Herein, we explore the psychology of deontic reasoning through the
presentation of a heterogeneous natural logic combining inference
schemas with a preference-based model-theoretic semantics such as
those typically found in various formalisms for nonmonotonic
reasoning. We conjecture that the heterogeneous approach is a
generalization of various other hypotheses concerning deontic
reasoning, and provides a robust framework for explaining semantic
intricacies which are present in so-called ``deontic paradoxes."
As an initial investigation, two theories were tested: The first
hypothesis states that people represent an obligation as a
conditional statement which explicitly includes the concept of
violation, and the other postulates that people not only prefer
deontically perfect situations to less-than-perfect situations,
but also have preference between these sub-ideal situations. Two
sets of experiments were conducted in order to gain some insight
regarding these two ideas, and the results show strong evidence
supporting our initial intuitions.
Conflict between groups of armed men is as old as
recorded history. Effective reasoning and decision-making are
fundamental to the successful execution of military operations.
These activities are of paramount importance, given the
high-stakes nature of conflict; most especially in this modern
era of asymmetric threats, and unconventionally armed rogue
states. Yet as high as the stakes are, there does not exist a
sufficiently formal military theory of reasoning and
decision-making that instantiates modern warfighting doctrine.
Large bodies of knowledge on reasoning and decision-making exist,
but they are not integrated, and they (to the author's knowledge)
have not been cast effectively into a military context. Herein, I
describe a new theory of military rationality which fully
captures the reasoning and decision-making processes of homo militius, military man. The goal of the third
generation wargaming effort at the Air Force Research Laboratory's
Information Directorate is to produce a high-fidelity simulation
of conflict environments in order to facilitate a new brand of
highly immersive training for our warfighters and supporting
personnel. This environment will be populated by a new breed of
intelligent agents that we affectionately call ASC-ME's (Advanced
Synthetic Characters for Military Environments). I shall briefly
highlight the philosophical foundations for the construction of
such entities, and the formal techniques by which they may be
modelled and engineered.
Computational models that give us insight into the behavior of
individuals and the organizations to which they belong will be
invaluable assets in our nation's war against terrorists, and
state sponsorship of terror organizations. Reasoning and
decision-making are essential ingredients in the formula for human
cognition, yet the two have almost exclusively been studied in
isolation from one another. While we have witnessed the emergence
of strong traditions in both symbolic logic, and decision theory,
we have yet to describe an acceptable interface between the two.
Mathematical formulations of decision-making and reasoning have
been developed extensively, but both fields make assumptions
concerning human rationality that are untenable at best. True to
this tradition, artificial intelligence has developed
architectures for intelligent agents under these same assumptions.
While these digital models of "cognition" tend to perform
superbly, given their tremendous capacity for calculation, it is
hardly reasonable to develop simulacra of human performance using
these techniques. We will discuss some the challenges associated
with the problem of developing integrated cognitive systems
for use in modelling, simulation, and analysis, along with some
ideas for the future.