Points of Articulation
Seth Scott
2025-26
Points of Articulation is an artistic research project exploring the politics and poetics of synthesised speech. It takes as a starting point the deep history of speech synthesis, which has roots in classical antiquity and encompasses West-African percussion music, Medieval literature and Renaissance automata, as well as modern day innovations in robotics and AI. The history of speaking machines and our changing attitudes towards them illuminate a longstanding fascination with animating the inanimate; through creative engagement with these tools I hope to explore complex and timely questions about technologies of knowledge production and the human body.
We live in a cacophony of artificial voices; a chorus of ‘smart’ assistants, customer service avatars, navigation systems, advertising voiceovers and deepfakes which become more difficult by the day to distinguish from the real thing. Whilst the ubiquity of these digital, incorporeal voices is reasonably new, fantasies of artificial bodies are not; they featured in Hellenistic scientific writing and mythology, in literature and historiography in the Middle Ages, and they began to take shape in Renaissance Europe in the form of sophisticated and imaginative automata. Tales of celebrated scholars combining arcane knowledge, technical innovation and diabolical magic to create clairvoyant talking heads illuminate complex relationships between nature and artifice, mechanics and magic, and between different types of intelligence and knowledge in pre-modern thought. They also resonate, I think, in quite an uncanny way with current discourse around technology, the body, creativity and labour in the wake of recent developments in machine learning and generative AI. In Points of Articulation I weave together historical and contemporary narratives, using sound and music to explore the politics and poetics of synthesised speech.
For the main artistic output of this project, I will produce an experimental radio piece incorporating elements of historical fiction, algorithmic writing, sound poetry and electroacoustic music. The spoken component will consist of seven short vignettes, each exploring a different historical paradigm of speech synthesis (eg. Middle-ages, Renaissance, 20th Century etc…) Around this historical narrative, I will compose an electroacoustic score made from synthetic vocal sounds, and a more abstract, poetic text exploring the themes described above. Below are some compositional excerpts made using AI-powered speech synthesizers:
Below are excerpts from Lina-Speech Studies, a series of 40 short pieces composed using AI-powered speech synthesis tools during an Artistic Research Residency at IRCAM in Spring 2025. Whilst this material will not feature directly in my piece for radio, I will employ similar creative processes in both the writing and musical composition. The texts were derived from the CMU_ARCTIC database, a phonetically balanced database designed specifically for unit selection speech synthesis research. I transformed the text contained within this database in a variety of ways, to create the pieces you hear below. All sounds were generated using the IRCAM software Lina-Speech.