Hi Claudia,
In this game, I recommend using the languages (R, S, T, and Y) as the base because of the certainty associated with the languages. As you mentioned, we know exactly how many researchers can learn each language, but we do not know exactly how many languages each researcher can learn. This makes the languages an intuitive base for the game. Additionally, it is easier to visualize and diagram the rules when using the languages as the base than it is to do so when the researchers are the base (for example the LG not block compared to a different representation of the rule when the researchers are the base).
Your inferences are still correct regardless, which makes sense since the inferences shouldn't change even if your base changes (after all, the rules are the same no matter how you visualize the game)! It's just easier to use the languages as the base for the aforementioned reasons
If you encounter a similar question in the future, consider the certainty of the different variables when determining a base. For example, if you know that there are three groups people could fit into, but you don't know how many people go in those groups, make the groups the base! You can also consider how the rules are phrased and what variables they focus on, as this can provide further direction for determining a base. Each game is a little bit different, and some even have a somewhat intuitive component (such as selecting the buildings trucks are going into as the base instead of the trucks being the base because it's much easier to think about trucks moving as opposed to buildings!).
I hope this helps!
Kate