More impressions from our visit to the Yucatan

Hi All,

Mexico was great. Though this time it was not about the gorgeous beaches of Cancun, or a mid stop for Havana. In fact, street drivers kept on offering “I’ll take you to Cancun, to Chichen, to Coba” and looked puzzled when we declined time and again. We had other exciting plans! This time our goal was to meet, interview, record, and teach a class to Yucatec and Spanish speakers!

Thanks to Florian’s knowledge and advice, I taught a class in psycholinguistics in Spanish at the Universidad de Oriente using the space and time of Teresa Miyar’s class (Thanks Tere!). The slides I used are here. The class was a great experience as part of giving back to the community while we were there. Equally enriching was interacting and recording Yucatec and Spanish speakers. All the people we talked to were super cooperative, helping us in any way possible they could think of, and even feeding us *delicious* high level yucatecan cuisine prepared by the cooking school conveniently located few doors down our recording lab.

At our peak, we were running as much as 10 subjects on 8 computers at once. This was quite intense, even with prerecorded spoken instructions in Spanish with Yucatec examples. People reported to have enjoyed the experiments, and were happy to participate for the multiple experiments. Speakers had different levels of bilingualism. In fact, I expect to find some code switching between Yucatec and Spanish, since many of them reported to not knowing how to name certain things in the stimuli, and the Spanish word instead. Transcription and annotation is currently underway, so we will know more about this later.

We run the Fruitcarts experiment in Spanish language. This involved a naive speaker and a confederate. The speaker would instruct how to move, rotate, and paint objects on the screen according to an end-state world map given the speaker can look at for reference. In the English data we saw a striking difference between a strategy that would select an object first, and then apply an action to it (we termed this bi-clausal strategy), versus simply applying the action to the object and thereby selecting it (we term this mono-clausal strategy). We attributed this difference to the amount of information needed to be placed in one single clause. If an object is harder to describe due to it’s contrasting context, speakers tended to plan for a bi-clausal production; while simpler descriptions were usually planned in a mono-clausal plan (Gomez Gallo, Jaeger and Smyth 2008).

During a preliminary glance at the Spanish data for the Fruitcarts experiment, a number of interesting issues came up. In particular, speaker’s consistent use of monoclausal plan, heavy use of relative clauses, and a Gricean approach to the description of object attributes. I’ll explain.

Spanish speakers were consistently using mono-clausal plans regardless of the complexity of the object description. This took us by surprise since, indeed, a bi-clausal plan is perfectly available in Spanish. Instead, a typical sentences was “mueva el triangulo que es pequeno que tiene una estrella en la esquina hacia el distrito federal” (move the triangle that is small that has a star on the corner to the federal district). Spanish speakers made heavy use of relative clauses to describe the object attributes necessary to identify the object under consideration.

Even though English speakers could also use postnominal attributes within relative clauses, in this case, they would prefer a bi-clausal plan in which attributes are produced in a single separate clause preceding the move action. A possible hypothesis which we need to test once things are fully transcribed and annotated is that prenominal adjetives in English may be a driving factor in the bi-clausal choice. In Spanish, most adjetives are postnominal (only few can be prenominal, and those would be a marked construction changing the meaning of the referential expression, for example: mi viejo amigo versus mi amigo viejo). This structural difference may allow Spanish speakers to chose the move verb early, elicit the object type to be moved (‘triangulo’) without worry to early about the particular complexity of the object under consideration.

In English, even when postnominal relative clauses are possible, they may carry a penalty over prenominal adjetives (for example: the small triangle versus the triangle that is small). Such penalty may not be as strong or existent in Spanish.

There was evidence that speakers relied on comprehenders’ interpreting their state of knowledge within context. In the Fruitcarts maps, there are certain triangles with no decoration on them. Some speakers would say “move the triangle to the federal district” which by the minimal use of description, that is no diamond on the corner (no relative clauses or posthead adjetives), speaker intention was to move the plain triangle with nothing on it.

More as things are annotated!