it is said (i did not look for the aricle) that after ~20 minutes of hearing an unknown language,
your neural activity shifts to another pattern.

If we restrict the language to have only 3 syllables,
and talk for some minutes, we start giving names to things,
and relating sounds to mood tags,
and to communicate.

Say the three syllables are bu ba co,
one utters:
"bubaco co baco bubu ba!"
Someone else:
"baco bubu ba baba. co baco bubu!."

And this process continues on.
I have not tried this through text chat,
only in person sometimes within periods
that ranged from 20 min to hours.

You come up with a language by doing so
and we might obtain some insights about
what is crucial for a minimal conlang.