Divergent and convergent thinking tasks provide insight on how humans are operating in open and closed problem spaces. Both have been connected to specific modulations in the EEG alpha band (10-12 Hz). However, these tasks often operate in different knowledge domains (e.g., verbal, spatial-figural) and are based on entirely different task designs (e.g., drawing, verbal, written). At the same time, prior research showed that both divergent and convergent thinking heavily rely on working memory (WM) capacity of the subject. Therefore, directly comparing both processes might result in an overestimation of their shared variance due to the impact of WM. We used three tasks for divergent and convergent thinking as well as a WM task that target verbal processing by using highly comparable stimulus material. We investigated divergent and convergent thinking at early, intermediate and late stages of each self-paced trial to account for the temporal variability due to the different cognitive stages in the problem-solving process. As a first step, we used theta band oscillatory activity as an indicator of WM-related activity. In a second step, we removed this oscillatory activity from the data collected during divergent and convergent thinking. Our results show a strong upper alpha synchronization during the divergent in contrast to the convergent thinking task, especially for fronto-parietal regions. Our study provides a first attempt to investigate mechanisms of divergent and convergent thinking beyond WM for the verbal knowledge domain.