Aristotle wants to demonstrate that virtue and vice are appropriate objects of praise and blame respectively.
The link to praise and blame is useful for Aristotle because he frequently appeals to endoxa about what we’d praise/blame, to back up his claims about the virtuous person.
Making this link requires showing that we have responsibility and agency over virtue & vice.
Which in turn brings us to voluntariness: as Watson (1982) puts it, S’s freely doing φ is necessary for S to be morally responsible for φ.
Aristotle would reject the consequentialist decoupling of responsibility and praise/blameworthiness.
According to him, laws are only effective at incentivising good behaviour when they reward or punish people for actions that they do in fact bear responsibility for. (III.5, 1113b21–30)
An analysis of voluntariness is also relevant to grasping how we should judge a person’s character based on their actions (Sauvé Meyer 2006).
Although prohairesis is a better indicator of character than actions are (III.2 1111b4-6) – since the same action could result from very many motivations – we can still get some signal about character from actions (and they’re far easier to observe).
But not all actions are relevant to assessing character – only voluntary ones.
Aristotle’s position on the precise connection to praise and blame is muddy. He seems to think voluntariness is necessary, yet not sufficient for praise/blame.
“[feelings and actions] receive praise or blame if they are voluntary, but pardon, sometimes even pity, if they are involuntary”
Aristotle doesn’t think voluntariness is sufficient because of the case of children and animals.
They’re able to act voluntarily, but they can’t make decisions. (III.1, 1111a24-6). And it seems like deliberation is required for responsibility, in addition to voluntariness.
So, since responsibility is required for praise/blameworthiness, children/animals aren’t appropriate recipients.
Although involuntariness is sufficient for pardon, it’s not necessary. Specifically, if human nature is “overstrained”, voluntary actions might be pardonable.
Aristotle says “presumably there are some things we cannot be compelled to do” (III.1, 1110a26-29): they’re so terrible they’re voluntary (because you could accept death instead), but it’s nonetheless pardonable to have carried it out.
There are limits to this pardon, though. e.g., Alcmaeon killing his mother under the supposed compulsion of the gods is so heinous it can’t be pardonable. (And that surely only makes sense if it was voluntary.)
There are three categories of action: voluntary, involuntary and non-voluntary.
An action is voluntary if it meets two conditions:
(i) The principle (arche) is in the agent
(ii) The agent knows the particulars of the action
Note that (i) is an extremely weak requirement. E.g. if you’re held at gunpoint, that would count as the principle being in you.
The agent must “contribute nothing – if for instance, a wind or people who have him in their control were to carry him off” (III.1, 1110a3-4)
For (ii), ignorance of universals, i.e. moral rules / “what is beneficial”, isn’t sufficient to make an action involuntary.
The specific particulars to be known are: who is acting, what they’re doing, what it concerns, the instrument, the result, and the manner (III.1, 1111a3-6)
If an action is done in ignorance but because of other reasons like anger or drunkenness, it might still be voluntary & blameworthy, because it is up to an agent whether they pay attention to things.
“A drunk, for instance, pays a double penalty; for the principle is in him, since he controls whether he gets drunk, and his getting drunk causes his ignorance.” (III.5)
So for (ii) to fail, the action needs to be caused by ignorance.
Aristotle discusses the idea of “mixed actions” which seem to involve voluntary and non-voluntary components, but it seems better to call them putatively-mixed – Aristotle just thinks that they’re voluntary (Nielsen 2007).
For example, throwing cargo off a sinking ship – or in general, intolerable dilemmas.
This doesn’t really line up with our usual concept of “voluntary”; Aristotle imposes extremely weak conditions for voluntariness.
But Aristotle isn’t saying that a mixed action can be at once right under the circumstances yet shameful. There are no “dirty hands” (required but morally problematic actions).
Recall that virtuous agents don’t feel shame. So of course they can’t feel remorse about their actions in coerced situations.
With the example of the man commanded by the tyrant to do something base, Aristotle really means that it is an action which is usually base, but in the circumstances in fact not so. When he says people are sometimes praised for enduring something shameful, this is in the sense of being shamed, not deserving it.
The agent who deserves blame for his mixed act is one who endures a shame for a trifling end.
Some non-voluntary actions can be further classified as involuntary, depending on the emotional response of the agent.
“Everything caused by ignorance is non-voluntary, but what is involuntary also involves pain and regret.” (III.1, 1110b18-19)
This is relevant for assessing someone’s character: we can’t observe it directly, but we can observe their feelings & actions.
The reaction of someone with good character to performing a bad action non-voluntarily will be one of pain & regret.
Aristotle thinks we can hold people responsible for their bad actions even if they’ve been habituated into that disposition, because they’re responsible for that habituation.
Recall that Aristotle holds that everyone sensible must be aware of habituation. So, you’re blameworthy for failing to foresee the consequences of getting “stuck” in a vicious state.
“the person who is [now] unjust or intemperate was originally free not to acquire this character, so that he has it willingly, though once he has acquired the character, he is no longer free not to have it” (III.5)
As Sauvé Meyer (2006) notes, this requires an additional assumption of transitivity of responsibility.
Aristotle isn’t directing his argument at people who’ve been raised poorly – so this isn’t as objectionable as it superficially seems.
Aristotle also thinks it’s your responsibility if you have a mistaken conception of the right ends, justifying his exclusion of ignorance of the universals as grounds for non-voluntariness.
People are responsible for their conception of the ends because it’s determined by their character, which in turn is voluntarily habituated.
“The sort of character we have determines the sort of end we lay down” (1114b23-5)
Partly Aristotle is responding to Plato’s idea that all wrongdoing is involuntary because it stems from ignorance of the good, but all right actions are voluntary.
There are two prongs to Aristotle’s rejection of the claim that because the apparent good depends on character, we have no control over it.
(i) If you’re responsible for your character then you’re responsible for how the good appears, and hence virtue/vice.
(ii) Even if you’re not, how can virtue be any more voluntary than vice?
So virtue and vice are either both voluntary or both involuntary.
By endoxa we praise & blame for those things, so presumably they both are voluntary.
(At a meta-level, it’d be unappealing to accept that we have no control over how the good appears to us, because then Aristotle would need to give up on praise/blameworthiness entirely.)
Aristotle does accept that external factors play a part – we are “jointly responsible” for our states of character along with nature and upbringing.