Religo: A Relationship System

Religo: A Relationship System
Rodríguez, Nuria, University of Málaga (Spain), nro@uma.es
Isolani, Alida, Scuola Normale Superiore (Italy), isolani@signum.sns.it
Lombardini, Dianella, Scuola Normale Superiore (Italy), dianella@signum.sns.it
Marotta, Daniele, Scuola Normale Superiore (Italy), marotta@signum.sns.it
Introduction
Over the years, digital libraries and textual archives have collected, described, and classified texts and multimedia objects. These kinds of repositories are effective in compiling, describing, and disseminating the cultural heritage such as the artistic and literary expressions. Also, many of them, following the developments of Computational Linguistics, have incorporated tools for textual analysis as part of their end-user services. Nevertheless, these systems are weak in terms of relationships. Of course, they are configured in such way that it is possible to relate the digital objects compiled; thus, for instance, it is easy to retrieve a set of visual artefacts sharing the same subject matter. However, these relationships are based on the traditional criteria of classification and description (metadata and keywords), without any intention of exploring the nature or specific characteristics of the relationships that the Art History discipline’s phenomena maintain among them.

We should not overlook the intrinsic relationship that exists among texts, concepts – or ideas –, words and visual artefacts in the construction of art-historical knowledge (Mitchell, 1994). As Heffernan (2006) argues for the case of words and images, this relationship should not taken as a simple reproduction of art works by a set of words, but rather the conjunction of visual artefacts and words generates new knowledge. This is one of the reasons for which we can contend that these relationships deserve to be treated as a research object in themselves.

Therefore, our intention is to explore the potential relationships that could be established among these entities: texts, concepts, words, and visual artefacts with the aim of investigating how these relationships are able to produce new significant knowledge or are able to open new understandings.

Religo
Religo is a system that enables the construction of interpretations based on relationships. According to the researcher’s needs and the art historian’s research habits and procedures, the question of text and image induces us to design a system to establish relations between various kinds of objects (texts, images, videos, etc.), to provide new possibilities for analysis and research, which are offered only partly by the state of the art.

Currently, in the most significant projectsImage Markup Tool (link) . Pinakes (link) . VLMA (link) . EPPT (link) . Flickr (link) . TextGrid (link) . Talia (link) . working on texts and images – although with interesting and useful features – some limitations can be found in terms of:

functionalities: only tagging or object manipulation;
usability: complex and without a user-friendly interface;
purpose: oriented to social participation (this feature often does not reach the entire scholarly community).
Taking into account this state of the art, Religo does not reduce the relationship to the concept of pure tagging or to the idea of simple connection between entities, but treats it as basic element for interpretation and analysis, making itself the subject of research in order to create new knowledge.

Religo relates the domain entities creating two logical levels: the expression one, consisting of digital objects on which the interpretation can be developed; and the semantic one, consisting of digital concepts (the relationships between digital objects and the predications on themselves) that allow the interpretation to be built (Buzzetti, 2004)This theoretical model has been used by Signum to develop a system for facilitating semantic research and text reading in Text and Semantics (link) ..

This means that, when a digital concept is the subject of interpretation, it is placed on the expression level, becoming a digital object itself.

An example is the Michelangelo’s masterpiece The Last Judgement, analysed and interpreted by the Spanish author Francisco Pacheco in his 17th treatise The Art of Painting (1649). In this case, The Last Judgement would be the digital object, and the interpretation given by Pacheco, the digital concept. However, insofar as the Pacheco’s interpretation is also subject of interpretation and analysis by the modern historiography, it in turn becomes a digital object.

The entities of the domain can be submitted to a number of general operations (such as selection of parts, links, free tagging or metadata encoding, etc.) and others more specific according to their particular features (specific operations on texts and image).

In every operation the centrality of the relationship is clear: from its creation, made easily by a simple drag and drop of selected portions of objects, up to reach the composition of documents as result of the different entities relationship that themselves constitute the new knowledge of the study process.

A more evident utility of this new use of relationships can be seen in the search and navigation functions, for example to improve the search capability because it ensures a higher degree of precision and recallPrecision can be seen as a measure of exactness or fidelity, whereas recall is a measure of completeness..

As an example, searching for the word emblema, Religo returns both entities containing emblema as textual occurrence (W) and those where emblema does not appear (NW) but which are related to W (Figure 1). Thus, the relationship gives relevance and importance to the entity NW, which otherwise, from a purely textual standpoint, would go unnoticed or simply would not exist.

Figure 1C. Ripa, Iconologia, overo Descrittione dell’Imagini universali […], Roma 1593, p. 96. G. Vasari, Ragionamenti di Giorgio Vasari pittore ed architetto aretino […], Firenze 1832-38, p. 1404.

Full Size Image

Relationship also affects the display of the entities: that is, in addition to the classical view as list, Religo provides a view as graph, creating a network between the various domain entities which on one hand allows the reconstruction of interpretive reading by simply moving the focus between digital objects, on the other hand enables a contextualized vision of every digital object. These features, typically used during the work process, are also useful in order to share and exploit the research results.

As a more concrete example, let us consider the following domain:

The Art of Painting by Pacheco discusses Michelangelo and his works, including The Last Judgement. It also treats other painters such as Velázquez or Tiziano, other works such as The Final Judgement painted by Pacheco himself, the portrait of King Phillip II painted byTiziano, and other artists such as Dolce, Paleotti, Lomazzo, or Céspedes.

Religo allows relationships to be created between these objects:

[Pacheco] author of [The Art of Painting]

[Michelangelo] painter of [The Last Judgment]

[The Last Judgement by Michelangelo], [The Last Judgement by Pacheco] described in [The Art of Painting]

[Phillip II by Tiziano] mentioned in [The Art of Painting]Notice the difference between describe, which implies a detailed explanation of the painting; and mention, which only means that the painting has been cited.

[The Last Judgement of Michelangelo] influences on [The Last Judgement of Pacheco]

[The Last Judgement of Michelangelo] used as example by [Dolce, Pacheco, Lomazzo, Céspedes]

[Dolce, Paleotti, Lomazzo, Céspedes] cited in [The Art of Painting]

[The Last Judgement of Michelangelo] illustrates concepts of [deviations of decoro, terribilità, movements and affetti]

[The Last Judgement of Michelangelo] described with terms [artificioso, espantoso, terrible, horribilidad, feroz]

These relationships themselves become new entities of the domain, forming an interconnected network and producing the following new knowledge level:

[Lomazzo] [Céspedes] cited by [Pacheco] to define the concept of [painting]

At this level, we can see how the result specifies the connection among the different theorists cited by Pacheco and the concepts that he defines in The Art of Painting until creating another richer level:

[The Last Judgement] used as example by [Pacheco, Dolce] to illustrate idea of [deviations of decoro]

[The Last Judgement of Michelangelo] used as example by [Pacheco, Lomazzo] to illustrate the idea of [movements and affetti]

[The Last Judgement of Michelangelo] used as example by [Pacheco, Céspedes] to illustrate the idea of [terribilità]

What we can deduce from this result is that, from the second half of 16th century, The Last Judgement by Michelangelo plays the role of universal reference to illustrate or exemplify a wide range of aspects concerning the visual arts, being used by each author in a different way. The Art of Painting, as an encyclopaedic treatise, brings together many of these interpretations, which Pacheco unifies into a single point of view.

The most interesting results arise when we use a complete repository of works, images, and texts. For example, if we consider a repository of Spanish 17th treatises, as ATENEA Project (link) , we might find the following types of relationships:

[The Last Judgement of Michelangelo] described by [Pacheco, 1649] and [Carducho, 1634] only mentioned by [Martínez, ca. 1675]

[The Last Judgement of Michelangelo] described by [Pacheco] and [Carducho] with coinciding terms [confusión, temor, horribilidad, terrible].

Conclusions and Future Developments
As an initial task, Religo is provided with all the typical features to operate on texts and images in terms of combination of interacting tools for example to describe and catalogue visual artefacts, to analyse images, to manipulate images, or to annotate images (whole or partly).

Together with the standalone version, an online should be allowed in order to ensure content sharing and social tagging in expert contexts of usage.

Moreover, the system would be generalized for use in other different domains and would have the capacity to handle other types of entities such as audio and video.

References:
Buzzetti, D. 2004 “Diacritical Ambiguity and Markup, ” Augmenting Comprehension: Digital Tools and the History of Ideas, D. Buzzetti, G. Pancaldi, and H. Short (eds.) Office for Humanities Communication London-Oxford

Heffernan, J. 2006 Cultivating picturacy: visual art and verbal interventions, Baylor University Press Waco, TX

Mitchell, W.J.T. 1994 Picture Theory: Essays on Verbal and Visual Representation, Chicago University Press Chicago