Table 1 Principles derived from major word learning theories
Task | Theory | Description | Example Reference(s) |
|---|---|---|---|
Mutual exclusivity assumption | Constraints theory | A speaker provides novel labels for several familiar and novel objects. Participants tend to apply novel labels to the novel object as opposed to the familiar object | |
Whole object assumption | Constraints theory | A speaker labels novel objects with various unfamiliar parts. Participants tend to apply novel labels to the whole object as opposed to a part of the object | Markman & Wachtel, (1988) |
Taxonomic bias | Constraints theory | A speaker presents participants with a target object (e.g., dog), followed by thematic associate (e.g., bone) and a taxonomic associate (e.g., another dog). The target object is labeled with a novel word (e.g., “wug”). When asked to find another “wug”, participants are more likely to choose the taxonomic associate | Markman and Hutchinson (1984) |
Shape bias | Constraints theory | A speaker labels a novel object (e.g., “wug”). Participants are then shown three other objects that match in size, color, or shape. When asked to find another “wug”, participants are more likely to choose the object that matches in shape | |
Pointing during learning | Sociopragmatic theory | A speaker labels a novel object while pointing at or not pointing at the novel object. Participants are more likely to learn the label of the novel object when the speaker is pointing at the object | Booth et al. (2008) |
Looking to visible object | Sociopragmatic theory | A speaker labels a novel object while looking at or not looking at the novel object. Participants are more likely to learn the label of the novel object when the speaker is looking at the object | Booth et al. (2008) |
Looking to non-visible object | Sociopragmatic theory | A speaker plays with a novel object and places it in one of two buckets in front of the participant. The novel object is no longer visible to the participant. The speaker labels the object while it is in the bucket. Then, the experimenter removes the labeled object from the bucket and another object in the second bucket that was not labeled. Participants are able to accurately identify the novel object that labelled while not visible | Baldwin (1995) |
Overheard speech | Sociopragmatic theory | Two speakers label a series of novel objects in the presence of the participant. The labeling events are not explicitly directed at the participant. Nonetheless, the participant can accurately map the labels to the objects | Tomasello et al. (1995) |
Massed vs. spaced learning | Domain-general theory | A speaker provides novel objects and labels on a massed (i.e., in immediate succession) or spaced (i.e., distributed across time) schedule. Participants display higher retention of novel objects presented on a spaced schedule | |
Cross-situational word learning | Domain-general theory | A speaker provides several novel objects and labels on-screen in a single trial (“This is a wug. This is a dax.”). At first, it is ambiguous which word corresponds to which object. Across the learning phase, however, words and objects co-occur in a reliable manner. Participants can accurately map words to objects presented cross-situationally | Smith and Yu, (2008), Vlach and DeBrock, (2017) |
)Same vs. varied context | Domain-general theory | A speaker provides novel objects and labels on a consistent background (e.g., same patterned cloth) versus a varied background (e.g., different patterned cloths). Participants display higher retention of novel objects presented on a varied background than the same background | Smith and Rothkopf (1984), Smith et al. (1978), Vlach and Sandhofer (2012a) |