Architectural settings create interest by their ability to arouse a variety of perceptions and imaginative readings in appropriately engaged visitors. Underlying this observation is a belief that particular vantage points, and conditions of illumination are at least partially responsible for varying percepts and readings of the same space, but this is not an easy phenomenon to test. The main issue is to capture what a visitor actually perceives in a space in ways that are comparable across visitors or that can be quantified. Recording and tracking of saccades is one possible means to do this, but is not very satisfactory since it only captures the points at which foveal vision is directed and attention can move independently of saccadic movements and with different radii of focus. In a study below we test a procedure that tries to capture differences in perceived configuration of a given space by using verbal reports to identify the object of attention. This approach depends on some key findings in studies of perception - that attention is necessary for conscious perception, and that verbalizing can identify objects of attention without preventing the subjects to switch from spatial to object attention or vice-versa. We tested this approach in an experiment that used selected views of Tadao Andoâ’s Pulitzer Foundation for the Arts (St Louis, Missouri) as a case. The views were grouped into 4 distinct exploration routes of the same setting, each route described by a set of photographs. Each subject was assigned to one of the 4 routes. The subjects were asked to visualize the space presented to them in the photographs and to report in a prescribed format of single sentences a narrative of their progress through the visualized space. The aim was to test if the differences in the routes correlated with differences in the description of the mentally traversed space. Comparing the sentences, we have found systematic and statistically significant differences between the routes in the number and types of prepositional phrases that subjects used, and in the order in which they selected, and therefore, attended to, the different features of the scene presented to them. The analysis of experimental results supports the hypothesis that different exploration routes influence and alter the observer‘s attention that results in a distinct complexity and nature of the observer’s understanding of the setting. In the proposed paper, we summarize the results of the experiement conducted in two different settings, one in Atlanta (in English) and one in Budapest (in Hungarian), and offer some reflections on certain methodological challenges that were raised in the course of the work. Comparing the Hungarian and English experiments, we discuss how using verbalization in specific formats provides a simple and useful way to capture the objects of subjects’ conscious attention and we offer some insights into the relationship between language and space.