Why do some courts (like those in America) decide through the dictum "guilty"(g) or "not guilty"(~g) instead of using the term "innocent"(i) for "not guilty"(~g)?

I think the polar opposites g/~g seem more suitable, because:

  1. Answers directly the prosecutor's question and faces it correctly as a completely ignorant question;
  2. It avoids any involvement with the problem of giving a truth value to the identity "~g=i";
  3. It takes into account that a trial's decision is a pro vero statement, and that court does not claim itself as a ruler of truth.
  • 52
    Because "not guilty" does not mean innocence, it means acquittal, i.e. that the proof of guilt was insufficient beyond reasonable doubt. Trials do not establish innocence because it is presumed, see presumption of innocence.
    – Conifold
    Commented May 19 at 11:31
  • 22
    Innocence is a matter of fact, not guilty is a matter of proving in a court of law. The difference between a false claim and an unproven claim.
    – Conifold
    Commented May 19 at 12:33
  • 9
    Think of it as proven guilty (beyond reasonable doubt) and not proven guilty (possibly innocent; possibly reasonabledoubt).
    – TripeHound
    Commented May 19 at 21:51
  • 5
    Not guilty in law. In England & Wales, a suspect does not have to prove their innocence: the prosecution must prove their guilt, and if they cannot do that the suspect is not guilty. In Scottish law there is a third verdict: not proven. Commented May 19 at 22:09
  • 6
    "not guilty" means "we could not prove that he committed the crime", which is certainly not the same thing as "we are certain he did not commit this crime". Commented May 20 at 1:45

7 Answers 7


It is a matter of convention. In many legal systems, the threshold of proof for a guilty verdict is set very high, so if the prosecution fails to attain it then the jury must delivery a not-guilty verdict even if, say, they think it is more likely than not that the defendant was guilty. Not guilty can therefore mean 'possibly guilty- we can't tell for sure', which is not the same as innocent. Someone can be not-guilty and innocent, but the two terms are not synonymous.

  • 1
    You can define anything to mean anything. In this answer you state that "not guilty" means "possibly guilty- we can't tell for sure" in a legal context, but it doesn't mean that in a literal context, so it's a re-definition of the phrase in a legal context. One could just as easily define "innocent" to mean "possibly guilty- we can't tell for sure". I don't think you have a convincing argument here. Commented May 21 at 13:56
  • 3
    @AykhanHagverdili He's not making an argument, he's just stating the plain fact that the reason court systems use "guilty" and "not guilty" is because of convention, not for a deeper philosophical reason.
    – Idran
    Commented May 21 at 14:17

Because the defence does not need to demonstrate innocence; whilst the prosecution must demonstrate guilt beyond a reasonable doubt.


One of the difficulties with the logic here is the definitions of "guilty" and "innocent" you're using. There's actually a number of different concepts entailed within those terms, and it's rather easy to equivocate and mix things up.

The two main underlying concepts within "guilty" & "innocent" are A) did the accused in fact commit the crime and B) was the accused found culpable of the crime by the legal system. The difficulty is that while "is found legally culpable" is easy to check, "did in fact do it" is in general hard if not impossible for a third party to verify. As such, the two are often conflated, with culpability being a proxy for "did in fact do it". However, all four combinations are possible in practice.

The quadrant of "culpable & did it" is normally what one thinks of when thinking of "guilty". Also, most everyone would agree that "not culpable & didn't do it" would count as "innocent".

Where to put the two "mixed" combinations is why it's difficult to take "innocent" as a direct equivalent to "not guilty". A third party can't know for certain whether someone did in fact commit the crime, so all they have is the culpability. So it's potentially expeditious to equate "guilty" and "innocent" to "culpable" and "not culpable", respectively.

However, the conventional meaning of "innocent" is usually taken to mean more "didn't do it". So it feels like a misnomer to lump the "not culpable but did actually commit the crime" subset into the "innocent" category. As such, making "not culpable" equate to "not guilty" instead is more descriptive.

But wait, what about "didn't do it, but found culpable anyway" subset? In post-enlightenment (Western) legal systems, this is considered to be a severe miscarriage of justice, and an attempt is made to strenuously avoid it. As such, it's normally assumed that this combination rarely happens. (It's also assumed that if anyone was proven to be in this subset, their culpability status would be rapidly reassessed.) Given the difficulty of assessing commission-in-fact, if the accused is found legally culpable of the crime, it's taken that they almost certainly did in fact commit the crime. So for the small number of people who do fall into this category, you end up in a strange situation where the legal system considers them "guilty" (i.e. in the "culpable & did it" quadrant), but their advocates consider them "innocent" (in the "did not in fact commit the crime" sense, regardless of their culpability status). It's less a dispute about what to call this quadrant, and more a disagreement about whether it actually applies to any particular person.


In the American judicial system, defendants are assumed to be innocent, and the prosecution must prove that they are guilty. There are philosophical rationales for that, of course, e.g.:

  • Logically speaking, it's difficult to prove a negative: in this case, to prove that one did not do a specific crime
  • Defendants rarely have the resources or skillset needed to investigate the crime they've been accused of and exonerate themselves
  • The law would risk de-evolving into a tool of harassment and vengeance, where people could be accused of crimes simply to force them through the time, expense, and exposure of showing themselves innocent

So when a judge or jury returns a verdict, the actual question they are answering is: "Has the defendant been proven guilty beyond a reasonable dounbt?". The answers they can give are:

  • Proven guilty
  • Not proven guilty

which is shortened to:

  • Guilty
  • Not guilty

In essence, the verdict says nothing whatsoever about the defendant, who may or may not have committed the crime regardless of the verdict. The verdict is a judgement on the prosecution's case: on whether the prosecution effectively proved its case before the court.

  • 1
    +1 "The law would risk de-evolving into a tool of harassment and vengeance, where people could be accused of crimes simply to force them through the time, expense, and exposure of showing themselves innocent". Due process only discourages egregious cases. Legal history is rife with examples that a clever prima facie accusation achieves the same effect.
    – J D
    Commented May 20 at 15:13
  • Nice approach, but I do see an ocean between "not proven guilty" and "(proven) not guilty". Also I do not find any special logical difficulty to prove a negative -e.g. reductio ad absurdum is a common method fo that. I am still all nice attempts on this page and I will be back! : )
    – SK_
    Commented May 21 at 11:59
  • @SK_: It's not "(proven) not guilty"; it's "not (proven) guilty". The confusion is understandable — and widespread — because people tend to focus more on persons than on arguments. But in court what matters is the argument (the proving) not the incident (the guilt). Commented May 21 at 16:44
  • @SK_: As far as proving a negative goes… We cn prove a lot of things in pure logic, but pure logic is a strictly limited domain that has little applicability in real life. Specifically, pure logic has no temporal component, and time really mucks things up. For example, say your SO accuses you of having an affair. It's easy enough to prove you DID (if you did) because evidence exists of the act (which any seedy PI can dig up). But how do you prove you DIDN'T? Can you account for every moment you are not in your SO's presence? Can you prove you didn't sneak out while s/he was asleep? Commented May 21 at 16:54
  • 1
    @SK_: There is no definitive evidence of non-action, except to except to account for every action you took at every possible moment of every day you might have had an affair. Could you (or anyone) do that for even a few days past? Commented May 21 at 17:00

The defendant is accused of committing a crime. The defendant can be innocent and yet held guilty in court. The defendant can be held not guilty and yet be not innocent.


The legal system here is somewhat similar to the reasoning behind a hypothesis test in statistics.

We have a null hypothesis (the defendant is innocent). We then set an decision threshold. In the case of statistics, this is often p<0.05, in law its "beyond reasonable doubt". We then test the hypothesis against the evidence and the threshold. If the evidence meets the threshold we reject the null hypothesis. If the evidence isn't sufficient, we don't reject the hypothesis. But not rejecting the hypothesis doesn't mean it is true.

If p=0.06, its still the case that its is very unlikely that the we would see this evidence in a world where the null hypothesis was true, but it doesn't quite make our threshold. Similarly, the evidence could suggest a defendant was probably guilty. But probably guilty isn't enough, it has be beyond reasonable doubt, so we can't find the defendant guilty. However, it would be perverse to say that a defendant was innocent, if the evidence added up to them "probably" being guilty.


In short: "innocent" is a claim you need to prove just like "guilty".

the 'non guilty' verdict means as much as "the evidence is inconclusive" or "we can't prove he defendant is guilty"

to be judged 'innocent', a lot more work is needed. this sometimes happens, and innocent always means "not guilty", but the inverse isn't true by definition.

if it very much possible for someone to be judged 'not guilty', but when new evidence comes up, there is now enough to claim the defendant 'guilty' at a second pass.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .