The aim of this dissertation was to develop a novel behavioural method to assess young children’s capacity for episodic foresight. Specifically, I developed a marble game paradigm to assess whether 3-, 4-, and 5-year-old children would save marbles for future enjoyment. Study 1 demonstrated that preschoolers will save more spontaneously when they can draw on a recent past experience compared to when the context is completely novel. Results from study 1 also suggest that preschoolers are sensitive to the relative value of future rewards. Study 2 revealed that providing children with a verbal prompt alerting them to their possible courses of action (i.e., saving or spending) facilitated saving, that 5-year-olds saved more than 3-year-olds, and that children’s performance on this behavioural (i.e., nonverbal) measure of saving was independent of their language proficiency. Children’s saving was not associated with other capacities hypothesized to relate to episodic foresight (theory of mind, inhibitory control, and working memory). Implications for children's saving, the development of episodic foresight, and future research are discussed.

目標解讀在許多不同的模仿理論中均受到重視，並被認為是模仿學習中的關鍵。其中，目標導向的模仿理論認為在模仿過程裡，觀察者會將動作要素以階層關係重組為主要目標和次要目標，在資源有限的情況下，次要目標在模仿中容易被忽略，只重現主要目標。本研究修訂過去研究嬰兒目標導向模仿之作業，探索示範中的動作方向及口語提示線索如何影響嬰兒的目標導向模仿。在實驗1中，嬰兒看到示範者以跳躍或滑行的動作方式將玩偶移動至桌面或盒子裡，結果發現嬰兒會在沒有盒子的情境中模仿不同的動作方式，並在有盒子的情境裡模仿示範者的位置選擇。實驗2修訂實驗1的程序，延後呈現運動方向提示目標的時間，結果發現移動路徑的改變會使嬰兒在有盒子情境裡模仿位置選擇的正確性下降。實驗3在示範開始前加入口語提示協助嬰兒區辨目標位置，並使用與實驗2相同的移動路徑，結果發現口語提示無法增加有盒子情境位置選擇的正確性，反而使無盒子情境中模仿動作方式的表現減少。嬰兒的模仿行為不能完全用目標導向模仿理論中目標的階層排序解釋，示範情境、溝通互動以及語言等不同線索皆可能影響嬰兒推理目標的方式，改變模仿的傾向。 / The goal-directed theory of imitation claims that infants imitate an action by decomposing it into separate and hierarchically organized goals. When resources are limited, infants ignore less important goals to reproduce main goals. The evidence of this theory is that infants prefer to imitate action outcomes over styles when an external goal was present. In contrast, infants take action styles as major goals when there was no observable outcome. In this research, we investigated how movement direction and verbal information influence the goal-directed imitation process. In Experiment 1, we replicated the goal-choice imitation task used in previous research. 18-month-old infants observed an adult moving a toy animal in different action styles (slide or hop) into one of the two boxes (box condition) or onto the table (no-box condition).The results showed that infants imitated the action styles in no-box condition and matched the location choice in box condition. In Experiment 2, we modify the task by delaying the timing of movement direction cue for goal choice. Infants imitated the box choices less accurately after observing the modified demonstration. In Experiment 3, we verbalized the actor’s goal to investigate whether the goal choice errors in Experiment 2 are due to the lack of goal salience. We found no increase for the accuracy of matching location in the box condition but a decrease for the imitation of action styles in the no-box condition. In addition to the tendency to imitate different goals in different conditions, the study suggests important roles of movement direction and verbal cues in infants’ goal-directed imitation.

Dog trainers sometimes teach verbal cues by saying the cue as the dog is performing the desired behavior. However, there is disagreement about when to say the cue. In this study, a pet dog was trained to go to three different apparatus, the cue for each of which was given at a different time, in a multi-element design. The cue "hoop" was given just as the dog began to move to the hoop apparatus. The cue "carrier" was given as the dog was stepping into the carrier apparatus. The cue "platform" was given after the dog was sitting on the platform apparatus. To test if the dog had learned the cues, the trainer had the dog sit and gave the cue. During testing, if only the correct apparatus was present, the dog responded to all three cues. However, when all three apparatus were present, the dog only responded correctly to the "hoop" cue. This suggests that giving the cue just as the learner is beginning to perform the desired behavior is the most effective teaching method.