Christopher Hitchcock
The Other Trolley Problem: Causation and Counterfactuals in Berry v. Sugar Notch

This talk focuses on the the American Civil case of Berry v. Sugar Notch (Pennsylvania State Supreme Court, 1899). The facts of the case are very simple, and almost everyone agrees with the decision in favor of the plaintiff. In this case, the defendants made an argument based on the familiar ‘but for’ test for causation. While everyone who hears this argument dismisses it as sophistry, it is actually very difficult to say precisely what is wrong with it. I will use this case to raise a number of issues about causal and counterfactual reasoning in the law.



Paolo Garbolino
Italian Law and Token Causation

A rule of the Italian Supreme Court has fixed some criteria for attribution of causal responsibility in case the agency increases the probability of the harmful outcome. The paper analyzes whether these criteria can be satisfied by probabilistic theories of token causation.


Philip Dawid
Statistical Bounds for the Probability of Causation

In a toxic tort case, the Law might be concerned with whether or not an individual’s exposure to a certain chemical was causally responsible for her illness. However, given even the best possible data on the statistical dependence of this outcome on exposure in the population, we can typically only provide bounds for the “probability of causation” in the case of a specifi c individual. We describe these bounds, and show how they can be improved if population information is available on additional variables, such as covariates or mediating variables.




Claudio Pizzi
The Franzese Judgment: A Look at Open Problems

The so-called Franzese judgment (July 11, 2002) introduced an important step in Italian jurisprudence by fixing clear ideas on the following points:

  1. the causal nexus between token-events must be ascertained by applying the so called “formula of the conditio sine qua non”;
  2. the truth of the involved counterfactual relies on presupposed laws of nature or so-called “experience maxims”;
  3. responsibility in penal law must be established with a very high degree of “rational credibility”, which is not coincident with statistical probability.

The aim of the paper is to give an evaluation of the mentioned points in the light of the philosophy of probability and of the recent developments of the logic of counterfactual conditionals.




Silvia Pellegrini
Behavioral Genetics in Criminal Trials: Where Do We Stand?

The recent introduction of behavioral genetics evidence in criminal trials has prompted a lively debate across the world concerning whether the genetic profile of an individual may affect ability to control one’s own behavior to a point to reduce responsibility in criminal acts. In my presentation, I will discuss results from scientific studies that provide evidence in support of the role of specific genetic variants in enhancing risk for the development of antisocial behavior. In the first place, none of these variants exerts any deterministic effect, but rather they act by modulating the impact of environmental factors on behavioral traits. As a matter of fact, recent studies indicate that these “risk-genes” may act more as “plasticity” rather than “vulnerability” genes. That is, these genetic variants would influence the individual susceptibility to both negative and positive environments. In light of these findings, the real issue is to what extent genetic profiles, by modulating the individual response to environmental factors, may ultimately affect their social behavior. A new perspective on the role of the environment on human behavior will be proposed, and the potentially relevant ethical and forensic implications will be examined.





Giuseppe Di Pellegrino
The Neural Basis of Moral Behavior




Andrea Cavanna
Voluntariness of Movement in Hyperkinetic Disorders




Giovanni Tuzet
Abduction and Causation

As an inferential tool for generating or suggesting explanations of given facts, abduction has a significant role in everyday life as well as in scientific practice and in more particular domains such as legal reasoning. In this domain abductive inferences aim at explaining some legally relevant facts, that is, the facts that generate a form of legal liability (civil or criminal). And more in particular legal abductions usually aim at reconstructing the relevant causal chains that led to those facts. To this purpose, one of the claims of this paper is that abduction works at its best when it is part of an “inference to best explanation” process where the selected explanation points at a causal difference in the relevant facts. Examples will be given of this. But sometimes that is not sufficient to establish legal liability, since we also need counterfactual reasoning to determine things like negligence.





Alberto Mura
Contingent Counterfactuals

It is argued that the logic of conditional is essentially epistemic. This  tenet, as maintained by R. Stalnaker, does not prevent using a possible-worlds semantics: it suffices, given a conditional of the form (A → B) to consider the set of “worlds” that are compatible with a given deductively closed stock of (bivalent) sentences whose truth is (albeit provisionally) taken for granted. This allows the characterisation of counterfactuals as conditionals with respect to a hypothetical (indeed counterfactual) stock K* obtained by contracting the current stock K so that K* := K-{A}. Indicative conditionals are simply conditionals with respect to the current stock (that does not contains A, so that K*K). A new compositional semantics for conditionals, based on de Finetti’s idea of tri-events, is adopted. It contains a logical consequence relation that encompasses Adams′s p-entailment and extends it to compounds of conditionals, avoiding Lewis’s triviality results. According to this theory (in contrast with Stalnaker′s and Lewis′s theories), those counterfactual conditionals that are neither analitycally true nor analytically false lack a truth-value at the actual world. However, they are essentially contingent, since there are K*-worlds at which they are true and K*-worlds at which they are false. The probability of countefactuals is defined along current lines in terms of the totality of K* worlds. No triviality result applies to them, since conditionals, meant as tri-events, are not bivalent propositions.





Paolo Francesco Piras
La causalità nella giurisprudenza

The Italian Supreme Court of Cassation intervened in 2002 (Cass. Pen., S.U., ud. 10.7.2002, dep. 11.9.2002, n. 30328, Franzese) on the way the causal relationship between conduct and event has to be assessed. The Court suggested a method based on two distinct assessments, made consecutively: the first assessment is made in terms of statistical probability, the second in terms of logical probability. The first assessment establishes what percentage the dutiful conduct prevents the event. For example, in case of placental detachment, timely Caesarean prevents fetal death in 84% of cases, as highlighted in a clear judgment (Cass. No. 32291-2009). By contrast, the second assessment considers the peculiarities of the single case under examination, in order to determine beyond a reasonable doubt whether it is included among those cases that succeed in preventing the event. In the judgment just mentioned, it was considered, by an assessment based on logical probability, that the case succeeded in preventing the event, due to severity of non-detachment, lack of suffering fetal heart at admission, fetus’s resistance to hypoxia for about six hours and the fact that the newborn lived nine days despite the delay in the practice of placental cutting. The Court concluded that if timely Caesarean had been made the death would not have occurred, so that there was a causal link between the omission of timely intervention and fetal death. However, since the method of assessment based on logical probability is not grounded on statistical matters of fact, does not guarantee a solid causal judgment, so that it is often only formally invoked in law, and hardly is at work in judicial decisions.





David Miller
Neither Pascal nor Bacon: Using Evidence More Justly

This lecture is part of two different research projects that have obligingly come together in the question of how the mathematical theory of probability bears, or should bear, on legal judgment.  The two projects are (a) to work out how the sceptical and anti-justificationist philosophy of critical rationalism can accommodate the predilection for proofs that is so evident in the law; and (b) to evaluate some alternatives to the conditional (or relative) probability  p(c | a) as a measure of the degree of deducibility of a conclusion c from an assumption (or premise) a.  Most of the lecture will be devoted to (b), to which its title obliquely alludes.
The function q(c | a) = p(a’ | c’) (the prime ‘ means ‘not’) is also a measure of the degree of deducibility of c from a. Its lower limit is not impossibility, as is p’s, but (maximal) independence, which obtains when a and c share no non-tautological consequences.  Unless a is itself a tautology, the value of q(c | a) is 0 when none of c’s content is included in a’s, and is 1 when it all is (that is, a implies c).  It will be argued that q does a better job than p in grading the relevance of evidence a to a conclusion c. This is especially so in the criminal law.