Over the last few years, Conversation Analysis researchers have moved well beyond the American English origins data that the founders used; in ROLSI, for example, it’s quite normal that English be only one of three or four different languages studied in any one issue. But what is really exciting is when a research team takes on a big, cross-cultural project, and I’m very happy to have Giovanni Rossi, N. J. Enfield and Mark Dingemanse tell us about their admirable new collection – and it’s open-access, too.

Researching talk and other conduct in social interaction is hugely challenging at all stages of the process, from data collection to transcription to analysis. So it is no wonder that most studies limit the scope to a single language. But what do we do when we want to compare interaction across languages and cultures, to probe questions of human universals versus cultural variation?

Interaction researchers now have access to more diverse data and analysis than ever before, but when different language studies are designed and carried out independently, rigorous comparison is impossible. Upon the publication of our open-access book “Getting others to do things: A pragmatic typology of recruitments,” we want to share our experience of collaborating with an international team in designing and carrying out a major comparative project.

A Bit of Background. In their introduction to their 1989 “Cross-cultural pragmatics” [1], Blum-Kulka, House, & Kasper explained their approach to comparing speech acts across languages. They used a written discourse-completion test because, they said, collecting an adequate sample of comparable data from natural interaction in multiple countries “would have been virtually impossible under field conditions”. Two decades later, we set out to meet this challenge, as members of the ERC project Human Sociality and Systems of Language Use (HSSLU), hosted at the Max Planck Institute for Psycholinguistics from 2010-2014. One of the three HSSLU subprojects was on recruitments: how people get others to do things. Together with our broader team of collaborators, we collectively developed the notion of “recruitment” to capture sequences of interaction in which one person’s behavior leads to another’s help (whether or not such help was intentionally sought). Recruitment sequences are diverse in kind, including request-like and offer-like exchanges, forms of assistance and collaboration, realized through verbal, nonverbal, and composite practices [2]. This table lists the researchers and languages studied:

Language	Location	Researcher
Cha’palaa	Ecuador	Simeon Floyd
English	UK, US	Kobin H. Kendrick
Italian	Italy	Giovanni Rossi
Lao	Laos	N. J. Enfield
Murrinhpatha	Northern Australia	Joe Blythe
Polish	Poland	Jörg Zinken
Russian	Russia	Julija Baranova
Siwu	Ghana	Mark Dingemanse

Through the course of this team project, we developed and refined a protocol for team comparative research on language in social interaction.

A first challenge was to ensure that the data were maximally comparable, by minimizing any obvious cultural influence on the practices being observed. Because cultural variation is most pronounced in institutional and formalized genres of interaction, we limited the data to maximally informal interaction in home, family, and village settings. This is the foundational context of human sociality, in which all humans face the same essential problems of social coordination [3].

A second challenge was to obtain enough cases in our data. While task-focused interactions such as cooking together are teeming with recruitment sequences, talk-focused interactions may go by with hardly any. So we needed corpora representing a broad range of settings and activities. Also, we needed high-quality video recordings. Such data are hard to come by even for major languages like Italian, Lao, Polish, and Russian, let alone for minority languages like Cha’palaa, Murrinhpatha, and Siwu. We therefore set out to collect our data from scratch, each team member traveling to their respective field sites to build the needed video corpora. 

Recording was of course only the beginning. Then there was transcription and translation, working with native speakers in the field. This work presupposes significant background qualifications on the part of each researcher, who had already spent years establishing the needed connections with host communities and knowledge of the languages and their cultural settings. Back from the field, we embarked on a sustained period of team data sessions and huddles. We became steeped in each other’s empirical materials and developed a deep and shared understanding of the languages, and of the practices under examination. Our collaborative qualitative analysis of cases in all languages provided a foundation for subsequent comparative and quantitative analyses, feeding into the design of a coding scheme to define formal and functional measures. We piloted the scheme and revised it in multiple rounds, before team members independently coded each language’s dataset for comparison. To confirm that our coding was done consistently, we carried out a coding reliability check. The final step involved statistical modeling of the coding results to understand dependencies among our observations. 

To summarize our six-step method for comparative team-based research: 1) Record, 2) Transcribe, 3) Confer, 4) Code, 5) Check, 6) Model. This protocol requires a team science approach. Given the demanding combination of fieldwork (Steps 1, 2), expertise in linguistics and conversation analysis (Steps 3, 4), and quantitative methods (Steps 5, 6), this could never have been done other than by an open-minded, interdisciplinary team.

Results of this team collaboration, and the shared conceptual work underpinning it, have appeared in several publications since results first emerged, for instance at ICCA 2014, and in Language, ROLSI, and Royal Society Open Science. The edited volume published makes the coding scheme available and provides outlines of the recruitment system in each of our eight languages. We are exceedingly happy to present the project available to all in an open-access book.

References:

[1] Blum-Kulka, Shoshana, Juliane House, and Gabriele Kasper, eds. 1989. Cross-Cultural Pragmatics: Requests and Apologies. Ablex.

[2] Floyd, Simeon, Giovanni Rossi, N. J. Enfield, Julija Baranova, Joe Blythe, Mark Dingemanse, Kobin H. Kendrick, and Jörg Zinken. 2014.“Recruitments across Languages: A Systematic Comparison.” ICCA 2014.

[3] Dingemanse, Mark, and Simeon Floyd. 2014. “Conversation across Cultures.” In The Cambridge Handbook of Linguistic Anthropology. CUP.