To acquire the meanings of attitude verbs like know and think, a child must marshall many linguistic and extralinguistic abilities, including an understanding of (i) the relevant concepts, (ii) the relationship between a verb's meaning and the syntactic frames it appears in, (iii) how verbs are used in context, and (iv) speakers' intentions. Past research on the acquisition of think and know, in particular, has suggested that children do not have an adult-like understanding of these verbs before the age of 4, and this might suggest that younger children are unable to synthesize all the relevant sources of data on their meanings. In this talk, I will present findings from a new experiment which indicate that children can acquire adult-like meanings for these verbs by 3 (which is an age that they are typically failing classic false belief tasks and demonstrating non-adult pragmatic abilities), but that there is much individual variation in mastery of know specifically. In order to further investigate the source of this variation, I will propose a rotation study that examines the effect, on comprehension of know and think, of differences in syntactic and pragmatic input for the verbs. The proposed study will bring together findings from a corpus study of differences in child-directed speech to children of low and high socioeconomic status (SES) backgrounds, and a behavioral study on how children of low and high SES backgrounds might differentially understand the verbs know and think.