Hi, my name is Liam Blything. I’ve recently returned to the UK after a 2 and a half year stint at University of Alberta in Canada to work with LuCiD as a postdoctoral researcher on the work package investigating how children learn pronouns. In this blog, I’m going to introduce some of the research that I did in Alberta on how children and adults understand pronouns, and I’ll touch on how that transitions to the LuCiD project.
When do we use pronouns?
With the goal of communication in mind, speakers can choose whether to talk about someone using a proper name (e.g., Serena Williams), or a reduced name like a pronoun (e.g., she, her, herself). When we introduce someone as the current ‘topic’ of conversation, we’ll typically use a proper name, as with Serena in (1a):
(1a) Serena is the favourite for this year’s Wimbledon tennis championships.
However, the more someone is established as the topic within the context, the more a pronoun such as she is preferred for their next mention: the listener would find it more difficult to interpret a follow up sentence to (1) that repeats the proper name (as in 2a) than one which uses a pronoun like she (as in 2b).
(2a) Serena would be the first Mum to win since 1980.
(2b) She would be the first Mum to win since 1980.
Popular research questions about pronouns.
How do listeners infer that the reduced information offered by she in (2b) communicates the same information as the proper name Serena in (2a)? This is ‘no mean feat’ given the lack of meaning offered by a pronoun alone, as well as the speed that is required to make that interpretation with more incoming speech!
One way to investigate this is to use ambiguous pronouns: the context often offers two or more plausible referents. Take a further example:
(3a) Serena won most of the points against Venus.
(3b) She started to reminisce about all those times they had played as children.
Here, Serena and Venus are both plausible referents to she. That said, the listener would usually think that she refers to Serena and not Venus. Broadly speaking, there are a number of language clues that can combine to make one person more topical and therefore more likely to mean she. One established example is called a “first mention” clue, which contributes to Serena being preferred as she in (3b).
Another example clue is provided by the meaning of a verb, which tells us whether a character is the cause of an event. For example, in (4a) the first mentioned character Serena is most causal because she is doing something to frighten Venus; whereas in (4b), Venus is causing Serena to be fearful and becomes the more likely meaning of she. The explanation here is that a causal meaning is associated with making a character more identifiable in memory and therefore more likely to be the topic that the pronoun refers to.
(4a) Serena frightened Venus during their match. She was hitting very powerful shots.
(4b) Serena feared Venus during their match. She was hitting very powerful shots.
My research on children’s understanding of pronouns.
One of my studies, recently accepted in The Journal of Child Language, revealed that, unlike adult pronoun interpretation which appears to be driven by cues like verb meaning, children appear to rely more heavily on straightforward cues like first mention. I’m interested in what learning mechanisms are involved in how children acquire an adult like ability. At Liverpool I am asking how children acquire a combination of complex language ‘clues’ (or ‘constraints’) and how this is learned through discourse and context.