The postdoc’s speech production is regulated by the algorithm below:
- Estimate probability distribution over potential current locations.
- Take maximum of said probability distribution and retrieve k most likely languages.
- Reweight k most likely languages according to a) distance to a border; b) historical and geopolitical situation of the area.
- Activate top language. Get worried about own level of proficiency and interference effects from current reading language, current thinking language, last text message language, last night’s nighmare’s language.
- Generate utterance. Usually something simple like ‘hello’, ‘I’m sorry, I’m not from here’, or ‘just a second, my ticket is between my computer cable and that banana I took from Italy to Sweden and am now taking back to Italy’.
- Realise predicted location distribution from step 1 was way out.