One of the fundamental questions in the field of language acquisition is a learnability problem, which considers how learners acquire certain aspects of language which are not directly provided in the input or whose referents are not readily observable. This dissertation investigates Japanese children's acquisition of various linguistic phenomena, each posing a learnability challenge due to either the absence of direct evidence or the unobservability of the referent in the world. The first focus is on the acquisition of verbs referring to mental states and those encoding a speaker's empathy. Acquiring the meanings of verbs is inherently complex, especially when its referent is not obvious. Verbs that refer to mental states (e.g., "think" in "John thinks that it is raining outside") present a particular challenge as their referents, being internal mental states, are not visible, thus questioning how children acquire their meanings. Similarly, empathy verbs in Japanese also present a potential learnability challenge. Empathy verbs, such as verbs of giving and receiving in Japanese ("ageru" 'give,' "kureru" 'give,' and "morau" 'receive'), encode speaker's empathy to specific discourse referents. These verbs can be used as simple transfer verbs (e.g., "John-NOM Mary-DAT flower-ACC give-PST" 'John gave the flower to Mary'), and also as benefactive auxiliary verbs to add a benefactive meaning to another verb (e.g., "John-NOM Mary-ACC praise give-PST" 'John praised Mary (for the sake of Mary)'). Importantly, which empathy verb to use depends on which discourse referent the speaker empathizes with (feels close to) the most: When the speaker empathizes with a giver/benefactor subject argument, the speaker has to use ageru 'give,' when the speaker empathizes with a recipient/beneficiary non-subject argument, the speaker has to use kureru 'give,' and when the speaker empathizes with a recipient/beneficiary subject argument, the speaker has to use morau 'receive.' Empathy verbs present a particular challenge for learners because which discourse referent the speaker empathizes with is determined in the speaker's mind, which is unobservable from outside. Moreover, "ageru" 'give' and "kureru" 'give' have the same subcategorization frame (e.g., giver/benefactor as the subject). Therefore, the difference between "ageru" 'give' and "kureru" 'give' is not overtly marked in the sentences. The second focus is on the fact that both mental verbs and empathy verbs license long-distance interpretations of the Japanese reflexive zibun 'self.' Previous acquisition study reported that Japanese children are hardly exposed to zibun 'self' referring to a long-distance antecedent. This poses a serious learnability challenge for children: How do children (come to) know that "zibun" can refer to long-distance antecedents even if there is almost no evidence in the input? Moreover, it has been reported that children generally have difficulty in accessing long-distance interpretations with a reflexive, including zibun 'self,' and the cause of this difficulty is not clear. The first investigation of this dissertation (Study 1) is corpus analyses examining how frequent empathy verbs are in child-directed speech and what kinds of cues for speaker's empathy children receive. I show that Japanese children are exposed to linguistic cues and non-linguistic cues for speaker's empathy with the empathy verbs fairly robustly. Subsequently, Study 2 shows that Japanese children come to master the empathy-encoding properties of the empathy verbs by around age 6yrs, but some children have already mastered them as young as age 4yrs or possibly younger. Moreover, based on the findings in the corpus analyses, Study 3 experimentally investigates whether children can really use the linguistic and non-linguistic cues for speaker's empathy to learn novel empathy verbs, and the experiment found that the linguistic cue is far more helpful than the non-linguistic cue in learning the novel empathy verbs. This provides evidence in favor of the Syntactic Bootstrapping Hypothesis (e.g., Gleitman, 1990; Landau & Gleitman, 1985) which states that children are predisposed to use syntactic information to solve learnability problems involved in the acquisition of mental verbs. I further investigate whether children extend their knowledge of the empathy verbs to other empathy-related phenomena, which are not provided in the input. My third experiment (Study 4) examines how early Japanese children (come to) restrict an antecedent of "zibun" 'self' to the long-distance subject when it is required by an empathy verb. The result shows that 6- year-old children are almost at ceiling, but some younger children are able to restrict "zibun's" antecedent to the long-distance subject with the empathy verb, even if there is almost no direct evidence in the child-directed speech. The final experiment (Study 5) examines children's acquisition of long-distance interpretations of "zibun" 'self' licensed by one of the mental verbs, the belief verb "omou" 'think.' The results show that children's interpretation of the belief verb, "omou" 'think,' significantly affects their interpretation of "zibun" 'self.' This result supports my hypothesis that one of the causes of children's difficulty with long-distance interpretations of zibun is their non-adult-like interpretation of the belief verb, "omou" 'think. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]