While studies have shown the importance of listener feedback in dialogue, we still know little about the factors that impact its quality. Feedback can indicate either that the addressee is aligning with the speaker (i.e. ‘positive’ feedback) or that there is some communicative trouble (i.e. ‘negative’ feedback). This study provides an in-depth account of listener feedback in task-oriented dialogue (a director–matcher game), where positive and negative feedback is produced, thus expressing both alignment and misalignment. By manipulating the listener’s cognitive load through a secondary mental task, we measure the effect of divided attention on the quantity and quality of feedback. Our quantitative analysis shows that performance and feedback quantity remain stable across cognitive load conditions, but that the timing and novelty of feedback vary: turns are produced after longer pauses when attention is divided between two speech-focused tasks, and they are more economical (i.e. include more other-repetitions) when unrelated words need to be retained in memory. These findings confirm that cognitive load impacts the quality of listener feedback. Finally, we found that positive feedback is more often generic and shorter than negative feedback and that its proportion increases over time.

The standardization account predicts short message service (SMS) interactions, allowed by current technology, will support the use and conventionalization of ideographs. Relying on psycholinguistic theories of dialogue, we argue that ideographs (such as emoji) can be used by interlocutors in SMS interactions, so that the main contributor can use them to accompany language and the addressee can use them as stand-alone feedback.

Comprehenders predict what a speaker is likely to say when listening to non-native (L2) and native (L1) utterances. But what are the characteristics of L2 prediction, and how does it relate to L1 prediction? We addressed this question in a visual-world eye-tracking experiment, which tested when L2 English comprehenders integrated perspective into their predictions. Male and female participants listened to male and female speakers producing sentences (e.g., I would like to wear the nice…) about stereotypically masculine (target: tie; distractor: drill) and feminine (target: dress; distractor: hairdryer) objects. Participants predicted associatively, fixating objects semantically associated with critical verbs (here, the tie and the dress). They also predicted stereotypically consistent objects (e.g., the tie rather than the dress, given the male speaker). Consistent predictions were made later than associative predictions, and were delayed for L2 speakers relative to L1 speakers. These findings suggest prediction involves both automatic and non-automatic stages.

Interlocutors tend to refer to objects using the same names as each other. We investigated whether native and non-native interlocutors’ tendency to do so is influenced by speakers’ nativeness and by their beliefs about an interlocutor's nativeness. A native or non-native participant and a native or non-native confederate directed each other around a map to deliver objects to locations. We manipulated whether confederates referred to objects using a favored or disfavored name, while controlling for confederates’ language behavior. We found evidence of audience design for native and non-native addressees: participants were more likely to use a disfavored name after a non-native confederate used that name than after a native confederate used that name; this tendency did not differ between native and non-native participants. Results suggest that both native and non-native speakers can adapt to the language of non-native partners through non-automatic, goal-directed mechanisms of alignment during cognitively demanding communicative tasks.