Okay here goes my hypermess on point 2, “language for thinking” VS “language for communicating”:
In language for thinking, the main benefit seems to be that “only if something is written down it is fixed enough to be discussed independently from the author” (Ahrens, 2017).
When we’re just on our own, we can use symbols, metaphors, analogies, coined words that are meaningful to us, anchored to our brain but that only we understand. We can work closer to the real associations of the mind, and thus not overload the short term memory bottleneck with terminology. We can also be fuzzier, maybe?
I suspect this may work less well if we’re talking to our future self, where maybe the mental associations (context) have changed.
In language for communicating I’d separate the classic linear form from the hypertextual form.
In classic linear form, the art is not only about thinking about our audience and the terms they would understand, but choosing what’s the adequate throughline that connects the ideas we want to express. In other words, in which sequence we paint the colours in the other person’s mind so that she’ll understand what we’re going to introduce later on. This is what Ted Nelson calls gradually condensing your flying pages.
In hypertextual form, I feel in unexplored territory. We do have wikis and other sorts of generally unidirectional hypertext, but I don’t know good examples of hypertext conceived as a communication form (from “I wrote a blogpost” to “I built a mesh of text, here’s an entry for you”).
Regarding how these two ways of looking at language (and associated workflows) should shape tools differently, @szymon_k what do you have in mind? The only slightly related thing that comes to my mind is: tools don’t help you much moding yourself in creative exploration mode (divergent) or judgemental edition mode (convergent).