Thesis Type: Postgraduate
Institution Of The Thesis: Orta Doğu Teknik Üniversitesi, Graduate School of Informatics, Cognitive Science, Turkey
Approval Date: 2011
Student: İSMET ADNAN ÖZTÜREL
Supervisor: HÜSEYİN CEM BOZŞAHİNAbstract:
Semiotic dynamics and computational evolutionary musicology literature investigate emergence and evolution of linguistic and musical conventions by using computational multi-agent complex adaptive system models. This thesis proposes a new computational evolutionary musicology model, by altering previous models of familiarity based musical interactions that try to capture evolution of songs as a co-evolutionary process through mate selection. The proposed modified familiarity game models a closed community of agents, where individuals of the society interact with each other just by using their musical expectations. With this novel methodology, it is found that constituent agents can form a musical agreement by agreeing on a shared bi-gram musical expectation scheme. This convergence is attained in a self-organizing fashion and throughout this process significant usage of n-gram melodic lines become observable. Furthermore, modified familiarity game dynamics are investigated and it is concluded that convergence trends are dependent on simulation parameters.