Tiffany Yu
Linguistic Landscapes
Language, Perception, Space
This thesis explores the intricate interplay between Language, Perception, and Space, proposing that language serves not merely as a descriptive tool but as a formative force that shapes our understanding and interaction with architectural and urban landscapes. Drawing on the works of influential architects and theorists like Oswald Mathias Ungers’ City Metaphors and Bernard Tschumi’s Manhattan Transcripts, which explore the poetic and narrative aspects of architecture, my research delves into the transformative potential of Natural Language Processing (NLP) and Artificial Intelligence (AI) as catalysts in a unique design workflow. This workflow aims to convert the abstract language of design prompts and interdisciplinary conversations into meaningful 2D curves and 3D forms.
The project employs machine learning based image generators to convert abstract linguistic conversations into complex 2D curves. These curves are then simplified and visualized using Fourier Transform. As a drawing tool, it reveals the underlying mathematical harmonies of the shapes inspired by the complexities. This visualization acts as a bridge, making the leap from abstract linguistic concepts to concrete spatial forms more transparent and dynamic. It enriches our understanding of how subtle shifts in language can lead to a kaleidoscope of architectural possibilities, thereby making the design process versatile and vibrant.
In this vision of architecture, form is not just a product but a process—a journey from narrative to translation, and finally to observation. There is beauty in how things can be described in a nondeterministic way, challenging conventional architectural paradigms. The methodology transcends traditional architectural form generation by interpreting, reimagining, and materializing conversations and perspectives into tangible forms. It thereby expands the architectural vocabulary to include not just physical structures but also the intricate layers of meaning, emotion, and perception that shape them.
By establishing a dialogue between architecture, artificial language processing, and spatial cognition, this thesis reinforces the idea that architecture is a deeply interdisciplinary field that benefits from drawing on a wide range of knowledge and perspectives. This research serves as an experimental platform to offer a new approach to architectural design—one that balances the subjectivity of human creativity with the objectivity and scalability of AI algorithms.