This dissertation is concerned with the acquisition of evidential markers by Turkish-speaking children and their non-linguistic source monitoring abilities.

Evidential markers encode the speaker's source for the information being reported in the utterance. While some languages like English express evidentiality by lexical markers (I saw that it was raining vs. I heard that it was raining) about one sixth of world's languages express source information through grammaticalized evidential markers.

In this dissertation, focusing on evidential markers in Turkish, we would like to examine systematically how children advance in the process of acquisition of evidentiality. Specifically, we would like to answer the following questions: How do children associate source information with evidential markers? When do they start producing them in a meaningful way? What is the relationship between the children's non-linguistic source reasoning development and their acquisition of evidential markers? Do the non-linguistic source reasoning abilities precede the linguistic ones or vice versa or do they go hand in hand? We report findings from three studies conducted with Turkish learners between the ages of 5 and 7 that test the acquisition of both the semantics and pragmatics of linguistic evidentiality (Exp.1-3) and three studies that reflect the same children's source monitoring abilities (non-linguistic) associated with evidential markers (Exp.4-6). The first three studies systematically target the full range of evidential meanings (direct vs. indirect: hearsay/ inference) encoded in the Turkish past tense system and attempt to chart their developmental timetable. The last three experiments are designed to be closely parallel to the linguistic experiments and they investigate children's non-linguistic source monitoring abilities. The goal of this investigation is to detect children's early difficulties with evidentiality and -if there are any- the nature and scope of these difficulties.

We conducted a general comparison between all three linguistic experiments and all three non-linguistic experiments. Our findings revealed that children performed better on the non-linguistic experiments than they did on the linguistic ones.

Moreover, one-by-one comparisons between each linguistic experiment and its non-linguistic counterpart revealed that the results of each of the non-linguistic experiment were better than the results of the relevant linguistic experiment. Our results show that the non-linguistic source concepts are developed much before the acquisition of linguistic evidentiality. Therefore, our data does not provide support for the "Conceptual Change Hypothesis", that predicted that conceptual development is closely tied to linguistic development and therefore, children's non-linguistic performance cannot be better than their linguistic performance (unlike what we have found to be the case).

Moreover, we have observed that children's performance at the Direct Evidence tasks was better than their performance at the Indirect Evidence tasks. This finding was persistent not only in language but also in cognition. Hence, we conclude that children's conceptual development for direct evidence precedes the one for Indirect Evidence. Children -at least at the beginning of the acquisition process- may be going through a period in which they are developing the relevant concepts and this conceptual development may not happen simultaneously for all the members of one category (In this case the category is source of information). (Abstract shortened by UMI.)