Systemic insight : the interplay between interactivity, incubation and transfer in insight problem solving

Henok, Niyat (2018) Systemic insight : the interplay between interactivity, incubation and transfer in insight problem solving. (PhD thesis), Kingston University, .


Classical perspectives on problem solving are embedded in computational models of insight problem solving, such as the information-processing model (e.g., Newell & Simon, 1972). Problem-solving activity is viewed as a product of information-processing in which people see or hear problem information, think about the solution, then produce the solution: see or hear, think, then act. More recently, Ohlsson (2011) suggested people solve problems by mentally restructuring the problem information. Hence, insight comes about as a consequence of restructuring (Weisberg, 2014). As such, the origin of insight is commonly understood as a mental experience. However, the traditional frameworks explaining the insight experience commonly overlook the influence of reasoners’ immediate environment. Systemic cognition frameworks such as the Extended Mind Thesis (Clark & Chalmers, 1998), Distributed Cognition (Hollan, Hutchins, & Kirsh, 2000) and the Systemic Thinking Model (Vallée-Tourangeau, Abadie, & Vallée-Tourangeau, 2015; Vallée-Tourangeau & Vallée-Tourangeau, 2017) assume information-processing is augmented when spread across mental and physical resources. When presented with a physical representation of a task, making changes to that physical representation, even arbitrary ones, may offer cues to new strategies, enabling better planning and efficiency in progressing towards a goal. Accordingly, the opportunity to interact and coordinate with the immediate environment enhances insight performance. This thesis sought to explore insight performance from a systemic cognition perspective. The research program investigated how the level of interactivity influenced solution rate in the Cheap Necklace Problem (de Bono, 1967; Silveira, 1971). Across four experiments, participants attempted to solve the problem either in a low interactivity condition, using only pen-and-paper and relying heavily on mental restructuring, or in a high interactivity condition, with a physical model of the problem with constituent elements they could manipulate while attempting to find a solution. The results across the experiments confirmed that increasing the level of interactivity resulted in enhanced insight performance. Incubation and transfer are often upheld as key determinants for insight performance. Thus, in addition to exploring the impact of interactivity, the experiments investigated how interactivity may interact with incubation and transfer to promote insight. To measure incubation effects, participants in the first two experiments reattempted the same problem after a two-week break. There was evidence of an incubation effect as performance substantially improved on the subsequent attempt. To explore transfer, a new Cheap Necklace Problem variant was introduced, which participants in the final two experiments attempted following the original version of the problem. Transfer was evident as participants were able to successfully transfer their solution to solve the new variant. Moreover, overall performance improved on the subsequent problem. Across the four experiments, the level of interactivity offered on the second problem attempt was important: When the problem presentation changed (low interactivity to high interactivity or high interactivity to low interactivity) performance only improved when working in a highly interactive task environment second. Thus, insight through interactivity fosters stronger performance on both the initial and subsequent task. This thesis further explored how interactivity prompts insight in a dynamic agent-environment by recording and analysing participants’ actions. One important finding from these behavioural analyses was the fact that those who spent the largest proportion of their time reconfiguring the task environment, thus making the most of the malleability of the artefacts available, were also most likely to reach insight.

Actions (Repository Editors)

Item Control Page Item Control Page