Memory for speaking and listening

We learn through language, sometimes by ourselves, through reading or watching the news, and sometimes with others, through conversation. Perhaps it would make sense for us, then, to remember what we read or hear especially well. However, evidence from memory research in the last few decades suggests that people remember the language they produce themselves better than the language they comprehend. In my doctoral thesis, I asked whether the above positions can be reconciled by taking into account natural language and communicative contexts, which have generally been overlooked in memory research. First, I ensured that, in non-communicative contexts, people remember their own speech better than speech they read or hear. I used different picture naming tasks, in which people say the names of objects they see on a computer screen, to collect speech that is fairly natural and includes conceptual processing. In this way, I identified two factors that improve memory during word production: People remember words they think of themselves better than words they hear and they remember words they say aloud better than words they say silently. These findings are consistent with well-known memory phenomena, the generation effect and the production effect. I then looked at short question-answer exchanges between two people. I found that participants listening to recorded question-answer exchanges remembered words mentioned in the answers better than they remembered words mentioned in the questions. Presumably the reason for that is that people consider answers more important than questions because they fulfil the communicative goal of the exchange, i.e. to gain information. I also found that, in studies with two participants, people still remember what they say better than what they hear, but that difference shrinks if what they hear answers one of their questions. As such, people remember different things from conversations depending on their roles