My research moves across several levels of abstraction. At the most general level, I am interested in the philosophy of inference, in the relation between observation and theory formation, and in the conditions under which regularities in the world can become objects of scientific knowledge. Economics is a particularly difficult and therefore interesting case. It is an observational science whose subject matter is historically mutable, institutionally structured, and inescapably self-referential. Human beings form expectations, respond to theories, revise their behavior, and in doing so alter the very patterns one is trying to explain. The problem is not merely to measure an independent reality, but to understand how order, regularity, and intelligibility arise in a world populated by interpreting agents.

At a more formal level, I work on problems in probability, logic, and statistical inference. I take the Laplacian/Bayesian interpretation of probability to be the most compelling account of quantified uncertainty. Degrees of belief are always conditioned by the information available, and for that reason I understand Bayesian probability as inseparable from information theory. Much of my formal work has therefore centered on the relationship between Bayesian inference, constrained entropy, and statistical mechanics. I am interested in the logical structure of inference under limited information, in the role of entropy as a quantification of uncertainty, and in the way probabilistic reasoning can be grounded in questions of information rather than metaphysics. 

At the applied level, I use these methods to study economic problems that have long interested me, especially those treated with unusual seriousness by the classical political economy of Adam Smith and Karl Marx. What matters in that tradition is the recognition that a social division of labor binds individuals together through dense networks of interdependence, while generating outcomes no one intends and few can fully comprehend from their local vantage point. Market economies are not orderly because individuals are orderly. Quite the opposite. They are turbulent, adaptive, and conflict-ridden systems whose large-scale regularities emerge through decentralized interaction, competition, institutions, and historical constraint. The philosophical problem is how such regularities are possible. The scientific problem is how to infer them without pretending that the noise can simply be wished away.

This perspective informs my work on statistical equilibrium in economic systems. I am interested in how robust macroeconomic and distributional patterns can emerge from heterogeneous agents whose behavior is locally variable, historically contingent, and mediated by institutions. In different projects, I have developed information-theoretic and statistical-mechanical approaches to problems in economic fluctuations, inflation, profit-rate dynamics, firm behavior, and the organization of production. A recurring theme in this work is that aggregate order is not the negation of disorder at the micro level, but one of its consequences. What requires explanation is not why economies are perfectly coordinated, since they plainly are not, but how persistent structures and distributions arise in spite of fragmentation, rivalry, and incomplete knowledge.

A related strand of my work concerns the methodological limits of causal inference. I am interested in the extent to which modern statistical practice asks counterfactual questions that are often poorly aligned with the actual inferential problem. In many settings, the more serious task is not to recover a metaphysically purified causal effect, but to reason carefully under uncertainty with incomplete information. This has led me toward Bayesian and information-theoretic approaches that emphasize conditional inference, partial information, and the representation of uncertainty rather than the fantasy that formal statistical machinery can spare us the burden of judgment.

In the end, my research is unified by a single question. How can one form credible theory about complex social systems when the data are limited, the institutions matter, and the observer is never quite outside the world being observed? My answer has been to work across philosophy, probability, and economics at once. That is less tidy than staying in one lane, but the subject has never shown much respect for neat departmental boundaries.