Desiderata for the (Possible) General Theory of Interpretation

It currently¹ seems plausible to me that something like a general theory of interpretation might be a necessary building block for how minds work in general. Interpretation has been a recurring theme in my thinking about these topics, as well as a major source of confusions (perhaps even a root source or, even more, the root source) and it seems to present a difficulty that, once resolved, might open the floodgates of insights. Here I point to examples of relevant stuff that matches at least our (my?) intuitive concept of interpretation and what the path forward might look like.

Some motivating motivations

Interpretation of observations/messages/statements

Building a model of the world based on observations involves, first, interpreting those observations as meaning something about the world.

This does not necessarily mean that the observations are treated as messages from an agent intended to mean something or refer to something. However, they are a potential source of relevant knowledge about something.

Interpretations of other agents

When Alice observes Bob doing something, often Alice can interpret Bob's behavior as being driven by some beliefs, desires, preferences, other mental states, attitudes, proclivities, and habits.

Interpretation in underterminacy/indeterminacy of the good

Sometimes it's not clear whether something is good in a way that is not reducible to any kind of uncertainty (empirical, indexical, computational, or whatever). Rather, the agent is encountered with some genuinely radical novelty and is thus forced to interpreted it a certain way, even if how it is interprted is left as an open case.

Interpretation of self

Humans have some degrees of freedom in how they interpret themselves, their past behavior,² what has happened to them etc, in a way such that it seems rather janky to try to reduce it all to rearrangement of probabilities or even with a modified ontology.

Speaking of ontologies…

Ontological crises

Faced with an ontological crisis, the agent needs to interpret its past belief structure into the new ontology. To the extent that the new ontology is underdetermined at the moment of the crisis, Everything needs be re-interpreted, including the ontology itself.

"It's just rearrangement of probabilities"

One could say that "Yeah but all of this could be modelled as the agent rearranging it's probabilistic beliefs when faced with some evidence etc.".

Glib responses:

• Even if it all, in the end (e.g. for the purpose of deciding on the course of action), cached out in terms of a belief update, there would be a lot of important stuff going on between the arrival of (generalized) evidence and that update; a lot of stuff that is worth studying. I would be claiming that this "lot of stuff" includes (i.a.) interpretation.
• I don't expect probabilism to be adequate, either descriptively or prescriptively or modelling-wise, as the foundation of a theory of intelligence/agency/etc; in particular, in this specific case.

(Going a bit meta, a statement like "It's just rearrangement of probabilities." is itself a kind of dismissal of the problems of interpretation by the means of interpreting them as not worthy of consideration.)

Synonyms

• "treat X as Y"
• "take X as Y"

Relevant

• There's a notion of interpretation in model theory which might be relevant.
• Hermeneutics?