed references consist of
Iranian scholars’ works from four main fields of architectural history, historical
geography, archeology, and restoration, and include classes of objects, events,
processes, states of affairs, attributes of objects, parts of objects, segments of
processes. The selected scholars and their compositions are considered references
and major sources within Iranian architectural history.
Methodology
Noy & McGuinness (2001) proposed an outline for
developing ontologies, which involves the following seven steps: 1) Determine the
domain and scope of the ontology; 2) Consider reusing existing ontologies; 3)
Enumerate essential terms in the ontology; 4) Define the classes and the class
hierarchy; 5) Define the properties of classes — slots; 6) Define the facets of
the slots; 7) Create instances. In developing the IArchHist, the methodology
proposed by Noy & McGuinness is applied by simultaneously performing the
suggested steps and including additional measures when necessary. Following this
strategy, the methodology in building the IArchHist ontology is mapped out into
the following five stages:
- Steps 1 & 2: Determining the purpose and the scope; selecting the key references;[5] choosing the appropriate ontology language and software.
- Steps 3 & 4: A taxonomy is an orderly classification for
a defined domain. It comprises controlled vocabulary terms (generally only
preferred terms) organized into a hierarchical structure. Each term in a
taxonomy is in one or more parent/child (broader/ narrower) relationships to
other terms in the taxonomy [Harpring 2010]. In order to create
the taxonomy for IArchHist ontology, two steps were necessary: These steps are explained in detail in section 4.
- 1. Forming the inventory: cataloging key concepts from information provided by references;
- 2. Categorization: grouping related concepts together, organizing them in a hierarchical order, and defining upper-level classes.
- Steps 5 & 6: In order to describe the internal structure of the domain concepts outlined in the previous step, we must identify and draw semantic relations between concepts, which have been suggested and specified by domain scholars. Each link must connect at least two concepts producing precise, unambiguous text definition for every pair of concepts that can be expressed in short logical phrases.
- Step 7: The defined concepts in the previous steps mainly cover general entities indicating each concept’s types and genre. In order to make the IArchHist ontology richer and more detailed, we must assign instances to each general concept. Such instances inherit the relations previously outlined between their prime concepts.
- Visualization: As soon as the ontology is built, the challenge of how to use it, how to explore it, and what data to recall arises. Visualization is mainly based on mapping the information to a graphical representation to facilitate data interpretation. It also provides ways to limit the amount of data users receive while keeping them “aware” of the full information [Silva et al. 2012]. The type of visualization depends on the user’s intention; however, concerning IArchHist ontology, direct 2D graphs was deemed the most acceptable since it is more familiar and direct for users.
IArchHist ontology
The Taxonomy of IArchHist Ontology
By implementing the conventional methods of building ontologies in the domain of
Iranian architectural history, a representative model of the domain was created,
which will be thoroughly described in the next section.
The IArchHist ontology is an OWL 2 ontology created in Protégé 5 editor,
represents Iranian architectural concepts by taking into account different aspects
of the subject material. This ontology was developed in Persian. Capturing the
main concepts of the IArchHist ontology required building a large lexicon of
terms. Here, each term was significant because it referred to a concept of the
domain. Concepts, also called “Class” or “Type,” or “Universal,”
are fundamental building blocks of the IArchHist ontology, as any other. Each
class represents a set of instances (also known as individuals) that share one or
more common properties. These properties are identified according to the
information provided by key references. Whether physical or nonphysical, spatial
or temporal, each item within the domain has been thoroughly represented in the
ontology.
For example, the class of masjid-i jāmiʿ (Friday mosque)[6] includes all the
individuals that have the following properties:
- Is a type of masjid (mosque)
- Is a place for duʿā (prayer)
- Is a place for namāz-i Jumʿa (Friday prayer)
- Has a part minbar (pulpit)
To accurately record classes and their related properties, each phrase of the
mentioned texts containing related information was documented in a separate
datasheet. These phrases were labeled under entries named by the exact terms
utilized in the text in the original language.[7] The terms were
further employed as IArchHist ontology classes and subclasses. An example is shown
below (Table 1).[8]
| Text | Citation |
| A darāygāh (entrance) contains faḍā (space)s such as pīsh-utāq (anteroom), sardar (portal), dargāh (doorway), hashtī (vestibule), keryās (lobby), dehlīz (entrance hall), dālān (corridor). | [Pirnia 2004, 143] |
| Physical parts of the masjid (mosque), shaped around a ḥayāṭ (courtyard) and a dargāh (doorway), with a sardar (portal), and a pīshṭāq (arched portal) are adequately linked together, creating the architecture of the masjid (mosque). | [Pirnia 2004, 287] |
| Each bāgh (garden) comprised a darāygāh (entrance); a building sometimes called sardar (portal). This building was similar to bīrunī (the public or male quarters of wealthy households) part of traditional khāna (house)s, which was the place for padhīrāyī (receiving guests). | [Pirnia 2004, 432] |
| In sangī (stone) kārvānsarā (caravanserai), sardar (portal)s were made from ājur (brick). | [Pirnia 2004, 480] |
| masjid-i jāmiʿ-i ʿAbbāsī (Abbasi Friday mosque) has a sardar (portal) decorated with kāshī-yi tarāsh (mosaic tile) | [Pirnia 2004, 282] |
Table 1.
Example of information about sardar (portal)In Table 1, related information to the term sardar (portal) is shown.
A sardar (portal) is the main façade of some buildings in Iranian historical
architecture containing various elements used as a complex entrance system. Each
cited phrase contains related concepts to Sardar, which can be listed as follows
(Table 2):
| darāygāh (entrance) | ḥayāṭ (courtyard) |
| faḍā (space) | pīshṭāq (arched portal) |
| pīsh-utāq (anteroom) | bāgh (garden) |
| sardar (portal) | bīrūnī (the public or male quarters of wealthy households) |
| dargāh (doorway) | khāna (house) |
| hashtī (vestibule) | sang (stone) |
| keryās (lobby) | kārvānsarā (caravanserai) |
| dehlīz (entrance hall) | ājur (brick) |
| dālān (corridor) | masjid-i jāmiʿ-i ʿAbbāsī (Abbasi friday mosque) |
| masjid (mosque) | kāshī-yi tarāsh (mosaic tile) |
Table 2.
Concepts related to sardar (portal)These terms were imported to Protégé in a class-subclass order, which was
developed by employing a middle-out approach of classification.[9] The class hierarchy of sardar
(portal) is shown in Figure 1.
Figure 1.
Class hierarchy of sardar (portal)Interconnections of IArchHist ontology
The main challenge of developing the IArchHist ontology was describing the
concepts via their interconnected relations. This process passes beyond a
taxonomic description, which only lays out the type of each entity by a
hierarchical relation to its upper (broader) class. The ontological model was
further enriched by the description of the internal structure of concepts, which
was accomplished by determining each concept's properties.
According to OWL, properties may appear as data properties or
object properties. Data properties include all descriptive and
behavioral aspects that relate to the entities under consideration – such as their
geometrical, physical, and behavioral features – which are defined by specific
values associated with those attributes. Instead, the object properties represent
the connections that exist between each entity and the others, within and between
the involved knowledge domains [Acierno et al. 2017].
Since the key references provide information in natural language, to identify the
properties of each concept, we need to rewrite the information in logical phrases
so that the building units (concepts and relations) of each phrase can be inferred
and transferred into protégé. For example, in the case of sardar (portal), the
following statements are inferred:
- sardar (portal) is one of the architectural elements of a darāygāh (entrance).
- Some masjid (mosque)s have a sardar (portal)
- Some bāgh (garden)s have a sardar (portal)
- Some khāna (house)s have a sardar (portal)
- Some sardar (portal)s are made of ājur (brick)
- Some sardar (portal)s are decorated with kāshī-yi tarāsh(mosaic tile)
Relations in an ontology can be categorized into two groups: first, the general
relations (like hierarchy or whole-part), which are not specific to a particular
domain and can be found in any domain of knowledge. Second, are specific
relations, which are appointed to a particular domain. For example, Having
Material or Being Decorated By are two properties probably
only directed to the concepts found in the domain of Iranian architecture
history.
The above list includes some properties that relate a certain concept to sardar
(portal) and specify how it is related to sardar (portal). That is how an entity
called “sardar (portal)” participates in connections to other concepts. One
instance is explained below:
It has been said that in the masjid-i jāmiʿ-i ʿAbbāsī (Abbasi Friday mosque), the
sardar (portal) has been decorated with a type of tile (kāshī) called kāshī-yi
tarāsh (mosaic tile). According to this statement, a sardar (portal) (like the one
in Abbasi Friday mosque) can be decorated with kāshī-yi tarāsh. Therefore, to make
the intended connection, units of the statement can be separated:
- Concept 1: sardar (portal)
- Concept 2: kāshī-yi tarāsh (mosaic tile)
- Link (OWL object property): IsDecoratedWith
Figure 2.
Representative model of the relation between sardar (portal) & kāshī-yi
tarāsh (mosaic tile)Here, the property IsDecoratedWith describes a binary relationship between two
individuals of sardar and kāshī-yi tarāsh. Now consider all of the individuals
that have an IsDecoratedWith relationship to some other individual.
We can think of these individuals as belonging to the class of individuals with
some IsDecoratedWith relationship. What is noteworthy here is the
fact that “relationships” are also capable of defining a class of
individuals. In OWL, we can define such classes by using restrictions [Horridge 2011]. Using restrictions signifies how each supposed
relationship applies to the instances of each concept. For example:
- Do all instances of sardar (portal) have this property? (Is this property universal?)
- Do specific instances of sardar (portal) have this property? (Is this an existential property?)
- Quantifier Restrictions
- Cardinality Restrictions
- “hasValue” Restrictions
- An existential restriction describes a class of individuals with at least one (some) relationship along with a specified property to an individual that is a member of a specified class.
- Universal restrictions describe the set of individuals that, for a given property, only have relationships with other individuals that are members of a specific class [Horridge 2011].
Classes and Sub-classes in IArchHist ontology
By following the mentioned strategies, the first draft of IArchHist ontology was
developed covering 1924 concepts, interconnected by 103 object properties. The
critical component of the ontological model, being the concepts and their semantic
relations, provided an overview of the knowledge in the field of Iranian
architectural history and an understanding of the branches of concepts in this
domain.
According to IArchHist ontology, two upper-classes were defined to capture the
concepts of the field: “IArchHist concept” and “IArchHist related
concept” where the former encompasses the central concept of Iranian
architectural history and the latter those such as rituals, beliefs, periods,
which act as intermediary concepts of the field.
The class IArchHist concept contains the following subclasses (Figure
3):[10]
Figure 3.
IArchHist concept- Relic: this class includes objective concepts, which refer to the remained relics related to architecture. Each architectural relic has been the result of an agent's work (a maker); it was accomplished through specific actions and technologies; and has certain characteristics such as form, measurement, material, and layout. Architectural relics are categorized as follows (Figure 4):
Figure 4.Architectural relics - Characteristics: this class includes concepts that are somehow assignable to relics. Such concepts are widely used when describing certain architectural relics (Figure 5). The characteristics are either physical (such as measurements) or nonphysical (like functions).
Figure 5.Characteristics - Agent: This class includes the agents that participate in the development of an architectural relic. Agents are divided into two categories: human agents and non-human agents (Figure 6).Human agents can be further categorized regarding the state of their interaction with the relics. Some agents have a role in erecting buildings (including the architect, the contractor, the tile worker, the carpenter, etc.). Some help in the maintenance of the relics (guards, lightkeepers, etc.). Some are the user of relics (travelers, pilgrims, etc.), and some have instigated the formation of a relic through the needs or beliefs of the social group they belong to (like Christians, Europeans). Additionally, some agents have affected the formation or continuity of relics, despite their non-human origin (like gods or goddesses).
Figure 6.Agent - Action: This class includes all the actions involved in the creation or usage of a relic. It is further categorized into interaction and creation (Figure 7).
Figure 7.Action
Resolving challenges in IArchHist ontology
As we mentioned briefly in the “Interconnections” Section, The
main challenge in the development process of IArchHist ontology was representing
relations based on the data derived from sources. It is important to note that in
languages such as OWL, we are mainly able to draw binary relations (meaning we can
use properties that link only two individuals). However, in reality, the domain of
Iranian architectural history contains certain concepts that require in their
representation, a type of relationship that links an entity to more than just one
other entity or value. These relations are called n-ary relations [Noy and Rector 2006]. N-ary relations are more complex types of links, and
drawing them is usually quite tricky and often requires specific tactics.
Consider the following examples:
In this case, one might view the relationship as a binary relation between the individual Friday mosque of Fahraj and its ceiling and represent it as shown in Figure 8. However, here we see additional aspects of that relationship, including its ornamental attribute (being detailed) and the spatial quality it creates (being human-scaled) associated with it.The ceiling of Friday mosque of Fahraj is quite detailed but human-scaled.
In most ontologies, especially in humanities, such instances of a relation cannot
be viewed as an instance of a binary relation with additional attributes attached
to it. Instead, it is an illustration of the individual Friday mosque of Fahraj
and the complex object representing various facts about its constituents.
Following the general guidelines posed by Noy & Rector
(2006), in the IArchHist ontology, we represented such problematic
relationships by adding an intermediate class involving all sub-relations
associated with the broader relationship (Figure 9).
Figure 8.
Representing the relationship between the Friday mosque of Fahraj and its
ceiling with a binary relation
Figure 9.
Representing the relationship between the Friday mosque of Fahraj and its
ceiling with N-ary relationsIarchhist outcomes
As we have shown in the course of this paper, the nature of some fields, such as
architectural history, imposes somewhat theoretical regard to ontologies as knowledge
representation tools rather than their typical goals involving the establishment of a
shared understanding. The challenges faced when developing IArchHist, which for the
most part was related to ambiguity and indefiniteness of concepts, gave rise to a
strategy for outlining a methodology for developing the intended ontology, a
methodology which, unlike other domains, is more simultaneous rather than a
step-by-step procedure.
The development of the IArchHist ontology prompted three outcomes that would
significantly benefit scholars and researchers of Iranian architecture history. The
first outcome is the data collected while developing the ontology. The lack of
structured databases or other forms of organized data collections in the focused
field was a problematic research challenge. All of the field's main sources are
paper-based books and articles. The process of developing IArchHist ontology, as
explained earlier, simultaneously led to collecting data in a structured manner,
which included describing and representing various and, in some cases, several
relationships between each pair of data.
The second outcome is producing relatively precise descriptions and definitions for
each concept in the scope of IArchHist ontology. Since drawing the relationships
between concepts required uncovering each concept's nature and what it logically
encompasses, a descriptive set of literal definitions was created during IArchHist
ontology development. For example, when attempting to represent the concept of
"mosque" and drawing its relations to the concept of "dome," a series of logical
questions were raised: what is the nature of the relationship between mosques and
domes? Must all mosques encompass a dome? Should a building necessarily include a
dome in order to be called a mosque? If any building possesses a dome, can we then
call it a mosque? Attempting to answer these questions and others along the same line
resulted in uncovering the relations between the concepts, such as mosque and dome,
and ultimately made it possible to acquire a comprehensive and exclusive definition
for each concept.
The third outcome is that by objectively relating our concepts, we are able to
identify new patterns and pose new questions or find new explanations for formerly
resolved issues. For instance, the concept of architect is among the sub-classes of
agent in IArchHist ontology. Typically, vernacular architects are considered to be
responsible for building design. However, in IArchHist ontology, the class architect
is related to sub-classes of relic and characteristic through the following
properties:
- Architect designs building
- Architect erects vault
- Architect renovates building complex
- Architect calculates building geometry
Conclusion
This paper presents an overview of developing an ontology in the field of Iranian
architectural history named IArchHist ontology. The main challenge here was
responding to the question of "how knowledge representation with ontologies helps
organize data in a field consisting of a plethora of ambiguous and unstructured
information scattered in resources. As explained in the course of this paper, an
exclusive strategy was established, which was based on the development methodologies
of conventional ontologies. The results of this research promote two key ideas:
- Methodological approach: In addition to resolving the fundamental issues of the particular field, the process of IArchHist development provokes a new practice of understanding and studying knowledge, which takes place while describing ambiguous concepts and uncovering their logical relations. Accomplishing such purposes demands revealing the complex semantic network that draws the explicit or implicit connections between concepts.
- Developed product: The outcome of this research is basically a bounded part of Iranian architectural history ontology, which can be visualized and analyzed thoroughly with different techniques and further refined under field experts. That act of operating the different tasks, from describing to classifying and linking data sets at the same time, makes the process of developing IArchHist ontology much more valuable than the physical outcome. In order to reach an agreeable description for each concept, a constant reciprocating strategy was needed to improve the representative model of domain concepts continuously.
In conclusion, this research opens up new possibilities for further investigation
into the use of ontology as a tool for organizing concepts in the field of Iranian
architectural history by displaying their ambiguities and disorders. Furthermore, the
suggested classification system could contribute to thinking more clearly within the
focused field of knowledge through clear and unambiguous terms and concepts.
Acknowledgements
This research has been developed within the master thesis Project of “Ontology of Iranian Architectural History: based on contemporary
scholarly publications” in Shahid Beheshti University (2017) under the
supervision of Dr. Niloofar Razavi and consultation of Dr. Mehrdad Qayyoomi Bidhendi.
We would like to express our personal gratitude to all those who have helped in
carrying out the research.
Notes
[1] To avoid verbosity, from this
point forward, we will refer to “Iranian architectural history ontology” by
its shortened form “IArchHist ontology.”
[2] The first
academic program for training architectural historians was established in Iran in
2005 (A program called “Iranian architectural studies”
at the university of Shahid Beheshti, Tehran, Iran). For a detailed discussion on
the history of Iranian architectural historiography See reference 6.
[3] The point is clarified with examples in
section 4 of the paper.
[4] The following works were used: [Afshar 1995]
[Ayatollah Zadeh Shirazi 1980]
[Mostafavi 1982]
[Prinia 1992]
[Pirnia 1994]
[Pirnia 2011]
[Pirnia 2004]
[Sotoudeh 1992]
[5] As mentioned before, due to the lack and
inadequacy of structured resources such as metadata standards, thesauruses,
and databases in the domain of Iranian architectural history, an inventory
of written references was chosen under consideration of a group of domain
experts to extract the data necessary for building IArchHist ontology.
[6] As
demonstrated before, English equivelants and definitions of the Persian
architectural terms do not always encompass all their meaning. Therefore, in
this section, discussed terms are presented by transliteratation. However,
loosely translated English equivelants of terms are provided to make them
understandable for anglophone readers. Note that the original version of the
IArchHist ontology is entirely developed in Persian.
[7] Importing exact terms used in
sources written by prominent scholars would simultaneously portray the
vocabulary of research in Iranian architectural history.
[8] To better understand the process, in the following examples,
texts are translated to English, however IArchHist ontology captures Persian
terms.
[9] Middle-Out
approaches identify central concepts in each area/domain identified; core
concepts are identified and then generalized and specialized to complete the
ontology [Gandon 2002].
[10] Figures 3 to 7 are translated screenshots of IArchHist
ontology.
Works Cited
Acierno et al. 2017 Acierno et al. “Architectural heritage knowledge modelling: An ontology-based
framework for conservation process.” In Journal of
Cultural Heritage, Volume 24 (March–April 2017), pp 124-133.
Afshar 1995 Afshar (Afshār), Iraj. Yādgārhā-yi Yazd (Monuments of Yazd). Tehran: society for
the appreciation of cultural works and dignitaries, 1995 (in Persian).
Ayatollah Zadeh Shirazi 1980 Ayatollah Zadeh
Shirazi (Āyatollahzāda Shīrāzī), Baqer. “Masjid-i jāmiʿ-i
Ardistān (Friday mosque of Ardistān).” In Athar, volume 1, issue 1, 1980 (in Persian).
Blayney 2017 Blaney, Jonathan. “Introduction to the Principles of Linked Open Data.” In The Programming Historian. Vol 6 (2017), https://doi.org/10.46430/phen0068.
Brewster et al. 2004 Brewster, Christopher et al.
“Knowledge Representation with Ontologies: The Present and
Future.” In IEEE Intelligent Systems, Volume
19 Issue 1 (January 2004), pp. 72-81.
Gandon 2002 Gandon, F. Distributed artificial intelligence and knowledge management: Ontologies and
multiagent systems for a corporate semantic web. Scientific Philosopher
Doctorate Thesis in Informatics Defended Thursday the 7th of November 2002, INRIA and
University of Nice -Sophia Antipolis, Doctoral School of Sciences and Technologies of
Information and Communication, (2002).
Harpring 2010 Harpring, Patricia. Introduction to Controlled Vocabularies: terminology for art,
architecture, and other cultural works. Los Angeles: Getty Research
Institute, 2010.
Horridge 2011 Horridge, Matthew. “A Practical Guide to Building OWL Ontologies: Using Protégé 4 and
CO-ODE Tools.” Edition 1.3. The University of Manchester: 2011. Available
online: http://owl.cs.manchester.ac.uk/tutorials/protegeowltutorial/resources/ProtegeOWLTutorialP4_v1_3.pdf
Mojtahedzadeh 2018 Mojtahedzadeh
(Mujtahidzāda), Rouhollah. “Tārīkh-i tārīkh-i meʿmārī-yi Islāmī
(The History of Iranian Architectural Historiography: An Overview)” in
Art in Islamic Civilization: Architecture 1, Mehrdad
Qayyoomi Bidhendi (Qayyūmī Bīdhendī), (ed.), Tehran: SAMT, 2018 (in Persian).
Mostafavi 1982 Mostafavi (Mustafavī), Mohammad
Taqi. Āthār-i tārīkh-i-yi Tehrān (Historical Monuments
of Tehran: Tehran Society of the National Heritage of Iran, 1982 (in Persian).
Noy and McGuinness 2001 Noy, Natalya and Deborah
McGuinness. “Ontology Development 101: A Guide to Creating Your
First Ontology.” Stanford Knowledge Systems Laboratory Technical Report
KSL-01-05 and Stanford Medical Informatics Technical Report SMI-2001-0880 (March
2001).
Noy and Rector 2006 Noy, Natalya & Alan Rector.
“Defining N-ary Relations on the Semantic Web.” 2, 12
(2006). Available online: https://www.w3.org/TR/swbp-n-aryRelations/
Pirnia 1994 Pirnia (Pīrniyā), Mohammad Karim. “Çafd-hā va Ṭāq-hā (Arches and Vaults).” In Athar, volume 15, issue 24, 1994 (in Persian).
Pirnia 2004 Pirnia (Pīrniyā), Mohammad Karim. Sabk-shenāsī-yi meʿmārī-yi Īrānī (The Stylistics of Iranian
Architecture), Surūsh-i Dānesh , 2004 (in Persian).
Pirnia 2011 Pirnia (Pīrniyā), Mohammad Karim. Meʿmārī-yi Īrānī (Iranian Architecture), Surūsh-i Dānesh,
2011 (in Persian).
Prinia 1992 Pirnia (Pīrniyā), Mohammad Karim. “Gunbad dar meʿmārī-yi Īrān (Dome in Iranian architecture).”
In Athar, volume 12, issue 20, 1992 (in Persian).
Qayyoomi Bidhendi 2009 Qayyoomi Bidhendi, Mehrdad.
“Sukhanī dar manābiʿ-i maktūb-i tārīkh-i meʿmārī-yi Īrān va
shīva-yi justujū dar ānhā (On the Sources of the History of Iranian Architecture
and Its Research Principles),” in Gulistān-i
Hunar, no. 15, 2009 (in Persian).
Sennett 2008 Sennett, Richard. The Craftsman. New Haven: Yale University Press,
2008.
Silva et al. 2012 Silva, Isabel et al. “An integrated approach for evaluating the visualization of
intensional and extensional levels of ontologies.” In Proceedings of the 2012 BELIV Workshop: Beyond Time and Errors-Novel Evaluation
Methods for Visualization, (2012).
Sotoudeh 1992 Sotoudeh (Sutūda), Manouchehr. Jughrāfiyā-yi tārīkh-i-yi Shemīrān (Historical Geography of
Shemiran), Tehran: Institute for Humanities and Cultural Studies, 1992 (in
Persian).
Stanford 1998 Stanford, Michael. An Introduction to the Philosophy of History. Malden,
Massachusetts: Blackwell publishers, 1998.
Uschold and Grüninger 1996 Uschold, Mike, and Michael
Grüninger. “Ontologies: principles, methods, and
applications.” In Knowledge Engineering
Review, Vol 11, No 2. (Februrary 1996), p.93–155.
URL: http://www.digitalhumanities.org/dhq/vol/15/3/000559/000559.html
Comments: dhqinfo@digitalhumanities.org
Published by: The Alliance of Digital Humanities Organizations and The Association for Computers and the Humanities
Affiliated with: Digital Scholarship in the Humanities
DHQ has been made possible in part by the National Endowment for the Humanities.
Copyright © 2005 -
Unless otherwise noted, the DHQ web site and all DHQ published content are published under a Creative Commons Attribution-NoDerivatives 4.0 International License. Individual articles may carry a more permissive license, as described in the footer for the individual article, and in the article’s metadata.
Comments: dhqinfo@digitalhumanities.org
Published by: The Alliance of Digital Humanities Organizations and The Association for Computers and the Humanities
Affiliated with: Digital Scholarship in the Humanities
DHQ has been made possible in part by the National Endowment for the Humanities.
Copyright © 2005 -
Unless otherwise noted, the DHQ web site and all DHQ published content are published under a Creative Commons Attribution-NoDerivatives 4.0 International License. Individual articles may carry a more permissive license, as described in the footer for the individual article, and in the article’s metadata.
