Moral Tethers

Agents are either guided directly by the elicited or communicated preferences of their principals, including both the conditions or states that their principals value and the particular actions the principals prefer that agents use to realize those states (goal priority), or can research clues to shape their discretionary actions for their principals by examining the nature of their principals and/or the principals' contexts (principal priority). In the latter case, the principal is acknowledged as such - there is a structural tie - but the preferences of the principal have not been expressed to the agent. These are polar cases; most agency is, of course, a mix in which agents both act for the principal’s known preferences and use informed discretion to fill in the gaps. This paper will explore this simple contrast between the basic alternatives of choice of goal priority vs. principal priority. The choice of moral tether is actually a subset of the general problem of principal choice: how agents decide whom to serve. This choice is also one of the basic dilemmas inherent in the processes of normative referencing (see Mitnick 1995). Besides the case of the faithful agent, in which goal and principal priority coincide, three additional principal alignment problems result from the possibilities of goal and principal priority: tethering the agent, converting the agent, and importing the agent. The paper presents a simple theory of agency and information-seeking as well as a series of propositions dealing with the circumstances of resolution of the alignment problems given conditions of uncertainty, resource scarcity, cognitive limits, and certain relational and social contexts. Several illustrations and applications are described