In the context of the whole "guilters vs innocenti" separation, I would make this point: I don't actually think that it's at all unreasonable to come at this case (if one is a so-called neutral and disinterested bystander) from a starting position of assuming the guilt beyond a reasonable doubt of AK, RS and RG. This is a logical position based on their conviction in a criminal court, especially if one is asked to make an "either/or" choice.
However, it's also a logically indeterminate position to claim that SINCE a person was convicted of a crime, they SHOULD have been convicted of this crime. Granted, in the majority of cases this position in fact does hold strong - since most convictions are indeed sound. But one can't use a QED argument to automatically link the two.
If all people who were convicted in a criminal court, based on the weight of evidence presented in court (and the way in which that evidence was presented) were automatically presumed to have a sound conviction, then there would be no point in an appeals system (that goes without saying of course). And even for cases that go through numerous appeals, it still doesn't automatically mean that the convictions are inviolably safe and just.
The situation is further clouded by the very real issue of reactionary behaviour within the judiciary (and that applies to pretty much any country). By this I mean that appellate courts are instinctively inclined to "side" with rulings made by their lower court colleagues - since to overrule them clearly implies a failing at some level withing the justice system as a whole (even though, paradoxically, correct appeals rulings are obviously a key facet of a well-functioning system). The failures of appellate courts to overturn or refer lower court verdicts despite clear indications to the contrary have been well-documented over many years, resulting in many wrongly-convicted people spending far more additional time in prison than they ought to have done.
Another related point that I've made before, but which bears repeating, is that a conviction can be unsafe NOT ONLY if the person is actually innocent of the crime, BUT ALSO if the person actually did commit the crime, but in the absence of sufficient evidence to convict beyond a reasonable doubt. This latter (and often misunderstood) category has probably led to the greater number of miscarriages of justice. This is often because it's not uncommon for the suspect to "fit the bill"; for (s)he to have no alibi; for a certain amount of (sometimes conflicting and often weak) identification evidence to exist; for a certain (often small) amount of forensic evidence to be presented linking the suspect to the crime; and for there to be no other credible potential suspects on the horizon. In these sorts of circumstances, juries can often be persuaded to convict based on their "gut feelings" - feelings which are actually often correct, but which in law are not an acceptable reason to convict in the absence of sufficient hard evidence.
The famous US District Attorney (ie state prosecutor) Vince Bugliosi likens circumstantial evidence (ie all evidence except confessions in trial or eyewitnesses to the crime being committed) as the strands of a metaphorical rope which connects the suspect to a jury. At the start of the trial, there is no rope at all. Then, as the trial proceeds, the prosecution adds strands to the rope, in the form of pieces of evidence. Each strand can be either relatively thick (e.g. victim's blood all over the suspect's clothes, when the suspect claims no contact with the victim), or thinner (eg questionable identification of the suspect a mile away from the crime scene). The defence try to discredit each strand, and the jury ultimately decide which of the strands remain and which are removed. At the verdict, the jury tug on this metaphorical rope. If the rope is thick and strong, it will hold firm and the suspect can be judged guilty. However if there are only a few thinner strands on the rope (or perhaps only one thicker strand), the rope will break, signifying reasonable doubt. It's up to the trial judge to advise the jury what the breaking point of this metaphorical rope should be - in other words to guide the jury in the area of "reasonable doubt".
So, if we extend the analogy, an case that goes to appeal starts with the thickness of rope that convicted the person. At this point, one of two things can happen. The first (and less likely) thing is that the appeal court decides that the breaking point of the rope was incorrectly set (ie that the lower court judge misdirected the jury). The appeal court can rule that had the correct breaking point been set, the rope would in fact have snapped.
The second, and more common thing that can happen in appeal is that the composition of the rope can change. Some strands may be removed (ie pieces of evidence may be successfully challenged or refuted with counter-evidence), and of course some might even be added. After the composition of the rope is changed, the appeal judges pull the rope again themselves, to see whether it now breaks. If not, then the conviction is still safe. But if it does break, time for an acquittal or a retrial.
Here endeth my bedtime story for today. I hope you're all sound asleep by now
EDIT: I realise that references to the word "jury" are strictly incorrect when applied in Italian criminal cases. However, the principles are still exactly the same (although I suspect the trial judge would not really need to issue much instruction to the judicial panel).
