(Columbia University, New York, USA)

Keywords: artificial intelligence, knowledge representation, document analysis

This paper presents an innovative approach to the analysis and reconstitution of medieval architecture. Computing technologies from Artificial Intelligence used in conjunction with 3D modeling software provide new means to visualize lost or never built architecture. Natural Language Processing makes it possible to translate descriptions of architecture in a format that can be processed by a computer; Knowledge Representation enables a computer to reason about a described structure. These methods from computer science offer new possibilities for the analysis of architecture. For example, through studying how knowledge is embedded in documents describing architecture these technologies can serve to reconstitute and reason about the verbally depicted structures. Biblical glosses in which medieval patrons envisioned ideal – though never built – architecture can be analyzed with the assistance of a computer, providing us with new means to enter into the architectural mindset of medieval people. These technologies also can further our current knowledge on actual architecture. Archeological descriptions can be computationally analyzed to check their accuracy with regard to extant structures. The method has the potential of determining the modifications that modern restorations introduced into the original medieval architecture. Finally, computers could provide assistance when analyzing written accounts to recreate lost buildings. Focusing on Gothic cathedral architecture and on Biblical archetypes, this paper will introduce these new technologies and present a computational method that utilizes Knowledge Representation for the study of medieval architecture. The paper will present the challenges that a computer system faces when analyzing descriptions of Gothic architecture, show how formal logic provides answers to these challenges, and introduce the core of an actual working implementation of a Knowledge Representation for Gothic cathedrals.

Relevance conference / Relevance session:
This paper demonstrates innovative technological research for the study of cultural heritage in general.

The use of Artificial Intelligence to create a computational model of historic architecture is new.