Fundamental Science Inspired Metaphorical Approach to Creative Computing for Persuasion: An Ethos-Pathos-Logos Framework

This thesis proposes a metaphor-driven creative computing framework for the structured generation of persuasive messages. Fundamental-science metaphors are treated as computational operators that define states, constraints, and composition rules, yielding varied yet interpretable sentence surfaces. The framework comprises three coordinated stages. Ethos is aligned with chemistry to compose credential statements and resource offers via rule mappings over chemical affinity matching, neutralisation reaction, and endothermic reaction. Pathos is aligned with physics to manage affect through active noise cancellation that removes unwanted negative affect, acoustic resonance that aligns message intensity with the listener’s arousal band for efficient uptake, and amplification that raises the level of the same in-band signal without changing its spectrum or phase. Logos is aligned with a mathematics metaphor to structure evidence through Pareto’s rule that summarises the past as decided by vital decisions, golden ratio that validates the present by representing the option as an balance between internal and external factors of listener, and Fibonacci sequence that projects the future, with each step reusing the previous output and jointly covering numerical, qualitative, and logical evidence. A proof-of-concept system for UK university subject selection is presented to demonstrate the approach. The implemented pipeline is evaluated using topic and length matched prompts. Assessment is conducted with audit driven operator metrics. Results indicate full metaphor coverage and correct placement across messages. Science-aligned metaphors are established as both the creativity mechanism and the methodological framework for creativity generation in persuasive message. Then, a structured ethos–pathos–logos pipeline is implemented that separates and coordinates credibility cues, affect control, and evidential framing. Lastly, chemistry, physics, and mathematics are unified into reusable, interpretable rule libraries, instantiating creativity generation as an operator space that supports enumerability and recombination.