I use TPRS amongst other “Comprehensible Input” methods in teaching Spanish. One thing that seems to really be controversial about TPRS is the use of translation. Admittedly, I do use translation, especially in level 1. It serves a couple of purposes: comprehension checking and clarification.
When using a word that is new or unfamiliar, I may ask “¿Cómo se dice ____ en inglés?” just to make sure the class understands what I am saying. In TPRS, we strive to make the class as close to 100% comprehensible as we can.
When reading, I do have kids translate out loud in the beginning. This allows me to have a clear view of their level of comprehension. After we read in translation, we go back and have a Spanish discussion of what we just read. However, by the latter part of semester one, we are moving away from this model and more frequently just reading and discussing in Spanish.
I write words on the board frequently and write the English equivalent as well. When comprehension breaks down, I can point to the board to alleviate the confusion. I also will answer any student who asks what something means. Once students are able, I can define words in Spanish whenever possible.
Deciding to use or not use translation is a personal choice for the teacher. My philosophy personally is that I’d rather use translation when and if it expedites learning, thus allowing the class to quickly get back to more comprehensible input faster. I also view it as one of many tools available to increase comprehension. Furthermore, use of translation should decrease over time as students become more proficient.
If anyone knows of research studies that correlate incidental use of translation in the ways I have described here (I am not referring to the “grammar translation method” nor am I referring to teaching translation as a skill), I would be most interested in learning more!