The paper is based on 82 open-ended interviews conducted by as many students during 2006–2013.The respondents were presented with pictures of two artworks, The Persistence of Memory (1931) bySalvador Dalí and Which Link Fails First? (1992) by Teemu Mäki, a Finnish contemporary artist.They were asked to comment and compare the two pictures and tell which one they liked better. Therespondents’ spontaneous comments show different aspects of how an artwork is perceived and evaluated.The interviews were analysed both qualitatively and quantitatively. As the result of in vivocoding, 40 variables were created for use in a content analysis. The respondents focused on differentthings when evaluating the two artworks. When commenting Dalí’s painting, they paid attention onits affective and sensory characteristics, while Mäki’s work was discussed primarily in terms of itsmessage and perceived lack of professional quality. In parallel, a selection of interviews was analysedin order to reveal the temporal sequence of discussing and evaluating different aspects of the paintings.The analysis showed three ways of discussing, which were called naïve, scholarly, and deliberative.The temporally structured model of aesthetic appreciation and judgement suggested in 2004 byLeder and his co-workers was used as a heuristic device for an analysis of the shifts of attention thattake place when a discourse is created and anchored in perception. Both cognitive psychology andphenomenological sociology emphasize the dependence of perception on context and intention; thereis reason to take that theoretical starting point seriously.