(p.170) Appendix I Varieties of Knowledge and Ignorance
(p.170) Appendix I Varieties of Knowledge and Ignorance
Aristotle says various conflicting things, in the different ethical works, about the kind of knowledge that voluntariness requires. These different positions become intelligible on the interpretation for which I have been arguing. That is, if we suppose that Aristotle evaluates proposed criteria for voluntariness by considering whether they entail the voluntariness of actions for which the agent is praiseworthy and blameworthy, and if we suppose that his goal is to count as voluntary all activity of which the agent’s moral character is the cause, then we can understand why he says what he does about the voluntary agent’s knowledge.
It is reasonable to suppose that Aristotle’s articulation of the cognitive criteria for voluntariness and involuntariness should provide evidence of the theoretical goals of his inquiry into voluntariness. As we saw in chapter 2 Aristotle insists, in all three of the ethical works, that voluntary actions are those that originate in the agent, are up to the agent, and of which the agent is causally responsible (aitios). He uses these locutions interchangeably. That is, he indicates that voluntariness is a causal relation between an agent and an action, and that the nature of voluntariness can be captured by this single causal condition, alternatively described by these different locutions. But when Aristotle comes to define voluntariness he requires not only that the origin of the voluntary action be in the agent, or that it be up to the agent, but also that the agent know what he is doing.1 He does not indicate that in adding the requirement of knowledge he admits the simple causal criterion to be inadequate. So, it is reasonable to suppose that the (p.171) requirement of knowledge expresses Aristotle’s assessment of what it takes for the action to originate in the agent; that it expresses a necessary condition for satisfaction of the causal condition he initially mentions as being constitutive of voluntariness.2 If, as chapters 2 and 3 have established, Aristotle intends his account of voluntariness to capture the conditions in which agents are praiseworthy and blameworthy for their actions and thinks they are praiseworthy and blameworthy for the actions produced by their moral character, then we should expect that in articulating the cognitive requirements for voluntariness, he will be sensitive to these considerations. That is, we should expect that he will reject versions of the cognitive requirement that fail to count as voluntary praiseworthy and blameworthy activity, and that the activity he will judge praiseworthy and blameworthy will be activity that is produced by the agent’s moral character. If we examine the various cognitive requirements for voluntariness that Aristotle articulates in his different discussions of voluntariness, we will see that his articulation of the requirement is sensitive to precisely these considerations.
In his various discussions of voluntariness and involuntariness, Aristotle claims that the voluntary agent must know the “particulars” but need not know the “universals” concerning his action. He claims that the agent whose involuntariness is due to ignorance must not only act “in ignorance” (agnoōn) of what he does but also “because of ignorance” (di’ agnoiari). And he appears to be of two minds on the question whether the agent who acts in culpable ignorance of particular facts satisfies the cognitive condition for involuntariness. We will see that Aristotle requires knowledge only of particulars for voluntariness because no other kind of ignorance can keep the agent’s character from being productive of the action in question. His discrimination between action performed in ignorance and because of ignorance shows that he is willing to call “because of ignorance” only those actions of which the agent’s character is not productive. Finally, his remarks about culpable ignorance and its relevance for voluntariness and involuntariness are motivated by his desire to capture, with his account of voluntariness, those actions produced by the agent’s character.
1 Universal and Particular Knowledge
In MM i 16, when Aristotle concludes that voluntary action is what is done “from thought” (ek dianoias), he does not elaborate upon the sort (p.172) of thought he has in mind. His examples in fact indicate some confusion over whether the thought in question is something like premeditation, or rather something like knowledge of the consequences of one’s actions (1188b28–37).3 In the corresponding passage in EE ii 9, Aristotle interprets the requirement of ‘thought’ as a requirement of knowledge, and illustrates the sort of knowledge he has in mind as follows:
Knowing either whom or how or with what result (ē hon ē hō(i) ē hou heneka) – i.e., sometimes one knows this is one’s father, but one does not act in order to kill him but rather to preserve him, as the daughters of Pelias did; or one might know this is a drink but think that it is a love potion and wine when in fact it is hellebore. (EE 1225b2–5)
In EN v 8, a book common to the EE and EN that repeats an account of voluntariness similar in many respects to that of EE ii 9 and appears to be referring back to it, Aristotle gives a similar illustration of the sort of knowledge voluntariness requires:
knowing and not being ignorant either of whom or how or with what result (mēte hon mēte hō(i) mēte hou 〈heneka〉) – for example: whom one hits (tina) and with what (tini) and for the sake of what (tinos heneka) … It is possible for the person hit to be your father but that you know only that he is a man or one of the company and not that he is your father. (EN 1135a24–30)
In EN iii 1, as in EE ii 9, Aristotle requires knowledge for voluntariness. His list of the relevant types of knowledge, and his examples, coincide with his illustration and examples of the relevant type of knowledge in EE ii 9 and EN v 8::
The person who is ignorant of one of these things acts involuntarily. It is presumably not a bad idea to enumerate them, what and how many they are: certainly who acts, what is done, and concerning what, or in what; sometimes also with what (for example, with what instrument), and for the sake of what (for example, preservation), and how (for example, lightly or strongly). (EN 1111a2–6)
The examples he gives include:
One might intend to show off the catapult, but set it off. And one might take one’s son for the enemy, as Merope did; or think that the spear is (p.173) blunted, when it is really pointed; or that a rock is pumice stone; and one might make someone drink in order to save him, but kill him instead. And while intending just to touch someone, as wrestlers do, one might strike him instead. (EN 1111a10–15)
Unlike EE ii 9 and EN v 8, Aristotle here in EN iii 1 explicitly distinguishes the kind of knowledge necessary for voluntariness from a different kind of knowledge:
Certainly every wicked person is ignorant of what one must (dei) do and what one must refrain from (aphekteon), and because of such error people become unjust and generally bad. But “involuntary” is not intended to apply to the case of someone who is ignorant of what is beneficial (ta sumpheronta), for ignorance in one’s decision (prohairesei) does not cause involuntariness but rather wickedness. Nor yet does ignorance of the universal (for one is blamed on account of that), but rather ignorance of the particulars (hē kath’ hekasta) concerning and constituting the action.In these cases there is both pity and excuse, for someone who is ignorant of one of these things acts involuntarily. (EN 1110b28–1111a2)
Aristotle here identifies the sort of knowledge required for voluntariness by describing the sort of ignorance that can render an action involuntary.4 He describes the kind of ignorance that can make an action involuntary as ignorance of “the particulars” (ta kath’ hekasta). He contrasts this kind of ignorance with ignorance that he variously describes as: ignorance of what one ought (dei) to do (1110b28); ignorance of what is beneficial (ta sumpheronta) (1110b30–1); and ignorance in one’s decision (prohairesei) (1110b31).
Aristotle here invokes a distinction between two types of thought involved in the causation5 of action, a distinction he makes frequently both inside and outside the ethical works. He regularly distinguishes between two kinds of “premises” (protaseis) involved in the inference that explains the occurrence of an action. He regularly labels one kind of premise “universal” (hē katholou), the other premise “particular” (hē kath’ hekasta).6 The “universal” premise is a thought about what sorts of things are good or bad for one, and generally takes the form of a claim that one ought (dei) to do something, alternatively expressed in the gerundive.7 The “particular” premise is a thought that a particular object or alternative satisfies the conditions that the universal premise puts forth as desirable or to be done. The particulars are the sorts of things one can perceive.8 For example, the universal premise might be (p.174) that stagnant water is bad, and the particular premise that this water is stagnant (EN 1142a20–3). Both sorts of premise are involved in the causation of the action:
One can also consider scientifically (phusikōs) the cause 〈of action〉 in the following way. For one belief (doxa) is of the universal (katholou), while the other is of the particulars, which perception controls. When one thing results from them, it is necessary that the soul affirm the conclusion or – in the case of thoughts about production – it is necessary that action immediately follow. For example, if one ought to (dei) taste anything sweet, and this particular thing is sweet, it is necessary that the agent who is able and unimpeded do this 〈sc. taste this〉. (EN 1147a24–31)
Here Aristotle indicates that the combination of these two premises results in action of necessity, unless something impedes it.9 He makes the same point at MA 701a15–16.
Both kinds of premise, the universal and the particular, can be true or false. One can be mistaken in one’s objects of pursuit (i.e., think one ought to do a certain kind of thing when one really ought not to). Or one can be mistaken in one’s beliefs about the particular situation in which one applies one’s beliefs about what one ought to do. The difference between such mistakes is the difference between believing mistakenly that it is permissible to pull the trigger of a loaded gun, and mistakenly believing that the gun whose trigger one pulls is not loaded.
Aristotle in EN iii 1 indicates that voluntary action requires knowledge of the particular premise, but is compatible with ignorance of the universal premise. While he does not here use ‘universal’ (katholou) as the general label for this type of ignorance, his examples indicate that this is the sort of ignorance he has in mind. For example, he first describes this ignorance as ignorance of what one ought (dei) to do (1110b28) – which at EN 1147a25–31 he gives as an example of a universal premise (cf. DA 434al6–19).10 He next describes it as ignorance of what is beneficial (ta sumpheronta), explaining this (ou gar, 1110b31) as “ignorance in one’s decision (prohairesei)” (b31). The thought that something is healthy or bad for one are examples he has given elsewhere of the universal premise (1141 b 14–22, 1142a20–3). Being healthy and being bad for one are specific ways in which something can be (or fail to be) beneficial (sumpheron). Ignorance about such matters counts as “ignorance in the prohairesis” because such benefits are the goals (telē) in the light of which one deliberates, the result of such deliberation being a prohairesis; for example, health is the goal in the (p.175) light of which the doctor deliberates (EN 1112b12–20).11 It is somewhat surprising that Aristotle here uses the label ‘universal’ for another kind of knowledge added to the list, instead of a general category into which the preceding types of knowledge fall.12 However, this should not keep us from concluding that Aristotle here, in EN iii 1, unambiguously claims that knowledge of the sort to which he elsewhere applies the general label ‘universal’ is not necessary for voluntariness.
Aristotle appeals to considerations of blameworthiness and character when he explains why knowledge of the universal premise is not necessary for voluntariness. He begins his description of the types of ignorance that do not make actions involuntary by saying that it is the sort of ignorance the bad person (ho mochthēros) has (1110b28). Such ignorance, he goes on to say, is a cause not of involuntariness but of wickedness (1110b31–2). Ignorance of the universal is not a feature of an action that makes it involuntary but is rather a feature of an agent that makes the agent bad (mochthēros). Finally, he claims, ignorance of the universal is a reason for blame (1110b32–3). These considerations show that Aristotle intends his account of voluntariness to capture the conditions in which agents are praiseworthy and blameworthy for what they do. And they confirm the suggestion that an evaluation of the agent’s moral character is central to the judgements of praiseworthiness and blameworthiness.
Furthermore, if we reflect upon the different causal roles played in the causation of action by universal and particular thoughts, Aristotle’s restriction of the requirement of knowledge for voluntariness to knowledge of particulars supports the interpretation that Aristotle thinks actions produced by the agent’s moral character are voluntary. In claiming that ignorance of the universal is a cause of wickedness (1110b31–2), Aristotle tells us, reasonably enough, that an agent’s universal beliefs, his beliefs about the sorts of things it is worthwhile to do, are constitutive of his character. This is because “universal” beliefs are constituted by the agent’s standing desires – his dispositions to have certain occurrent desires in particular circumstances. Such dispositions are part of the agent’s character. The particular beliefs, by contrast, are not states of one’s persisting character. They are simply occurrent states that represent the circumstances in which the agent finds himself. Such beliefs supply the information that engages a particular standing desire, and produces an occurrent desire for the action. In order for the character, the set of standing desires, to be productive of an action in the appropriate way, the particular beliefs must be true, but the “universal” beliefs need not be.
(p.176) It is easy to see how actions involving and even depending on ignorance of what one should do do not thereby fail to be produced by and expressive of the agent’s character. Such ignorance is constitutive of the agent’s character. It is also easy to see how ignorance of particular facts can keep the action actually performed from depending on one’s character. For example, if Oedipus knew that the old man he met at the crossroads was his father, then the action he performed would have depended on his “universal” beliefs about how one should (dei) treat one’s father. In such circumstances his action would have shown him to be vicious regarding his parental obligations. However, in the circumstances in which he acted, he did not know that the old man was his father, and hence we cannot conclude from the fact that he killed his father that he has the corresponding vice. Given his ignorance of this crucial fact, we have no reason to suppose that his attitudes toward his father were engaged at all in the action. Since one’s beliefs about the particular facts about one’s action play an important role in the process whereby one’s character is productive of one’s actions, it seems reasonable for Aristotle, in his account of voluntariness, to require that these beliefs not be mistaken.
The best explanation of Aristotle’s account of the sort of thought or knowledge necessary for voluntariness is that he modulates the requirement in order to preserve the result that actions produced by the agent’s moral character, and indicative of that character, are voluntary.
2 “In Ignorance” vs. “Because of Ignorance”
The preceding explanation of Aristotle’s motive for restricting the sort of knowledge necessary for voluntariness to knowledge of particulars allows us to understand a further distinction Aristotle makes between types of ignorance: the distinction between acting “because of ignorance” (di’ agnoian) and merely acting “in ignorance” (agnoōn). He makes this distinction in two places – once in EN iii 1 (1110b24–7) and once in EN v 8 (1136a5–9). Let us first examine the distinction as Aristotle introduces it in EN iii 1, immediately before his distinction between universal and particular ignorance:
Acting because of ignorance is different from acting in ignorance. For the person who is drunk or angry does not act because of ignorance, but rather because of one of the things mentioned 〈sc. being drunk or being angry〉 – although not in knowledge but in ignorance. (EN 1110b24–7)
(p.177) Immediately after he makes this distinction, Aristotle proceeds to distinguish between universal and particular ignorance, claiming that only the latter makes an action involuntary (1110b28–33). He does not indicate that the second distinction, concerning the content of the ignorance, introduces a requirement additional to the requirement introduced by the first distinction: that the involuntary agent act because of ignorance. On the contrary, the second distinction begins in a way (agnoei men oun, 1110b28) that suggests it provides further support for the requirement expressed in the first distinction. It is therefore natural to interpret the distinction between acting in ignorance and acting because of ignorance as corresponding to the distinction between acting with ignorance of the universal and acting with ignorance of the particulars. On this interpretation, the angry person and the drunk are ignorant of what they should do. The angry person, for example, desires to harm the object of his anger, his anger mistakenly presenting this goal as permissible or required. The person who is drunk has his inhibitions lessened, and will undertake to do things he would think inappropriate if he were sober.13
Aristotle also distinguishes between acting in ignorance and because of ignorance at the end of EN v 8:
Some involuntary actions are excusable (sungnōmonika), while others are not. For those errors that people commit (hosa … hamartanousi) not only in ignorance (agnoountes) but also because of ignorance (di’ agnoian) are excusable. But those they commit not because of ignorance but rather in ignorance due to a feeling (pathos) that is neither natural nor human are not excusable. (EN 1136a5–9)
It is, however, hard to see how the distinction invoked here can be interpreted in the same way as I have suggested it is natural to interpret it in EN iii 1. The ignorance that here counts as ignorance “in which” but not “because of which” the agent acts cannot be ignorance of the universal, for Aristotle indicates here that the action performed in such ignorance counts as involuntary. The ignorance in question therefore seems to be ignorance of particulars. And it is easy to see why Aristotle should here want to draw the distinction between acting in ignorance of particulars and acting because of such ignorance. In the account of voluntariness and involuntariness presented earlier in the chapter, Aristotle describes the types of involuntary actions involving ignorance as actions which are unknown (agnooumenon, 1135a31–3), and his examples (p.178) make it clear that the ignorance in question is ignorance of the particulars (1135a28–30). Unlike EE ii 9 and EN iii 1, he does not explicitly require that such actions be “because of ignorance.” Hence it is reasonable to suppose that these remarks at the end of EN v 8 serve to distinguish between actions that are performed in ignorance of the particulars from those that are performed both in such ignorance and because of it.
So the distinction between acting “in ignorance” and acting “because of ignorance”, if Aristotle intends the same distinction in EN v 8 and EN iii 1, does not simply correspond to the distinction between acting with universal and acting with particular ignorance. Rather, Aristotle must intend the distinction to capture the contrast between some other pair of properties. The property he intends to capture with the label “in ignorance but not because of ignorance” must be a property common both to actions performed due to ignorance of the universal (as in EN iii 1) and to actions performed in ignorance of particulars due to “a feeling neither natural nor human” (as in EN v 8, 1136a8–9). In order to identify the common feature of these two cases, we need first to identify the feelings that count as “neither natural nor human” (1136a8–9).
A reasonable guide to the interpretation of “a feeling neither natural nor human” at 1136a8–9 is Aristotle’s remark that:
What an agent does in knowledge but without prior deliberation is an injustice (adikēma) – for example, actions due to spirit (thumos) and other feelings that are necessary or occur naturally to human beings.(1135b19–22)
Such cases are cases of voluntary, rather than involuntary activity, but the natural and necessary feelings referred to here are probably the same as the “natural and human” feelings referred to at 1136a8–9. Aristotle proceeds to say that such actions as these are not due to vice (dia mochthērian, 1135b24). He cannot mean by this remark that no action is due to vice if it results from any of the kinds of feelings for which we naturally have the capacity in our souls,14 for this would conflict with his account of virtue and vice as dispositions to have and act in the light of these feelings.15 Rather, he must have in mind a particular class of feelings which do not express the agent’s moral character. If he has in mind particular token feelings that anyone in the same situation would feel, feelings that are for this reason “natural and necessary”, then (p.179) he would have grounds for claiming that such actions are not due to vice. Rather, they would be due to human nature, and hence not proper to one’s character. Let us suppose, therefore, that these are the feelings to which Aristotle refers at 1136a8–9.
On this interpretation, ignorance that is of particulars but due to feelings that are natural or human (1136a8–9) would be ignorance that is not due to the agent’s character, since such feelings are natural and hence not within the scope of character. Ignorance of particulars that is due to feelings not of this type, however, is due to the agent’s character, since such feelings are not natural and hence are within the scope of character. On this interpretation, the actions that, according to EN v 8, are performed both in ignorance of particulars and because of such ignorance are cases of actions in which the agent’s character is not responsible for the ignorance. And actions that are performed merely in such ignorance but not because of it are actions in which the agent’s character is responsible for the ignorance and derivatively for the action. In this respect, the distinction between acting merely in ignorance and acting because of ignorance in EN v 8 corresponds to the distinction in EN iii 1. For in EN iii 1, the ignorance that fails to count as ignorance “because of which” the agent acts is, as in EN v 8, ignorance that is a product of the agent’s character.
In both EN v 8 and EN iii 1 the ignorance that fails to count as ignorance “because of which” the action occurs is ignorance that is either constitutive of or produced by the agent’s moral character. But why should the fact that the ignorance in question stands in this relation to the agent’s moral character give Aristotle reason to deny that the action is “because of ignorance”. After all, the ignorance in question plays a causal role in the action. Aristotle’s reason for denying that such ignorance is ignorance “because of which” the action occurs is intelligible if he intends “because of ignorance” to capture “because of ignorance and not because of the agent”. This, I propose, is what the distinction between acting in ignorance and because of ignorance means for Aristotle. Aristotle’s judgements about whether an action is because of the agent (rather than because of his ignorance) depend on his assessment of whether the ignorance in question is a product or feature of the agent’s character. Aristotle’s judgements about whether actions are due to ignorance are explicable on the hypothesis that he thinks voluntary actions are not due to ignorance and that actions produced by one’s character are voluntary.
3 Culpable Ignorance
(p.180) The discussion of EN iii 1 seems to allow that any action involving ignorance of particulars is thereby due to ignorance of particulars (1111a15–18). But EN v 8, on the interpretation offered above, gives reason to suppose that not all such ignorance renders the action “because of ignorance”. If the agents character is responsible for that ignorance of particulars, then EN v 8 does not count that action as “due to ignorance”. (I shall refer to such ignorance as “culpable ignorance”.) EN v 8 does not claim that culpable ignorance of particulars keeps an action from being involuntary. It simply claims that the culpability of the ignorance affects the propriety of giving “forgiveness” (sungnōmē) for the involuntary action. An action performed in ignorance of particulars is, according to EN v 8, involuntary, but such an involuntary action does not merit sungnōmē. But even though EN v 8 does not claim that culpable ignorance renders an action voluntary, one might wonder whether it does not provide a reason for such a claim. After all, if Aristotle thinks the activity produced by the agent’s character is voluntary, then action performed in culpable ignorance of particulars would seem to be voluntary. Aristotle explicitly discusses culpable ignorance in two places in his discussions of voluntariness: in EN iii 5, 1113b23–1114a3 and in EE ii 9, 1225b11–16.
In EN iii 5, Aristotle claims that an agent who does something bad is responsible for it unless the action is due to force or due to ignorance for which the agent is not responsible (di’ agnoian hēs mē autoi aitioi) (EN 111 3b23–5). It is not clear that the ignorance in question is ignorance of the particular facts rather than ignorance of what one should and should not do (ignorance of the universal). Since the chapter is one in which Aristotle argues that agents are responsible for their states of character, it is possible that the ignorance for which one is responsible is ignorance of what one ought to do. However, if the ignorance in question is supposed to be ignorance of the particulars, and if Aristotle continues to maintain the view that one is blameworthy only for voluntary actions, then these remarks indicate that Aristotle’s definition of voluntariness in EN iii 1, requiring knowlege of the particulars, needs to be revised.
Aristotle does not address this question explicitly in EN iii 5. This silence might be further evidence that particular ignorance is not the sort of ignorance he has in mind when he refers to “ignorance for which (p.181) the agent is responsible” at 1113b24–5. However, even if this ignorance is simply ignorance of the “universal”, there still remains an unclarity and tension in Aristotle’s account, for in his argument that our states of character (and hence our ignorance of the universal) are volun-tary (1114a4–13) he does not establish that we knowingly do what will produce these states of character in us; rather, he establishes simply that we have no excuse for not knowing this (1114a9–10). This claim would serve the conclusion that character development is voluntary only if culpable ignorance did not count as the sort of ignorance that makes an action involuntary.
A similar problem arises, in more explicit terms, in EE ii 9. In his preliminary discussion of voluntariness and involuntariness, Aristotle requires that the action whose involuntariness is due to ignorance must be “because of ignorance non-accidentally” (1225b5–6). Presumably the addition of “non-accidentally” to “because of ignorance” is intended to rule out cases in which the ignorance of particulars doesn’t play the relevant causal role in the production of the action. Aristotle does not indicate what sorts of particular ignorance he would take to be only accidentally involved in the production of the action. He might, for example, have in mind something like this. Oedipus strikes the old man at the cross-roads not knowing that the old man is a Theban. But his ignorance that the old man is from Thebes has no effect on whether he strikes him, since he would do the same even if he knew this. Even though Oedipus does not voluntarily strike the Theban, his ignorance that the old man is a Theban does not undermine the claim that he voluntarily strikes the old man at the crossroads.16
Aristotle, however, also might intend the remark, about being due to ignorance non-accidentally, to be interpreted in the light of certain remarks about culpable ignorance which he appends to the definition of voluntariness and involuntariness:
Knowledge and knowing are twofold – one being to have the knowledge, the other to use it. The person who has but does not use the knowledge might rightly be said not to know. But he might also not rightly be said not to know, for example if he failed to use it due to carelessness. Similarly too, the person not having the knowledge might be blamed, if he fails, due to carelessness or pleasure or pain, to have some knowledge that is easy or necessary to have. So let these things be added to the definition. (EE 1225b11–16)
Aristotle’s remarks here are clearly motivated by the thought that agents who act in culpable ignorance are blameworthy, as he remarks (p.182) explicitly in discussing the second type of culpable ignorance. He does not, however, appear to have clearly thought through the implications that these remarks are to be “added to the definition”. It is not clear what the second case is supposed to add, since it only claims that such agents are blameworthy, not that they are voluntary. Does he intend that such agents act voluntarily? If so, then he cannot require knowledge for voluntariness, as he has just done in the definition above. This would not be a simple addition to but an amendment of the definition.
Aristotle does not resolve these issues, or even address them explicitly. For our present purposes, it is sufficient to note that Aristotle’s assessment of whether a given type of ignorance keeps an action from being voluntary is sensitive to his judgement about whether the agent is blameworthy for the action in question. Aristotle’s unresolved difficulty here about whether knowledge is necessary for voluntariness, or whether culpable ignorance would suffice, is explicable in the light of the assumption that, in his account of voluntariness, he is seeking to capture the conditions in which an agent is praiseworthy or blameworthy. And, as we have had occasion to observe at various points in chapter 3, the judgements about praiseworthiness and blameworthiness against which Aristotle checks proposed criteria for voluntariness are closely connected to assessments of the agent’s moral character. That is, the praise and blame for which he thinks voluntariness is a necessary condition are the retrospective moral evaluations of concern to moral responsibility.
(1.) EN 1111a22–4; EE 1225b8–10; cf EN 1135a23–8. Only the definition in the MM states a single criterion: that the action be “from thought” (1188b26; cf. b38).
(2.) For example, the requirement of knowledge in EE ii 9’s definition of voluntariness is stated: “not in ignorance and because of oneself” (mē agnoōn kai di’ hauton, 1225b8–9). And the recapitulation of the definition at EE ii 10, 1226b30–2 shows that the “and” (kai) in this articulation of the condition is epexegetic, for the equivalent condition is there stated as: “because of oneself and not because of ignorance” (di’ hauton kai mē di’ agnoian, 1226b31–2).
(3.) In this respect, the MM’s definition reflects contemporary legal criteria, according to which pronoia (forethought) is considered necessary for voluntariness, but what counts as pronoia ranges widely (cf MM 1188b35). The MM cites a trial in the Areopagus as evidence in favour of its definition (p.183) (1188b31–7). Plato, in Laws ix counts as involuntary (akōn) agents who act unintentionally (865a–c), but counts as voluntary (hekōn) only agents who act from the deliberate forethought and plotting characteristic of consummate and incurable vice (869e–873c). He refuses to count as either voluntary or involuntary actions performed in knowledge, but prompted by anger (866d–868a), indicating that there is considerable controversy about the voluntariness of such cases (867b). These controversial cases would seem voluntary if mere awareness satisfied the cognitive condition for voluntariness, but involuntary if prior deliberate intent is required.
(4.) That is, if supplemented by pain or regret (EN 1111a20–1).
(5.) EN 1147a24–6 makes it clear that to state the two types of thought is to state the cause (aitia) of the action in the manner of the natural scientist (phusikōs, 1147a24).
(6.) The distinction invoking the terms ‘universal’ and ‘particular’ is made at EN 1112b33–1113a2; common book EN 1141b14–22, 1142a20–23, 1147a25–31. The same distinction is made, but without the labels ‘universal’ and ‘particular’ at DA 434a16–19 and MA 701a10–11.
(7.) Universal premise about what is good or bad: EN 1141b14, 1142a20; of the form dei (ought): EN 1147a29, DA 434a18; expressed in the gerundive (e.g., badisteon, poiēteon): MA 701a13, 14, 16.
(8.) EN 1142a27–30, 1143b5, 1147a26, 1113a1–2; cf. 1109b18–26, 1126b2–4.
(9.) On the absence of impediment, see Met. 1048a16– 21.
(10.) In the same phrase he also uses the gerundive (aphekteon) (1110b29), the form he uses in MA 7 to give the universal premise, although he does not there use the terminological distinction between universal and particular.
(11.) On error about the universal as error involved in deliberation: EN 1142a20. On the relation between the goals and the results of deliberation as the relation between universal and particular: EN 1112b33–1113a1.
(12.) I would explain this more restricted use of ‘universal’ as follows: to d’ akousion bouletai … en hois kai peri ha hē praxis (1110b28–1111a1) is a de-clause in contrast with the men-clause: agnoei men oun pas ho mochthëros … kakoi gignontai (b28–30). The types of ignorance listed in the de-clause (of the beneficial, in the prohairesis, of the universal) therefore are all supposed to be of the same type as the types of ignorance in the men–clause (of what one should (dei, -teon) do). The oud’ in oud’ hē katholou (b32) effects a contrast between ignorance of the sumpheronta (which is ignorance in the prohairesis, as indicated by the gar in b31) and ignorance of the universal. The intended contrast is probably between ignorance of the highest goals in the light of which one deliberates, and the subordinate goals that contribute (sumpherei) either instrumentally or intrinsically to those goals. The intended contrast is between ignorance of whether one should act justly and ignorance of whether returning this deposit is just. Good (p.184) intentions do not excuse, Aristotle indicates, if one is ignorant about how to execute them.
(13.) Further support for the interpretation that, in EN iii 1, actions performed in ignorance of particulars are thereby because of ignorance comes from 1111a15–18, where Aristotle indicates that he takes simply being ignorant of one of the particular facts to satisfy the condition of acting because of ignorance of such facts.
(14.) EN 1106a9; cf. EE 1220b7–8.
(15.) In Rhet. i 13 he claims explicitly that actions due to feelings can be due to wickedness (apo ponērias) (Rhet. 1374b9–10; omitted by one MS and William of Moerbeke).
(16.) This explanation was suggested to me by Terence Irwin.