Axioms connecting goal syntax and semantics
Determine what semantic implications, if any, a particular compositional account of goals must have, and characterize the relationship between the syntactic and semantic properties of a goal language.
References
Open Question: Axioms for Goals? However, these are just the choices of relations between goal syntax and semantics the community has explored so far. We suggest that this framing highlights a significant conceptual opportunity to explore alternative compositional and semantic accounts of goals. Moreover, it positions us to scrutinize the needed relationship—if any—across the syntactic and semantic properties of a goal language. If we embrace a particular kind of compositional account of goal, what must this imply, if anything, about our semantic view of goals?
Open Question: Syntactic affordances and semantic expressivity as constraints on cognition? Lastly, a syntax-semantics distinction for goals can provide a useful analytical frame for investigating goal representations in human and animal cognition. In our view, this is not unlike studying the same cognitive process from, say, the computational versus algorithmic levels of analysis [Marr, 1982, Anderson, 1990]. Specifically, certain properties of goals may be better understood in terms of their syntactic constraints (e.g., how well they afford efficient computation) while others in terms of their semantic constraints (e.g., how well they express rational behaviors in a certain environment). Characterizing how these distinct kinds of constraints interact with one another may also shed light on rational accounts of human cognition [Griffiths et al., 2024] and their computational plausibility [Rich et al., 2020].