The spectatorship of portraits by naïve viewers (beholders) was explored in a single experiment. Twenty-five participants rated their liking for 142 portraits painted by Courbet (36 paintings), Fantin-Latour (36 paintings), and Manet (70 paintings) on a 4-point Likert scale. The portraits were classified in terms of focused versus ambiguous nature of sitter gaze and the presence of salient features in the context beyond sitters. Participants rated portraits while having their eye movements recorded. The portraits were split into regions of interest (ROIs) defined by faces, bodies, and context. Participants also completed individual difference measures of attention and task focus. Results showed naïve spectatorship to be subject to attentional capture by faces. Paradoxically, the presence of salient features in the context amplified the attentional capture by faces through increasing participants liking of portraits. Attentional capture by faces was also influenced by sitter gaze and task focus. Unsurprisingly, the spectatorship of portraits by naïve beholders is dominated by faces, but the extent of this dominance is influenced by exogenous and endogenous attentional factors.