The acquisition of mental state terms in the third year of life is a developmental milestone in the transition from an implicit to an explicit Theory of Mind (Harris, 2015). In a longitudinal study (Sodian & Kristen-Antonow, 2016) children’s mental state vocabulary at the ages of 24 and 30 months was predictive of their understanding of sources of knowledge at 36 months, which, in turn, predicted false belief understanding at 50 months. To date, our assessment of mental state language is, however, solely based on parent-report data (Olineck & Poulin-Dubois, 2005; Kristen et al., 2012, 2014;  Chiarella et al., 2013) . In the present study, we developed an eye-tracking paradigm to assess children’s sensitivity for the distinction between “know” and “think” in 27- to 36-month-old children. Children saw animated cartoons in which two monkeys sequentially gave conflicting pointing messages about the location of a sticker, each followed by a voiceover which said  “The monkey knows (thinks) that the sticker is in this box”. Children were then prompted by asking “Where is the sticker?”. Looking  patterns at the images of the two monkeys and at  each of the two boxes were recorded at two seconds post stimulus in relation to a baseline period. Children’s discrimination between “know” and “think” trials was assumed to be correlated with their performance on a picture naming task (Bartz, Rowe, & Harris, 2015) which assessed verbal and non-verbal indications of uncertainty in response to familiar vs. unfamiliar pictures. Furthermore, we attempted to validate the eye-tracking task with a parent mental state language questionnaire. Findings will be interpreted with respect to the early development of children’s understanding of their own and others’ knowledge and ignorance, and with respect to the relation between language and conceptual development.