Categorisation in Infancy
In our studies on infant categorisation we address the perception of similarity and dissimilarity in the developing mind. One focus of our research in this area is the impact of labelling on category learning in preverbal infants: does hearing similar words for similar objects facilitate category formation?
Category formation during familiarisation: order effects
Whether infants are able to learn a category is established in familiarisation / novelty preference procedures. Current research is aimed at pin-pointing the role of familiarisation contingencies (order effects, primacy and recency effects) in determining infants' behaviour on a novelty preference test trial. This work focuses on integrating category formation as a continuous process where the time course of learning is a highly relevant factor.
Labelling and categorisation
Whether hearing labels helps infants form categories (because it provides additional information) or whether they disrupt category learning (because more sensory signals have to be processed simultaneously) is controversial: Past research has provided evidence pointing both ways, with no single result being conclusive. Our work currently focuses on factors that may determine whether labels are beneficial or increase processing load to the extent that they disrupt categorisation.
Feature variability and feature saliency
Several undergraduate projects currently focus on what is perceived as similar by infants between 6 and 12 months. In particular, we ask what makes a feature salient enough to be used as diagnostic for categorisation, and how are variable versus invariable features used in category learning.
Timing matters: The impact of label synchrony on infant categorisation
Some recent information on categorization can be found in this paper by Althaus and Plunkett (2015)
The study on categorisation asked the question of whether hearing names for new objects helps babies to group them into a category. For instance, imagine you’re going for a walk with your baby and you see a Golden Retriever, so you say something like “Look, it’s a doggie!”. While the baby starts learning about the word “dog” this is also telling her that this is an animal that’s similar to other “doggies” she knows (like, the neighbour’s dog, for example).
In this study, 12-month-olds were familiarised with eight images of objects from a “novel” category (see picture above) made up of leaves and shells. The timbos were constructed so that the shells were quite variable across different objects, but the leaves were highly similar. In addition to seeing the pictures, half the babies also heard a name for the objects, “timbo”.
The babies’ eye-tracking data showed us that babies who heard names were learning about the objects by looking more at the leaves, whereas the babies who only saw pictures were focussing more on the shells during learning. We think this means that hearing names helps babies find similarities, or “diagnostic features” between different objects – they learned that “a timbo is something that has a leaf on one side”.
While words can therefore be really helpful for learning about objects, they only have this function if the baby has enough time to look at the picture before hearing the word. When we played the word “timbo” at the same time as the picture came on the screen, the babies found it much harder to learn. This tells us that the integration of words and objects involves complex processes that are challenging for babies who are just starting to learn words.