noticing problems in communication

In a collaboration with Nic Fay of the University of Western Australia, we have challenged the claim that participants do not notice problems of miscommunication or incoherence in interaction. We use a referential communication game to show that participants do attend to problems of miscommunication, especially when their confidence in typed descriptions is low, and they use other-initiated repair to address these problems. This work is published in Cognitive Science and can be found here.

repair across modalities

I am working with a number of data sets from experiments on language evolution to show that repair is a ubiquitous feature - even when not being directly tested - and has an influence on the evolution of the communication system in question. I also hope to show that modality affordances operate on repair strategies such that similar modes of communication should exhibit similar strategies across experimental and natural data. Data sets include the following modalities:

• Spoken language and text-based communication

• Signed language and silent gesture

• Graphical communication (Pictionary)

Repair as a site for the cultural evolution of language

In collaboration with Sean Roberts (Bristol University), Justin Sulik (Royal Halloway University), and Hannah Little (University of Western England), I am developing a theory of interaction as a driver of the cumulative cultural evolution of language. I propose that repair sequences are a major site in the innovation and improvement of informative/efficient signals. This work has been presented at:

• EVOLANG12, Torun, Poland, 2018

• Inaugural Meeting of the Cultural Evolution Society, Jena, Germany, 2017

interaction vs. transmission

My previous work on repair in cultural evolution of language experiments has suggested that interaction may provide enough pressure on its own to make language learnable and expressive (where previous studies have suggested that vertical transmission to new learners is a main driver in this process). When interaction is not impoverished, e.g. when it allows for the bi-directional, contingent, and dynamic features of human interaction, then horizontal transmission might be sufficient for creating a systematic language.

Multimodal Repair Initiation

Looking at how repairs are initiated in silent gesture has resulted in a fine-tuned descriptive analysis. I have described how facial gestures, manual gestures, and the body build signals to miscommunication in strategy-specific ways. This work has been presented at:

• International Conference on Multimodal Communication, Osnabruck, Germany, 2017

iconicity in silent gesture

Using video data from a silent gesture experiment, I designed an online experiment to collect naive participants' judgments about individual signals and how much they believed the gestural signal represented the concept. This was done for both the gestures produced initially and after transmission over simulated generations. I used CrowdFlower to design the experiment and collect judgment data. This work has been presented at:

• International Symposium on Iconicity in Language and Literature, Brighton, 2017

• Iconicity Focus Group Workshop, Nijmegen, The Netherlands, 2017

repair in silent gesture

Building from a tradition of iterated learning experiments in the field of language evolution, I have integrated a gradual-turnover paradigm in a more interactive (i.e. face-to-face) setting to the original model. This way I have been able to examine the effects of certain features of interaction on the conventionalize process of a silent gesture communication system. I have found that repair and negotiation lead to efficient gestural systems more rapidly than without these mechanisms in interaction. This work has been presented at:

• EvoLang11, New Orleans, 2016

• International Society for Gesture Studies, Paris, 2016

• Center for Research in Language, UCSD, 2016

social learning, semiotics, and repair in child-caregiver interactions

I have collected video data of natural interactions between a child and his caregivers. With this data I have analyzed how the child uses multimodal resources to build meaning, child language choice (as a Spanish-English bilingual), and how child-caregiver interactions exemplify social learning through the use of talk, gesture, the body, gaze, and semiotic materials embedded in the interactive environment.

peer corrective feedback in the l2 classroom

For my MA thesis, I recorded peer-to-peer interactions in a Spanish L2 classroom. I analyzed the interactions for peer corrective-feedback, documenting how students would go about attending to the errors of their peers.