"There is an incompatibility between the symmetries of causal structure in relativity theory and the signaling abilities of probabilistic devices with inputs and outputs: while time reversal in relativity will not introduce the ability to signal between spacelike separated regions, this is not the case for probabilistic devices with spacelike separated input-output pairs. We explicitly describe a nonsignaling device which becomes a perfect signaling device under time reversal, where time reversal can be conceptualized as playing backwards a videotape of an agent manipulating the device. This leads to an arrow of time that is identifiable when studying the correlations of events for spacelike separated regions. Somewhat surprisingly, although the time reversal of Popescu-Rohrlich boxes also allows agents to signal, it does not yield a perfect signaling device. Finally, we realize time reversal using postselection, which could to lead experimental implementation."

"New measures are proposed for mutual and causal dependence between two time series, based on information theoretical ideas. The measure of mutual dependence is shown to be the sum of the measure of unidirectional causal dependence from the first time series to the second, the measure of unidirectional causal dependence from the second to the first, and the measure of instantaneous causal dependence. The measures are applicable to any kind of time series: continuous, discrete, or categorical."

"We propose two graphical models to represent a concise description of the causal statistical dependence structure between a group of coupled stochastic processes. The first, minimum generative model graphs, is motivated by generative models. The second, directed information graphs, is motivated by Granger causality. We show that under mild assumptions, the graphs are identical. In fact, these are analogous to Bayesian and Markov networks respectively, in terms of Markov blankets and I-map properties. Furthermore, the underlying variable dependence structure is the unique causal Bayesian network. Lastly, we present a method using minimal-dimension statistics to identify the structure when upper bounds on the in-degrees are known. Simulations show the effectiveness of the approach."

"Common methods of causal inference generate directed acyclic graphs (DAGs) that formalize causal relations between n variables. Given the joint distribution of all these variables, the DAG contains all information about how intervening on one variable would change the distribution of the other n-1 variables. It remains, however, a non-trivial question how to quantify the causal influence of one variable on another one.
Here we propose a measure for causal strength that refers to direct effects and measure the "strength of an arrow" or a set of arrows. It is based on a hypothetical intervention that modifies the joint distribution by cutting the corresponding edge. The causal strength is then the relative entropy distance between the old and the new distribution.
We discuss other measures of causal strength like the average causal effect, transfer entropy and information flow and describe their limitations. We argue that our measure is also more appropriate for time series than the known ones.
Finally, we discuss conceptual problems in defining the strength of indirect effects."

Michael preaches the Gospel According to Pearl; and very nicely too. (I would dispute however that DAGs don't give us a handle on mechanisms.)

"The Explanation of Social Action is a sustained critique of the conventional understanding of what it means to "explain" something in the social sciences. It makes the strong argument that the traditional understanding involves asking questions that have no clear foundation and provoke an unnecessary tension between lay and expert vocabularies. Drawing on the history and philosophy of the social sciences, John Levi Martin exposes the root of the problem as an attempt to counterpose two radically different types of answers to the question of why someone did a certain thing: first person and third person responses. The tendency is epitomized by attempts to explain human action in "causal" terms. This "causality" has little to do with reality and instead involves the creation and validation of abstract statements that almost no social scientist would defend literally.
This substitution of analysts' imaginations over actors' realities results from an intellectual history wherein social scientists began to distrust the self-understanding of actors in favor of fundamentally anti-democratic epistemologies. These were rooted most defensibly in a general understanding of an epistemic hiatus in social knowledge and least defensibly in the importation of practices of truth production from the hierarchical setting of institutions for the insane. Martin, instead of assuming that there is something fundamentally arbitrary about the cognitive schemes of actors, focuses on the nature of judgment. This implies the need for a social aesthetics, an understanding of the process whereby actors intuit intersubjectively valid qualities of complex social objects. In this thought-provoking and ambitious book, John Levi Martin argues that the most promising way forward to such a science of social aesthetics will involve a rigorous field theory."

"Probabilistic graphical models are a fundamental tool in statistics, machine learning, signal processing, and control. When such a model is defined on a directed acyclic graph (DAG), one can assign a partial ordering to the events occurring in the corresponding stochastic system. Based on the work of Judea Pearl and others, these DAG-based "causal factorizations" of joint probability measures have been used for characterization and inference of functional dependencies (causal links). This mostly expository paper focuses on several connections between Pearl's formalism (and in particular his notion of "intervention") and information-theoretic notions of causality and feedback (such as causal conditioning, directed stochastic kernels, and directed information). As an application, we show how conditional directed information can be used to develop an information-theoretic version of Pearl's "back-door" criterion for identifiability of causal effects from passive observations. This suggests that the back-door criterion can be thought of as a causal analog of statistical sufficiency."

Uncle Judea sounds a bit testy in this one, but no doubt anyone would be if they had to keep swatting down such pathetic misunderstandings passing for objections.

The abstract makes this sound closely akin to good old-fashioned Lloyd-Pagels thermodynamic depth.

Described by Uncle Judea as "A new survey paper, gently summarizing everything I know about causation (in only 43 pages)".

An illustration of the necessity for counterfactuals.

Damn you, Clark! (Not really.)

Note the extreme weakness of the sense in which Pearl is even "half-Bayesian"; the blessed St. Jerzy could agree with it.

Interesting sounding paper on using sex of congressmember's children as a definitely exogenous source of variation w.r.t. voting & campaign donations. To read and ponder

Not read yet, comes recommended by the orgtheory crowd

Nihat on an information-theoretic refinement of Reichenbach's common-cause principle

Andy Gelman reviews Brooks & Manza's interesting-sound book on this subject. Gentle but firm corrections re causal inference.

Concealed causal theories in apparently plain historical narratives. (Cf. Popper, _Poverty of Historicism_.)

Scarily, the abstract almost makes sense.