The Public Prosecutor: a Character in Search of an Author?

1. The last line, read first

“The judgment under appeal is quashed without remand because the act did not occur” (p. 41). It is worth starting from the end, because this is the formula that denies at the root. It does not say that the thing happened but the law does not punish it, as “the act does not constitute an offence” would; it does not say that a rule is missing, as “the act is not provided for by law as an offence” would; it does not send the papers back to another judge to write better reasoning, as quashing with remand would. It says that the act described by the rule — a mandatory act, the refusal to perform it, a refusal without justification — did not take place. The historical facts — the failure to acquire certain papers, the note of 5 March 2021 in which the two public prosecutors explained to their superiors why they were not acquiring them — nobody denies, and the Court says so (p. 32). What does not exist is the offence that two levels of judgment had built on top of those facts.

The judgment runs to forty-one pages, and its architecture is simple. The Ritenuto in fatto, the statement of facts, from p. 2 to p. 15, summarises the charge and the grounds of appeal: twenty-six grounds from the defendants (nineteen in the appeal, seven added later) and six from the Presidency of the Council of Ministers, which in Brescia had been ordered to pay damages as the party civilly liable. The Considerato in diritto, the reasoning in law, begins on that same p. 15 with four words, “The appeals are well founded”, and the part devoted to the offence ends on p. 38; then come two pages on who should answer civilly for an offence that does not exist, and the operative part. Between p. 15 and p. 38 the Court says three noes, one for each element of the offence under the first paragraph of Art. 328 of the Criminal Code: the act was not mandatory; there was no refusal; and even if there had been, it was not undue. It announces this on p. 16, where it writes that the three questions “must be assessed individually”, and it keeps its promise: it assesses them one by one, and closes all three in the same direction.

The people and the proceedings, for those who do not know the story, are few, and the judgment presents them in its first lines (p. 2). In Milan the Eni-Nigeria trial was under way, and the prosecution was conducted by two public prosecutors, De Pasquale and Spadaro. In the same office another public prosecutor, Storari, was co-holder, with the Deputy Chief Prosecutor Pedio, of a related proceeding, the one on the so-called “false Eni conspiracy”. In that proceeding Armanna had handed over his own telephone (p. 35); in the Eni-Nigeria trial — in the words of the charge, as the Court summarises it (p. 2) — he “had assumed the role of declarant against the principal defendants”. Storari passed to his two colleagues, first orally and then by emails of February 2021 (some forwarded by the Chief Public Prosecutor, some sent directly), information that the charge itself describes as “capable of demonstrating the falsity of the elements provided to the Prosecution by Armanna”, and urged them to bring a series of documents into the trial. The two public prosecutors did not do so, and on 5 March 2021 they explained why to the Chief Public Prosecutor, Greco, and to Pedio herself. From this, and from a video of July 2014 of which I shall speak at the seventh plank, arose the trial for refusal of official acts before the judges of Brescia, first the Brescia Court and then the Court of Appeal, which convicted them and which from now on I shall call, together, Brescia.

On p. 15, before the three noes, the Court delimits the subject of the proceedings with a sentence that must be read carefully, because it is easy to misread. “The subject of the proceedings”, it writes, “does not concern the assessment of the general propriety of the conduct of the two appellants or of the non-negotiable duty of loyalty and impartiality of the Public Prosecutor. The issue concerns criminal legality and the typicality of the offence”; and immediately afterwards: “Not every possible breach of the duty of propriety in itself constitutes an offence”. Anyone who stopped at this line might read into it a reservation about the propriety of the two prosecutors’ conduct: the Court would have acquitted two improper magistrates, because impropriety is not an offence. It is a not very honest reading, and the judgment does not allow it. On the merits of the choice, whether it was right or wrong not to acquire those papers, the Court abstains, and repeats that it abstains, because that is not the judgment that falls to it. On the way in which the two public prosecutors behaved, and on the law, the Court does not abstain at all: it deals with them in the pages in which, having to establish whether there was a refusal, it describes what they did. The decision to involve their superiors, which took the form of the note of 5 March, “represents an act not of refusal but of objective, dutiful transparency” (p. 33), and it will be read in full at the fifth plank. Their choice rested on “a conviction that was in fact shared both by the Chief Public Prosecutor and by the Deputy Chief Prosecutor” (p. 35). And the act they did not perform was not only not mandatory: as for the chats, performing it would have been contrary to the law, because their acquisition, the Court writes on p. 37, would have been “undue”.

The sentence on p. 15, then, concedes nothing. It says that the criminal judge is not called upon to hand out report cards on propriety, and then, where the conduct of the two magistrates enters the judgment, describes it in the words I have quoted.

If there is something contrary to the rules in the story the judgment tells, the Court places it precisely, and it is not an act performed or refused: it is an act unduly requested, not refused.

The acquisition of those chats, which have become iconic, of the bad faith of the “pentito” who was corrupting his witnesses to give credit to his version, had it been made in the OPL 245 trial, would have been “undue”, as their acquisition in the “conspiracy” file was “undue”.

Who requested it, and how, the judgment states by name and in two words: “those chats that Dr Storari was forcefully requesting” (p. 36). The reading is mine; the words are the Court’s.

The question of the origin of those items of evidence, of their probative value and in whose favour, and therefore of the realistic interpretation of the context in which the chats were extracted and used, is left in the shadow by the Court of Cassation, as it had to be.

The sentence on p. 15 shifts the judgment from the character of the two magistrates to the rule, and the rule is the point from which everything else is measured, the compass. Here the Court lets it be understood that the compass had gone haywire. Brescia had turned the needle: at the top, where north lies, it had placed the organ of justice, and from that idea it had derived, in order, an obligation, a refusal and an offence.

The Cassation puts the needle back in its place.

The trial is a trial of parties; the public prosecutor is one of them; his duties take the form of the rules of procedure, and the rules of procedure, after committal for trial, are written at the point that matters with the verb “may”: searching further is a faculty, filing what one has decided to search for is a duty, but it comes after the choice, not in place of the choice. The case in which an act imposes itself exists, but it is measured on the facts, and here the facts said that it did not exist.

This paper follows the judgment in the order of its pages, skipping what is not needed, to show two things.

The first: of the propositions of law with which Brescia, starting from Arts. 358 and 430 of the Code of Criminal Procedure, had built an obligation, then a refusal, then an offence, the Court saved nothing. The second: in which direction it reset the compass. I shall also say, out of fairness to the reader, what the Court kept, because it did keep something: and it is precisely what Brescia had used as a label, returned without the consequences Brescia had drawn from it.

And at the end I shall say what remains to be done. The Cassation has laid the first brick. To rebuild the edifice something more will be needed, the rational critical re-examination of a brief season of extraordinary injustice. Otherwise the wound will leave indelible traces.

2. The seven planks on which the convictions rested

To know what has fallen one must know what had been put up. The Brescia convictions rested on seven propositions of law, which I shall call planks because, laid side by side, they held up the floor on which the two Brescia judgments walked.

I draw them from the two judgments and, where the Cassation reports them, I give the page; the seventh the Cassation does not report, and it will be seen why it had no need to.

The first plank is the organ of justice. The public prosecutor, for Brescia, is always one, even at trial, because he is “appointed, in the interest of the community at large, to defend the legal order” (App., p. 128). From this it followed that Art. 358, the rule that requires him to establish, during the investigation, also what benefits the suspect, applied also during the trial, since the public prosecutor “always and in any case acts in the interest of justice” (words of the appeal judgment, which the grounds of appeal quote and the Cassation reports on p. 6); and that it applied in the name of equality of arms and the right of defence.

The second plank is Art. 430, the rule that allows the parties to investigate further after committal for trial. For Brescia it was the source of a duty and not of a power, and to trigger it the “pertinence” of the material “together with the possible interest for the other parties to the proceedings” was enough (App., p. 95). The third is the receiver: the public prosecutor who receives documents from a colleague is “obliged to file ‘automatically’ the documents received” (as p. 28 summarises the appeal judgment), without assessing them.

The fourth plank is the criterion by which the evidence was to be weighed: what matters is not that the evidence proves, that is, that it is evidence, but that it might be of use. “It is not a matter, in this context, of establishing whether the witnesses had actually been paid” (Trib., p. 103); the money had been received “perhaps even for purposes other than the corruption of witnesses, but which in any case deserved disclosure” (App., p. 109). The fifth is the refusal: not filing is refusing, even when one replies in writing to one’s superiors explaining why. The sixth is the container: the chats extracted from Armanna’s telephone were acquirable, because the seizure has as its object the “container”, that is, the telephone (App., p. 93, as the Cassation reports it on p. 37), and therefore the refusal was undue. The seventh concerns the video that entered the story undeservedly: the defence lawyers, some of whom had had it since 2018, do not count, because they “are not required to produce” (Trib., p. 114) and “the duty of discovery lies with the Public Prosecutor, not with the defence” (App., p. 126).

The Court sums up on p. 17 Brescia’s thesis on the mandatory nature of the act, and the very choice of words already says how it will end. For Brescia the filing was “a procedural act which, for systemic reasons linked to the purpose of the trial, to the sacredness of its fact-finding function, to the principle of equality of arms, to the adversarial principle, should have been performed in any event, in every case, without delay; a duty linked to principles so weighty that failure to observe it would have amounted to an undue refusal”. The source of this “self-evident mandatoriness” lay in Arts. 358 and 430 (p. 17). A Court which, summarising someone else’s thesis, chooses or leaves words such as “sacredness” and “self-evident mandatoriness” is not preparing to share it; and in any case it had already said so a page earlier, on p. 16: “a line of reasoning that in reality cannot be shared, for several reasons”.

3. First plank: Art. 358 lives in the investigation and serves the prosecution

To the first plank the Court devotes pages 18 to 26, and it is the finest part of the judgment, because it does not merely say that Brescia was wrong: it reconstructs where the error comes from, adopting, with the adverbs “correctly” and “lucidly”, a scholarly reading that it does not name (p. 18). The story begins with the enabling law of 1987, which had prescribed that the characteristics of the “accusatorial system” be implemented in the trial and had written, in directive no. 37, a “power-duty of the public prosecutor to carry out investigations for the purpose of exercising criminal prosecution and establishing specific facts, including the elements favourable to the accused” (p. 18). A single power, with exculpation inside it too: “including”. The Code translated that single power into two sentences side by side: the public prosecutor “carries out every necessary activity”, and “also carries out checks… in favour”. From that “juxtaposition” arose the idea that investigation in favour was a separate obligation, “detached from the decisions concerning the exercise of criminal prosecution” and supported by “other values and principles” (p. 18). It is Brescia’s idea, and the Court says so: “a construction that appears in some way to confirm the prosecution’s assumptions adopted by the trial judges” (p. 18).

Then the Court closes that door, with constitutional case law. Since 1991 the Constitutional Court has read Arts. 326 and 358 “in a unitary and not a disjoint manner” (p. 19): investigation in favour serves the completeness of investigations and the mandatory nature of prosecution, that is, it serves to curb practices of “apparent” prosecution on “investigations that are too superficial, incomplete or truncated” (p. 19). Order no. 96 of 1997 says it without circumlocution: the obligation under Art. 358 is “functional to a correct and rational exercise of criminal prosecution”, and is aimed neither at equality of arms nor at the right of defence. This is how the Cassation reports it, verbatim (p. 20). Judgment no. 58 of 2026, which the Cassation summarises over three pages, sets the seal: “mandatory” prosecution “is only prosecution exercised with prudence and responsibility, on the basis of complete investigations which have not neglected the evidence in favour of the accused” (p. 24). The Cassation draws from it that the duty under Art. 358 is “a duty that arises and develops within the phase of the preliminary investigation” (p. 25).

The conclusion is on p. 25, in four points that the Court marks with the letters a) to d), and that should be read calmly. Art. 358 “lays down a specific rule of conduct which… arises and develops within the phase of the preliminary investigation”. Its observance “is linked to the correct exercise of criminal prosecution”. Its breach “does not entail a specific procedural sanction”, but the risk of a prosecution “destined to be dismantled” and liability “on different levels”: civil, disciplinary, “criminal, while respecting, however, the typicality of the offences”. And the issue, “contrary to what is stated in the judgment under appeal, is not connected with the principle of equality of arms or with the right of defence” (p. 25).

Hence the verdict, on p. 26, with the adjectives the Court chooses. Brescia’s assumption “reveals its objective groundlessness”. Art. 358 “cannot constitute the source of an objectively different obligation”, that of a supplementary activity “mandatory, undifferentiated, all-encompassing, enforceable even during the trial hearing”. It is “a functional, unsystematic expansion, not permitted by the rule invoked, which has in reality been used improperly in order to make, distorting its function, a further procedural option binding and mandatory”. And the closing line: “The appellants… were by no means obliged, under that rule, to carry out supplementary investigative activity in that trial phase” (p. 26).

One thing must be noted separately, because it says how the Court understands the organ of justice that it nonetheless names. On p. 22 it writes that “a prosecution exercised without rigorous comparison with the elements in favour has greater diagnostic chances of being demolished before the judge, from its very birth”: the duty to look for exculpatory evidence is explained by the interest of the accuser in not losing. On p. 24, in the words of the Constitutional Court, it also explains it by the position of the accused, who only through complete investigations is put in a position to choose an alternative procedure, “an integral part of his right of defence”. The two explanations have one point in common, and the reading is mine: in both the duty lives in the investigation and serves the prosecution. It is the rule of a party that must accuse well. The organ of justice of judgment no. 88 of 1991, which the Court reports without reservation (p. 19), lives there, in the investigation, and serves that purpose. Of the first plank, Art. 358 as the source of an obligation at trial, nothing remains.

4. Second plank: Art. 430 is a “may”, and the second paragraph serves the first

The second plank falls in less than two pages (pp. 26-28), because it is enough to read the rule. After the order committing for trial “the parties may carry out ‘supplementary investigative activity’” (p. 26). Art. 430 “is structured in two paragraphs, the second of which is ancillary to the first” (p. 26). The first “confers on the public prosecutor a faculty - they ‘may’ carry out supplementary investigative activity - and, above all, establishes what the public prosecutor may not do” (p. 27). Supplementary activity, “since it is carried out, by definition, outside the hearing”, is “on the one hand, not enforceable and, on the other, not restrictable, because carrying it out is left to the exclusive choice of the person who holds the investigative power” (p. 27). It has a logical limit, “pertinence to the subject of the charge as outlined in the indictment” (p. 27), and it has a holder: “it is the public prosecutor who, in the exercise of his own autonomy and taking on the relevant responsibility, decides whether to carry out supplementary investigative activity” (p. 27).

On p. 28 the principle is written in the form of a rule. “The first paragraph of Art. 430 of the Code of Criminal Procedure confers on the public prosecutor a discretionary power and does not impose an obligation to carry out supplementary investigative activity such that his refusal would be relevant for the purposes of Art. 328, first paragraph”. “The mandatory nature of that activity does not derive from Art. 358 either”. The second paragraph “lays down a consequential rule: if the public prosecutor decides to carry out supplementary investigative activity, he must then, immediately, file the relevant documentation and notify the parties”. The two levels must be kept distinct: “what is mandatory is not, in fact, carrying out the supplementary investigative activity, but ‘only’ the conduct following the discretionary decision to carry out such supplementary activity” (p. 28).

Brescia had read the second paragraph without the first, and from the duty to file what one has decided to look for it had derived the duty to look. The Cassation puts the two paragraphs back in the order in which the legislature had written them; Brescia’s reference to Art. 430, it will say on p. 31, is “legally erroneous”. Of the second plank the verb remains: “may”.

5. Third plank: the public prosecutor is not an empty container

The third plank is the one on which all the rest rested, and the Court devotes to it its most severe pages (pp. 28-31). The problem is this: can supplementary activity be carried out by another public prosecutor, the holder of another proceeding, and is the holder of the trial obliged to pour its results into it? Brescia had answered yes, “automatically”, relying on a judgment of 2012, the Barbaro judgment. The Court: “This too is an erroneous legal approach” (p. 28).

The reasoning is in three steps. The first it takes in the words of a scholarly opinion that it declares correct and does not name: “it would be objectively unreasonable to hold that, with respect to acts and documents coming from sources external to the proceedings and to the trial - that is, from other judicial offices, from administrative offices, from associations, from individual citizens - the public prosecutor in charge does not hold that same power of discernment and assessment which indisputably belongs to him in the course of the investigation and is instead obliged, unconditionally, to ‘pour’ into the trial ad quem everything that is transmitted to him” (pp. 28-29). The second step identifies the subject: “the rules of conduct, impartiality, the need to assess with loyalty and to decide whether to carry out supplementary investigative activity can only refer to the public prosecutor who holds the proceeding, that is, to the magistrate who has responsibility for that proceeding” (p. 29). The third is the sentence that will remain: “the holder of the proceeding is not a mere recipient, an empty container that can be filled on the basis of the assessments of others, even where those are assessments made by the holder of a proceeding concerning connected offences” (p. 29).

As for the Barbaro judgment, the Court says at once that Brescia’s reliance on it is “objectively incongruous and erroneous” (p. 29), and then reports it in full (pp. 29-30), so that one may see what it was talking about. They are Barbaro’s words, not the Court’s, according to which the adverb “immediately” in the second paragraph “must be related to the concrete situation in which the Public Prosecutor finds himself operating”, who “may be in objective difficulty in immediately identifying the acts to be filed” (p. 30). Barbaro spoke of when to file what the public prosecutor has acquired, even if the material reaches him from another authority; it did not say that he was obliged to acquire it. The reading is mine, on the text the Court reports. What the Court draws from it lies in the four letters that follow.

The four letters on pp. 30-31 are the rule as it has always been, put back in writing. If the act comes from a different public prosecutor, the activity “becomes ‘supplementary’ only if there is, on the part of the public prosecutor who holds the proceeding, an act of reception, of acquisition of the material transmitted”. That act of reception “is not in turn automatic, but follows from an evaluative and selective activity”. The assessment has its limits, which are “pertinence”, “specific content”, the “usefulness of the elements that have emerged… with respect to the ‘framework’ constituted by the indictment”, and “the stage of development of the trial”. And at the end “the public prosecutor who holds the proceeding, at the outcome of that assessment, decides whether and what to acquire”. The colleague who transmits “merely transmits the acts and documents that he thinks may have value in the other proceeding” (p. 31): he thinks, he does not decide.

The consequence, on p. 31, is written with the colleague’s name. The act refused “was not an act that… ‘had to’ be performed for reasons of justice, that is, an act that could not but be performed by reason of the assessments of Dr Storari, because, in reality, it was a typically discretionary act which, as such, necessarily involved technical assessments on the part of the two appellants”. Assessments “of relevance, of pertinence, of genuineness, of consistency” (p. 31); I shall return at the end to one of these four words. Of the third plank there remains a container, empty, and the rule that no one can fill it from outside.

6. Fourth plank: possible usefulness and hindsight

The fourth plank was the criterion by which Brescia weighed the evidence: not that it proves, but that it might be of use. On this point the Court does two different things, and it is worth distinguishing them. It refrains, three times in explicit terms, from saying whether the choice of the two prosecutors not to acquire was correct: “correct or erroneous as it may have been” (p. 32), “rightly or wrongly” (p. 34), “in a way that may or may not be shared” (p. 35); and a fourth time, on p. 33, with a parenthetical clause, “beyond its intrinsic correctness”, which it refers to the decision to involve the heads of the office. The merits of that choice, it says, do not change the typicality of the offence. It is the abstention announced on p. 15, and it has a precise perimeter: it concerns the merits of the choice, not the way in which it was made nor the law. The note entered in the register remains an act of “dutiful” transparency (p. 33) and the acquisition requested remains “undue” (p. 37), as we read at the beginning and as we shall see shortly. On the weight of the material, on the other hand, the Court does rule; and it does so per relationem, that is, in the words of others: of those who had received that material first, the Chief Public Prosecutor and the Deputy Chief Prosecutor, and of the person who presided over the trial into which it was supposed to enter. The Cassation takes their statements from the appeal judgment.

The Chief Public Prosecutor, Greco, had received from Storari “a deluge of mile-long emails - put together - with a cut-and-paste technique from the files called falsità”, that is, in all likelihood, the texts in Word format attached to the emails that Storari had sent him between 14 January and 4 February 2021, as the appeal describes them (p. 4). He had found it “rather complicated and… practically impossible to read for someone who had little time available”. He had judged it “marginal stuff, not decisive”, still “in fieri” with respect to a trial which, in his judgment, was based mainly on documentary evidence. He had “passed it all on to De Pasquale”. He had received from the two public prosecutors the memo of 5 March, “containing a detailed examination of the situation”, and he had “entered it in the register”. He had considered that there were “legal problems” with the acquisition (p. 32).

The Deputy Chief Prosecutor, Pedio, was co-holder with Storari of the proceeding on the conspiracy. When Storari had proposed a precautionary measure in that proceeding, she, after examining the draft he had prepared, had given a negative opinion, “making use of the discretionary power” of her role, and had not changed her mind “not even reassessing it with hindsight” (p. 32). In the same matter she had herself been investigated, on the charge of not having “promptly” registered Armanna as a suspect for calumny “in order to avoid possible repercussions on the Eni - Nigeria trial”, and the case against her had been dismissed (p. 32).

The Presiding Judge of the Milan panel, that is, of the court trying the Eni-Nigeria case, had stated that Armanna’s credibility had already been undermined and compromised in the course of the trial, and that the documentation “proving the ‘fraudulent conduct of the declarant’”, if acquired, would only have “facilitated the effort put into the reasoning” (p. 34): he considered it probative, that is, and not decisive. From these assessments the Court draws the sum in one line: “That documentation which Dr Storari had pointed out was not self-evident and would not have changed the outcome of the trial” (p. 34).

There follows the statement of principle, which holds beyond this trial: “The typicality of the offence cannot change as a result of an indistinct posthumous assessment, which is affected by the level - greater or lesser - of the review that one wishes to exercise over the discretion of the appraisal entrusted to the public official” (p. 34). Brescia had judged in 2024 and in 2025 what two public prosecutors had decided in March 2021, and had judged it by the yardstick of someone who knows how it all ended. The Cassation says that the offence is measured at the moment of the conduct, by the yardstick of the person who decides, and not by that of the person who reads it again later. Brescia’s criterion was possible usefulness; the yardstick by which the Court measures, on p. 34 and on p. 37, is decisiveness. The reading is mine, the words are the Court’s. Of the fourth plank there remains hindsight, which is not a rule of law.

7. Fifth plank: the refusal that was not there

Refusal, for the purposes of Art. 328, is “a clear manifestation of the will not to perform the mandatory act” (p. 31), “a subjectively attributable omission, unequivocally referable to an act that must be performed without delay and that the agent unduly decides, instead, not to perform” (p. 34). It can also be implicit, and this “is not in dispute” (p. 16). What had happened here is not in dispute either: “It is not in dispute that the appellants not only replied to Dr Storari, but, on 5 March 2021, informed the Chief Public Prosecutor and the Deputy Chief Prosecutor - who was moreover co-holder of the connected proceeding - of their assessment - correct or erroneous as it may have been - that they did not consider it appropriate to proceed with the acquisition of those acts and documents, leaving, however, the decision to Dr Greco and Dr Pedio” (p. 32).

And the heads of the office? “Greco and Pedio had been informed and remained silent, raised no objection, asked for nothing, did not consider that the matter should be looked into further, called no meeting, did not ask for explanations either before or after” (p. 33). Their statements, “unless one wishes to hypothesise a general collusion to the detriment of Dr Storari and a generalised breach of the duties of loyalty and propriety in which - in some way - those who had responsibility for the Prosecutor’s Office would also have had a part, point in the opposite direction to the prosecution’s thesis accepted by the Court of Appeal” (p. 33). The Court does not need to say that there was no collusion: it says that to convict the two prosecutors one had to hypothesise it, and with it an entire office; and that on the thesis that they had decided “arbitrarily, and for instrumental purposes” “the judgment under appeal is essentially assertive and one-sided” (p. 33).

The note of 5 March, which Brescia had read against the two magistrates, as reinforcing the intent (this is how the appellants report it, p. 13), is for the Cassation the opposite of a refusal. The decision to involve the heads of the office “represents an act not of refusal but of objective, dutiful transparency; an act that reveals how the defendants did not act in a logic of hiding, of improbable concealment, of self-referentiality, but sought a physiological dialogue with the heads of their office” (p. 33). The conclusion: “the defendants, on closer inspection, refused nothing because they did not express a clear manifestation of the will not to perform the act, having left the decision to others who, in fact, shared it” (p. 34). Of the fifth plank there remains a note entered in the register, and two heads of office who never contradicted it.

8. Sixth plank: the undue, overturned

The adverb “unduly” has more than one reading, and the Court goes through them one after another (pp. 34-35). For one line of case law it is a requirement of special unlawfulness, which gives weight only to refusals that have no justification in the law. For another it is a superfluous element, which repeats the unlawfulness already written into the rule. In simpler words: either the adverb adds a requirement, or it adds nothing. The Cassation, for its part, had made it a feature of intent, that is, of what the agent knows, “the need for awareness of acting in breach of the duties imposed” (p. 35); and now, with an “in reality”, it brings it back to where it belongs, in the act itself: “it is an element that already characterises the objective typicality of the offence” (p. 35). However one reads it, the Court has “already said how the conduct attributed to the defendants is not arbitrary and is not accompanied by the awareness of omitting to perform an act that, instead, should have been performed” (p. 35). And it writes the line that marks the boundary: “Art. 328, first paragraph, of the Criminal Code does not make criminally punishable generic negligence, lack of caution or poor institutional sensitivity on the part of the public official, but the conscious refusal to perform an act that is objectively mandatory and that the public official knows he must adopt” (p. 35). It is a general line, written for the rule; in the case, the Court has just written that the conduct was not arbitrary (p. 35) and that there was no concealment (p. 33).

If one takes the first of the three readings, the one for which what matters is whether the refusal had a justification in the law, one reaches the point at which the judgment stops speaking of rules and recounts what had happened; it recounts it, it says, from the documents attached to the appeal and from its own judgment no. 22795 of 2022 (pp. 35-36). In the proceeding on the conspiracy the technical examination of the telephone handed over by Armanna concerned only the chats between him and two Eni executives; the lawyers for those executives had asked for access to the entire content of the telephone, and therefore also to the other chats, the same ones that Storari was asking to have poured into the Eni-Nigeria trial (p. 35). The Deputy Chief Prosecutor had said no; the judge had said no; and the Cassation, with judgment no. 22795 of 2022, had made the no final (pp. 35-36). The reason is the rule on evidentiary seizure, and the word “container” here refers to the device, not to the public prosecutor: “the object of the seizure is not the container of the sensitive data (the telephone), but rather only the data that are extracted from the mobile phone and that serve the function of establishing the offence being prosecuted” (p. 36). Not the device, not the forensic copy, “but only the data that are extracted from that copy because they are considered relevant for the purposes of proving the offence” (p. 36). “This, and only this, guarantees the proportionality of the measure” (p. 36).

Hence the consequence. “Those chats which Dr Storari was forcefully requesting to be acquired in the Eni - Nigeria proceeding could not be acquired, for the reasons indicated, in the Eni - Conspiracy proceeding” (p. 36); and their content “could not be poured automatically into the Eni - Nigeria proceeding either, because such acquisition would in any case have lacked a legal basis; that is, a new authorising order was needed” (p. 36). Brescia’s statement on the “container” “is seriously flawed” (p. 37). And the closing is the exact reversal of the convictions: the refusal “cannot be considered ‘undue’ because, in reality… that decision was in accordance with the principles and rules of law that govern evidentiary seizure. Conversely, the acquisition of the content of those chats, as requested instead by Dr Storari, would have been undue” (p. 37).

In this story the only undue act that the Cassation finds is the one that had been requested. Of the sixth plank there remains an adverb that has changed sides.

9. Seventh plank: the video that the defence lawyers already had

Point 5 of the charge concerned a video of 28 July 2014, which the judgment calls the “Bigotti video” and whose content it does not describe; I shall not describe it either, because for the Court’s reasoning it is enough to know that the Milan Prosecutor’s Office “would have had it available in April 2017” (p. 38). The Court establishes three undisputed facts: that the video “was relevant in several proceedings”; that “some defence lawyers were already aware of it in February 2018”; and that “it was then produced at trial, at the request of the defence, in July 2019, that is, well before the final deadline indicated in the charge ‘17 March 2021’, that is, two years before the decision in that trial” (p. 38). The charge, then, was one of delay.

The answer lies in a question. The Court asks it after setting aside two things: any procedural invalidity, “which, in reality, does not appear ever to have been raised in the course of the trial”, and the “further legal questions that might have suggested not proceeding immediately with the production of that video”. The question is this: “what is not clear is why, if that video had been so relevant, none of the defence lawyers, who had had that video available since February 2018, had requested its acquisition” (p. 38). The defence, the Presiding Judge of the Milan panel reported, expressed “a ‘veiled complaint and then… in closing argument an explicit one too’” (p. 38). And the closing: “it is not at all clear why there would have been an act of undue refusal, since neither the reason why that video should necessarily have been filed earlier nor why that acquisition had not been requested by the defence lawyers who had had that video available for some time has been made explicit” (p. 38).

Brescia had written that defence lawyers “are not required to produce” and that discovery is incumbent on the public prosecutor, “not on the defence”. The Cassation does not discuss the maxim, and for this reason had no need to report it; it asks a question about the defence lawyers: why, if the video was “so relevant”, they did not ask for it themselves, when they had had it for a year and a half. The rest is my reading: in a trial of parties, whoever has a piece of evidence and judges it decisive produces it; whoever keeps it in a drawer for a year and a half and then complains that the other party did not file it owes the Court’s question an answer. Of the seventh plank there remains the maxim, which the Court did not touch, stripped of the criminal consequence that Brescia had drawn from it; and there remains a question that the Court asked and that has received no answer.

10. The reckoning: what remains of the convictions, and what the Court kept

One can now do the reckoning, and one can do it with the Court’s adjectives, because they are its own and not mine, save one that I declare. Brescia’s reasoning “cannot be shared, for several reasons” (p. 16). The use of Art. 358 “reveals its objective groundlessness” and is “a functional, unsystematic expansion, not permitted by the rule invoked” (p. 26). The automatic obligation to pour material into the trial is “an erroneous legal approach” (p. 28); the reliance on Barbaro is “objectively incongruous and erroneous” (p. 29); the reference to Art. 430 is “legally erroneous” (p. 31). The criterion by which the evidence had been weighed falls under the statement of principle on the “indistinct posthumous assessment” (p. 34), which the Court writes in general terms and which, as I said at the fourth plank, I read as referring to Brescia. On the thesis that the two public prosecutors had decided “arbitrarily, and for instrumental purposes”, the judgment is “essentially assertive and one-sided” (p. 33). The statement on the container is “seriously flawed” (p. 37). On the video, “it is not at all clear” (p. 38).

The summary on p. 37 lines up the three noes with the terseness of an operative part. The act “was not a mandatory act, that is, an obligatory one, either under Art 358 of the Code of Criminal Procedure, or under Art. 430”. “It was not an act that, for reasons of justice, necessarily and automatically had to be performed without delay”. “It was an act that involved delicate discretionary assessments, proper to the public prosecutor who holds the trial”. Supplementary activity “is not enforceable and not restrictable and concerns the public prosecutor who holds the proceeding and cannot technically be carried out by a different magistrate of the public prosecution service - natural person”, who “may certainly point out” but “without being able to exclude, on the part of the public prosecutor who holds the proceeding, the existence of an autonomous power of assessment”. “There was no criminally relevant act of refusal”. And “the failure to acquire the chats pointed out by Dr Storari not only did not concern decisive exculpatory evidence, but was not undue”.

These conclusions are “absorbing on the logical and legal level” (p. 37), and the Court says so to explain why it does not examine the rest: the concrete availability of the acts, intent, the unfinished technical examination, the time limits set by the Milan Court (p. 38). Other grounds of appeal remain unexamined, and not even named, which the register at the end lists: there was nothing left to judge. Two of the sources written into the charge, Art. 73 of the Law on the Judicial System and Art. 111 of the Constitution, are not examined in their own right. The second the Court names on p. 16, as the third rule around which Brescia’s reasoning turned, and sweeps it away together with that reasoning; when on p. 25 it denies the link with equality of arms and the right of defence it does not name it, and the connection is mine. On the first it is silent. They have not been saved and they have not been demolished in their own right; once the two rules that were supposed to fill them had fallen, there was nothing left to examine: this too is my reading. The civil-law rulings fall (p. 38). And even the party ordered to pay falls: for an offence committed by a magistrate in the exercise of his functions the party civilly liable is the Minister of Justice, not the Presidency of the Council (p. 40), so that whoever had been ordered in Brescia to pay damages should not even have been in the courtroom.

The reckoning is this: seven planks; six rejected, with the words we have read; of the seventh the maxim has remained, the offence has fallen and the Court’s question has remained open. Of the propositions with which Brescia had built the obligation, the refusal and the offence not one has been left standing, and the Court did not even send the papers back to another judge so that he might try to rebuild them. It remains to say, as promised, what the Court kept; and it is not little. It kept the “non-negotiable duty of loyalty and impartiality” (p. 15). It kept the formula of the Constitutional Court of 1991, the organ of justice “obliged to seek all the evidence relevant to a just decision” (p. 19), placing it, however, in the investigation and tying it to the prosecution. It kept, on p. 27, the “general duties of impartiality and loyalty” and the rules that derive from them, and it named them, “acquisitions in favour of the person under investigation, filing of the relevant acts”: duties which, it says, accompany the public party from the start of the investigation to the end of the trial. It kept the category of discretionary acts that may appear necessary “by reason of the progressive development of the proceedings, of their outcomes, of the pursuit of the general fact-finding aims of the criminal trial” (p. 17). And it kept a “need for coordination of investigations, for transparency and for timely information to those who have responsibility for the office” (p. 29), which on p. 33 it calls “dutiful” transparency.

This list must be looked at closely, because it is Brescia’s label, the organ of justice with its duties, given back whole and stripped of what Brescia had derived from it: enforceability, automatism, criminal sanction. The duties remain, and take shape in the rules of procedure, which are written, as we have seen, under the sign of discretion. The duty to file remains; the duty to file what someone else has decided does not. The category of acts that may become necessary remains; that this was one of them does not. Coordination remains; and it had been done, with a note entered in the register. And there is one last thing, which closes the circle opened on p. 15. The Court does not hand out report cards, and it has said so; but of what the two prosecutors did it writes three things, the same ones I reported at the beginning: a “dutiful” transparency (p. 33), a conviction “in fact shared” by those who headed the office (p. 35), a decision, for the chats, “in accordance with the principles and rules of law” (p. 37). That those duties, applied to the two magistrates, turn out to have been observed is my reading of these three lines; the lines are the Court’s. It is not a judgment that corrects a conviction. It is a judgment that dismantles the mechanism by which, from a true duty, an offence had been manufactured, and that describes the conduct of those who had that duty with those three words.

11. The compass: a trial of parties

Dismantling is not enough, and the Court does not stop at dismantling. In the same pages it writes, with its own case law and with that of the Constitutional Court, where north lies. It writes it in six propositions, and it is worth isolating them one by one, because they are the rule that had been in the record for thirty-eight years and that two levels of judgment had read otherwise.

The first concerns the model, and the Court states it more with categories than with a name. It recalls that the trial is the one intended by the enabling law of 1987, with the characteristics of the “accusatorial system” (p. 18); it warns that that framework “profoundly changed in the following years”, with the principle of non-dispersal of evidence and with the three constitutional judgments of 1992, and that Art. 358 instead “remained unchanged” (p. 19), precisely the rule on which Brescia had built everything. The adjective accusatorial, in the sense of a model, the Court writes only once, on p. 18, and in quotation marks, quoting the enabling law; afterwards, the word recurs only in the sense of the prosecution’s thesis. In its own voice it uses “public party” and “role as a party” (p. 27), and takes up from its own case law “sui generis party” (p. 20): these are the words that count. Calling the resulting model accusatorial is my own choice of word: a trial of parties, in which the public prosecutor is one of them.

The second concerns the organ of justice, and the Court states it in the words of its own case law: the public prosecutor exercises the faculty of establishing facts in favour of the accused “also as an organ of justice, that is, as a sui generis party” (p. 20). The “also” and the “that is”, read together, say, in my view, that the organ of justice is a way of being of the party, not a nature that replaces it.

The third concerns the phases. At trial, the Court writes, “his role as a party is accentuated” (p. 27). It adopts, “in a way that may be shared”, the observation that impartiality “takes different forms depending on the phases”, and that the public prosecutor is “more impartial” when, in the investigation, he is more “alone”, that is, acts without an opponent before him (p. 27). But immediately afterwards it makes clear that the phase changes the manner, not the subject: the general duties and the rules that derive from them are not “limited and confined to specific phases of the proceedings or of the trial, because, instead, they always characterise the action of the public party, from the moment the investigation begins until the conclusion of the trial” (p. 27). The needle does not change its north with the phase.

The fourth is the most important. Those duties “are filled with specific content through the rules of procedure which, however, are inspired by the discretion, the non-restrictability and the non-enforceability of the action of the public prosecutor” (pp. 27-28). The duties of loyalty and impartiality exist, and the Court calls them “absolute” (p. 27); but they do not have the form of a virtue from which a judge can derive unwritten obligations. They have the form of the rules of procedure, and we have seen how those are written.

The fifth concerns discretion, which is responsibility and not arbitrariness. The public prosecutor decides, again p. 27, “in the exercise of his own autonomy and taking on the relevant responsibility”; the rules of conduct “are not automatic and unconditional but find their inevitable shape, case by case, in the light of the considered assessments that the magistrate in charge makes, in the exercise of his discretionary choices, and of the legibility of his considerations” (p. 29). Legibility: the duty is to explain, not to obey.

The sixth concerns sanctions. The errors of the public prosecutor in applying Art. 358 have their sanctions, and they are those that the law provides and that we read on p. 25: civil, disciplinary, “criminal, while respecting, however, the typicality of the offences”. The code of ethics too, whose Art. 13 the Court quotes, says that the public prosecutor “does not conceal from the judge the existence of facts in favour of the suspect or the accused” (p. 25). It is a duty, it has a name, and it has its judges, civil, disciplinary, criminal; the last, only where the offence as defined provides for it.

Lined up, the six propositions say one thing only, and they say it in the opposite direction to Brescia’s: none of the duties of the public prosecutor turns the party into an organ that must pour into the trial what others have decided for it. This is north. The Court did not invent it: it read it again.

12. The act that imposes itself, and why it was not there

There remains the question that every reader asks: if filing is a power, can a public prosecutor keep proof of innocence in a drawer? The appellants had raised it, because Brescia had used it against them, and had answered that in that case the correct exercise of discretion leads to filing, by acquiring the evidence, as the first paragraph of Art. 430 allows, and by notifying the parties, as the second requires (p. 5). The Court does not address the textbook hypothesis, and this must be said. What the judgment offers is not an answer to that question, but the yardstick by which, should the case arise, it would have to be measured; and the yardstick lies in three places, which must be read for what they are.

The first is on p. 17. The mandatory nature of the act, for the purposes of Art. 328, “exists not only in cases where, as in the one under examination, there is a subjective prompting that takes the form of a request in respect of which the conduct of the active subject constitutes an undue negative response, but also in those where, independently of the existence of a request or an order, there is a substantive urgency, which requires the act to be performed, made evident by facts so objective that the subjective inertia of the agent takes on the value of a refusal”. The Court places this case in the first hypothesis, that of the request; and for a request the negative response must be “undue”, that is, there must be a due act that is denied. The second hypothesis, substantive urgency “made evident by facts so objective”, is the act that imposes itself without anyone asking for it: the Court states it as a general rule, and has not needed to apply it here. Procedural acts “which are discretionary and which, nonetheless, appear necessary” for the “general fact-finding aims of the criminal trial” remain, it writes, “discretionary in nature”, and the law provides for them “not procedural invalidity but forms of control” over the efficiency, the legality and the responsibility of the public prosecutor (p. 17).

The second place is p. 29, which we have just read: the rules of conduct take shape “case by case”, in the light of the assessments of the magistrate in charge. The case by case belongs to the public prosecutor, not to the judge who tries him: it is he who, faced with what reaches him, must carry out the four assessments of p. 31, and must do so legibly. The third place is on p. 34. There the Court, from the assessments reported in the judgment under appeal, concludes that the act “did not concern evidence of self-evident decisiveness”; from this conclusion I derive, a contrario, the threshold: self-evident decisiveness.

I put the three places together, and the synthesis is mine. The act imposes itself when it is due on account of objective facts, when it is decisive in a self-evident way, and when it can lawfully be performed; and its first judge is the public prosecutor who assesses it, under his own responsibility and with the obligation to explain himself.

Here, on a reading of the judgment, the second and third conditions were lacking, and the first does not appear to have been measured by the Court in its own right. Not decisiveness: the material was not “self-evident and would not have changed the outcome of the trial” (p. 34). Not legality, at least for the chats: they could not be acquired without “a new authorising order” (p. 36), and acquiring them would have been “undue” (p. 37). As for objectivity, the judgment does not measure it in its own right: it reports that the person who received the material first judged it “marginal stuff, not decisive”, “in fieri” (p. 32), and that none of the heads of the office contradicted him (p. 33): these are statements, and on objectivity there is nothing else in the judgment. For the video, finally, one cannot see where the urgency of p. 17 could have come from, if those who wanted it had already had it since February 2018 (p. 38); and this too is my reading. The judgment offers no shelter to anyone who keeps proof of innocence in a drawer: it leaves before him the non-negotiable duty of p. 15, the code of ethics of p. 25, the substantive urgency of p. 17 and, while respecting the typicality of the offence, Art. 328. And to the judge who may one day have to measure that case it leaves the three noes of this judgment, read in reverse: that the act be due on account of facts and not of a colleague’s assessment; that it be refused and not assessed; that it be undue and not in accordance with the law. Brescia had convicted without any of the three.

13. The needle

I return to the compass. For two levels of judgment the needle had pointed to a north that is not in the code: an organ of justice that at trial pours in what is sent to it, because justice wants it so, without the right to assess, and without its assessment, shared by the heads of the office, counting for anything. The Cassation took the needle and put it back where the enabling law of 1987, the code of 1988 and the Constitutional Court of 1991, of 1997 and of 2026 had left it: the trial is a trial of parties; the public prosecutor is a party, as sui generis as one likes, but a party; his duties are written rules, with an object, a time and a sanction; discretion is responsibility, and it is explained; the act that imposes itself exists, and is measured on the facts. Here the facts said that it was not there.

Of everything Brescia had written, the Court kept the duties and threw away the offence; and of the conduct of the two public prosecutors it left the three words we have read: transparent, shared and, for the chats, in accordance with the law. Who was correct and who was not the Court does not write, and it says so (p. 15). It writes of an act performed that was transparent and shared, and of an act requested “forcefully” that would have been undue. The rest, as I said at the beginning, is my reading.

14. What remains to be done

An exegesis says what a judgment has decided and in what words; it does not say everything there is to know. The Cassation judged an offence, and to judge it the law was enough: the act was not due, it was not refused, it would not have been undue. But beneath the offence that was not there lie facts that were there. Brescia weighed them by the criterion of possible usefulness, which does not ask what they prove; the Cassation, which was judging the law, reported their weight in the words of others and by the yardstick of decisiveness (pp. 34, 37), without having to read them in its own right. No one, at the three levels, has yet judged them for what they prove, for their origin and for their genuineness. The next bricks are laid there: on the facts as they happened, before the facts as the trials have told them.

The questions are four, and they must be asked in this order. Who is at the origin of the evidence that was supposed to be poured into the Eni-Nigeria trial, and of the video that was supposed to have been filed earlier? Is the origin legitimate? Was the use made of it in the trials, in the Milan one and in the Brescia one, lawful? And how much was that evidence worth, had someone read it in full instead of weighing it by the criterion of possible usefulness? The second question is the one that the Court, on p. 31, calls “genuineness”, and which it wrote without emphasis; the present writer, who is a civil party in the trial on the conspiracy, cannot write it without, and declares this so that the reader may take it into account.

On the last question the Brescia appeal judgment offers the best example, and it is the one I placed at the fourth plank. In the Eni-Nigeria trial Armanna was a declarant against the defendants, and his credibility was at stake; evidence that he had paid witnesses would have weighed against him, and therefore in favour of the defendants. This, for the Brescia prosecution, was the exculpatory evidence that the two prosecutors were supposed to pour into the trial: the chats of December 2019, extracted from his telephone, would have shown, according to the Brescia charge and the two judgments, that Armanna had paid two Nigerian witnesses (Trib., pp. 102-103; App., p. 109). The Court of Appeal wrote that the money had been received “perhaps even for purposes other than the corruption of witnesses, but which in any case deserved disclosure” (App., p. 109). The Brescia Court, before it, had had the fact before its eyes: that money was linked to an “EFCC source file” (Trib., p. 102), that is, to a document coming from the Nigerian authority that investigates financial crimes; and it had chosen to draw nothing from it, writing that “it is not a matter, in this context, of establishing whether the witnesses had actually been paid” (Trib., p. 103). Here is the criterion of possible usefulness applied to a piece of evidence: one does not establish what it proves, one establishes that it might be of use. The present writer maintains, and will show in the next brick, the one on the records of interviews, that to understand what those chats were about it was enough to read the interviews given by Armanna in 2016, which had been in the case files for years: they were talking about that file, not about witnesses. If so, the “exculpatory” evidence that two magistrates were supposed to pour into a trial, under the threat of an offence, said something else; and its value could have been measured from the outset.

Perhaps less was needed: to read the record of the witness Eke at the hearing of 29 January 2020 in the OPL 245 trial: if this was a witness bought by Armanna, the least that can be said is that someone else bought him at a much higher price, of what nature I do not know.

The same four questions apply to the other chats, to the video and to the documents that the following bricks will present; and the answer, in each case, must be sought in the documents and the dates, not in the titles of those who handled them, which is the error from which this piece started. It is the work of the bricks to come: the genesis of the evidence, its passage from one office to another, its fate in the trials, its value. This first brick was meant to clear the ground of the law that the Court has rejected. The others are laid on top of it.²

Get each part by email as well, so it does not depend on any social network: whitecollarcrimes.it

Luca Santa Maria
whitecollarcrimes · 7 October 2026