論荷莉行動學習的 迴路共享模型

On Susan Hurley’s Shared Circuits Model of Action Learning

Hurley’s shared circuits model (SCM) is a cognitive architecture that specifies behavior-related sociocognition in terms of the dynamic interactions among perception, action, and the world. Despite its explanatory power for action learning, this model confronts a challenge: although Hurley introduces representation to the SCM to describe higher cognitive skills, she rejects classical representations with a domain-specific processor. The SCM has thus been questioned in terms of its ability to accommodate representation and computation. In this paper, I present a solution that integrates a motor selection mechanism into the SCM. I show how this integration, which requires neither an additional specialized processor nor representation in any classical sense, explains action learning and provides a basis for even higher sociocognitive skills in terms of general computational processes.

英國哲學家蘇姍‧荷莉所提出的迴路共享模型，是一個透過感知、運動與外在環境三者的動態系統來解釋人類行動的社會認知架構。雖然這個模型在解釋工具行動之學習上有很多優點，但由於荷莉並不接受古典表徵與相應的處理器，卻又引進表徵概念來處理高等認知能力，因此該模型不但被批評與動態系統不相容，其處理計算和表徵的能力也備受質疑。本文提出一個解決方案，將荷莉的模型與一運動控制機制相整合。在此整合架構下，不但無需預設古典表徵與處理器就能進行資訊處理，此架構也替未來描述更高層的認知能力（例如涉及命題、概念的認知能力）提供有利基礎。