Deontic and virtue conditions

A common theme among authors who have written about the moral lives of heavenly agents is to confine such discussions to the topic of right actions. This focus on the right, appropriate, or moral acts of those in heaven makes sense given the basic biblical and theological agreement in the history of Christendom that heaven is an environment in which there is no sin or evil, and that an at least partial explanation for this lack of evil is that heavenly agents cannot sin. Augustine argued that, in contrast to Adam and Eve's ability to sin in the garden, heavenly agents will have an inability to sin in heaven: ‘For when man was created righteous, the first freedom of will that he was given consisted in an ability not to sin, but also in an ability to sin. But this last freedom of will be greater, in that it will consist in not being able to sin’ (Augustine (Reference Augustine and Dyson1998), 22:30). According to Simon Gaine, ‘the traditional Christian answer is that the blessed cannot sin, cannot want to sin, but instead are impeccable’ (Gaine (Reference Gaine2003), 2).Footnote ¹ Now there may be some who are slightly bewildered by Gaine's use of the term ‘impeccable’ here, and this bewilderment is likely to be motivated by the fact that ‘impeccability’ usually appears in the literature in the context of God's nature, and thus, understood as a property, is something like ‘essential sinlessness’ (Morris (Reference Morris1983) and Carter (Reference Carter1985) ). Given that the context of this article concerns the moral character of contingently existing agents that have all had ‘sinful natures’ at one point in their existence, the modality attached to ‘sinlessness’ here is not going to be the same. Therefore the cannot in ‘cannot sin’ should be understood in a causal or temporal sense of modality, rather than a metaphysical or logical sense. The property of ‘impeccability’ is a contingently accessed property (for contingent agents) that, once attained, renders acts of sin impossible.Footnote ²

Many would also likely argue that it is a point of orthodoxy (or perhaps tradition) to affirm that no heavenly agent will ever be removed from heaven as a result of ‘unheavenly acts’.Footnote ³ According to Aquinas,

Gaine echoes Aquinas here:

Thus it is natural to assume from Christian tradition that no heavenly agent will ever sin or perform such a morally blameworthy act that qualifies as ‘sin’. Since the status of ‘impeccable’ refers to the moral feature of heavenly agents, we can say at this point that to be an impeccable agent requires a deontic component, that only those actions deemed right or permissible are possible for such agents; acts that are deemed wrong or impermissible are impossible for such agents.

I would like, at this point, to propose the notion that impeccability has more than just a deontic component. Typically, definitions of ‘impeccability’ are exhausted by the modal descriptions I mentioned above (‘an inability to sin’ or ‘essential sinlessness’) and are thus confined to discussions of obligatory, permissible, and impermissible acts. However, I want to argue that impeccability implies a virtue component along with the deontic. The first reason why I think virtue is relevant for impeccability is derived from the fact that numerous biblical figures appeared to believe that character was just as relevant to an agent's moral life as were his or her actions. Christ makes the point that actions are not the only litmus test for praise or blame, but character is also important.Footnote ⁴ Hating someone confers just as much blame as actually committing the act of murder, and lusting after someone who is not your spouse is just as sinful as going through with the act of adultery. Here Christ seems to be implying that one can be guilty of sin without performing any overt actions, and that the guilt is based on one's internal disposition, or character. Also, in many of Paul's letters, the apostle seems to be arguing that a successful or godly moral life requires growth in virtue (Gal. 5:22–23; Eph. 4:2–3; Col. 3:12–13, NASB). While Christ's point doesn't give explicit consent to the inference from having an inability to sin to having a virtuous character, his point does seem to allow the inference from having an inability to sin to having an inability to be disposed to sin. And while Paul does not explicitly state that heavenly agents will be admirably virtuous agents, he does imply that a godly or holy individual with also be a virtuous one.

In addition to Scripture, the second reason I have for believing that impeccability implies a virtue component is that church history seems to support the idea; there are some within Christendom who have argued that there will be a ‘maturing’ or ‘completing’ in heaven of what is lacking in the characters of the human agents. Aquinas, for instance, seemed to argue that once man experiences the divine light or vision (in heaven), his desire to manage his lower self according to reason is truly fulfilled, enabling the agent to live completely according to virtue.

Augustine also appeared to support the idea that heavenly agents will be virtuous, arguing that virtue will finally have success or victory over vice once the redeemed enter the final beatitude: ‘the virtues will not be called upon to strive against any vice or evil whatsoever. Rather, they will possess the prize of victory: the eternal peace which no adversary can disturb. This is our final happiness, our last perfection, a consummation which will have no end’ (Augustine (Reference Augustine and Dyson1998), 19.10).

Thus, while the term ‘impeccability’ may not explicitly say anything about the virtue component of morality, as it has typically been used, I do think I have warrant to treat the term as though it implicitly involves a virtue feature.

Now while it is fairly straightforward to say what the deontic condition amounts to (an inability to sin), the virtue or ‘heavenly character’ component may be a bit more difficult. The reason for the difficulty lies in certain tensions among virtue theorists concerning the appropriate way to understand the ‘virtues’ and how such virtues are expressed in actions.

There are two relevant issues for the purpose of this article, concerning the proper conception of the moral virtues: (i) the broadness of the virtues, and (ii) the relationship between virtues and action. There has been much controversy among virtue theorists on whether or not the virtues should be understood as ‘broad’ or ‘narrow’.Footnote ⁵ A broad conception of the virtues implies that for any given virtue, if an agent has the virtue, then the virtue is relevant or applicable cross-situationally. In other words, if the agent has the virtue ‘courage’, then the agent has the ability to act courageously across any and all situations the agent could possibly find himself in. A narrow conception, on the other hand, implies less applicability of virtue exemplification. To have the virtue ‘courage’ narrowly, typically implies the application or expression of courage in a certain set of circumstances; this ‘set’ is usually understood as the circumstances that the agent is well acquainted with, or at least not far off from the agent's ‘normal’ array of circumstances.

The other relevant issue concerns the relationship between the virtues and virtuous actions. The important question is, for any acquired virtue, what is the likelihood of expressing that virtue in action, assuming the expression of the virtue is appropriate in the particular circumstance? If Mike has the virtue ‘courage’ and he is in a situation in which acting courageously is appropriate, how confident can we be that Mike will act courageously?Footnote ⁶ Again, there are two relevant positions a virtue theorist may take on the question. Those who have a particularly robust conception of the virtues will say that if Mike has the virtue ‘courage’, and Mike is in a situation in which acting courageously is appropriate, then necessarily Mike will act courageously (assuming nothing external to Mike prevents him).Footnote ⁷ A more moderate position will say that if Mike has the virtue courage, and he is in a courage-appropriate scenario, then Mike will probably act courageously (assuming nothing external prevents him). Thus the issue is whether virtuous action necessarily follows from a virtue trait, or whether it probabilistically follows from a virtue trait.

I will be assuming that impeccability implies a strong conception of the virtues: the virtues should be understood broadly, and virtuous action should be understood to follow necessarily from virtuous traits. For any agent who has all the relevant virtues for a heavenly character (whichever virtues those may be), and has them strongly according to broadness and action-necessitation, such an agent should be understood as ‘perfectly virtuous’. This conception appears to be adopted by numerous theistic philosophers in their own defence or explanation of the heavenly character of an impeccable agent, and while these authors do not use the terminology of ‘broad vs narrow’ and ‘strong vs moderate’ in reference to heavenly characters, I do think it is clear that they support the moral perfection of heavenly agents, and thus the strong conception I am assuming here. For instance, some authors abstain from articulating the particulars of the heavenly character, but yet still refer to such agents as ‘morally perfect’ (Wall (Reference Wall1977), 353; Brown (Reference Brown1985), 447). Others, however, are a bit clearer on the implications of being morally perfect. Sennett, for instance, provides an explanation of moral perfection that coheres well with my description of ‘perfect virtue’:

Notice that it is the virtuous skills or character, determined freely by agents prior to heaven, that preclude the possibility of acting (or desiring) contrary to the good. Also, since such a character is formed prior to heaven, it seems to imply that the dispositions for virtue should be understood broadly; that is, such an agent should behave virtuously in any context. Swinburne argues in a similar way when he claims that God sends agents to heaven because they have already fitted themselves for such an environment:

Given that there are at least some that seem to endorse the notion of ‘perfect virtue’, I think this conception of impeccability is a good place to start. So under the virtue condition for impeccability, the agent not only will be unable to perform acts of sin or evil, but the agent's character will be so virtuous that he or she will be unable to form the desire, motivation, and intention to perform acts of sin or evil.

Perfect virtue is impossible

Obviously, if it could be shown that the virtue condition was impossible to attain, it would be a major problem for the doctrine of impeccability. Michael Slote (Reference Slote2011) has recently attempted to provide an argument that aims at demonstrating this exact end, that perfect virtue is an impossibility. Slote's primary reason for claiming that perfect virtue is impossible is that the relevant virtues for perfection, or at least some them, are actually partial virtues. However, so the argument goes, if even some of the relevant virtues for perfection are partial virtues, then the status of ‘perfectly virtuous’ is impossible to attain. In what follows, I will elaborate on Slote's argument, then finish with a few responses.

Whatever else is implied by a perfectly virtuous agent, Slote appears to assume that the bare minimum for perfection entails that the agent:

Given that the first claim implies a weighty and lengthy discussion into the area of the ‘unity thesis’, Slote spends very little time discussing it; the bulk of his book is spent on arguing for the second claim. Whether Slote is successful or not at proving his second claim, I do not think his assumption that the first claim is necessary for perfect virtue is problematic, even if it is somewhat controversial. It is obvious that philosophers have disagreed about a number of issues related to this claim: How many virtues are there? Which virtues should be understood as moral rather than intellectual? Does the acquisition of one virtue assume or entail the acquisition of more virtues (or all the virtues)? I do not think, however, that these questions need to be answered decisively for the purpose of determining the possibility of perfection. Slote avoids the controversy of the unity thesis given his position on ‘dependency’ in the first claim; this just leaves the issue of the number and particulars of the actual virtues. This also doesn't seem to be a problem for Slote, for he could easily amend his first claim to say: ‘Has all the relevant virtues for perfection, whether independently or dependently’. Such a modification seems innocent enough, and it also makes the claim much less controversial. Therefore, I will spend the rest of my analysis of Slote by focusing on what he says about his second claim.

Partial virtues for Slote are virtues that are naturally opposed to one another; such opposition implies that acting on one virtue, in some scenarios, would entail compromising or acting contrary to another virtue (Slote (Reference Slote2011), 28–35). Further, acting contrary to, or failing to express a virtue (in action) when that virtue's expression is appropriate, confers some amount of moral blame upon the agent (ibid., 31). So if there is an agent out there who is thought to have all the virtues, and to have them as completely as possible, and assuming that some of these virtues are partial virtues, we get the conclusion that there is always the possibility that this agent would find himself in a scenario in which acting rightly or appropriately in accordance with one virtue would entail the acting against or contrary to another virtue that is also appropriate or right for that scenario. In such an event, it is impossible for the agent to emerge morally unscathed; the agent will necessarily have some moral regret, regardless of the virtue according to which he acts.

The two virtues that Slote believes are the best candidates for partial virtues are frankness (honesty) and tactfulness (kindness). Slote's thought experiment involves friend ‘A’ asking friend ‘B’ for advice about a personal matter in A's life. In the example, being a good friend to A entails that B respond with either frankness (honesty) or tactfulness (kindness):

In this scenario, responding with frankness would probably be good for A in the long term, but will just as likely cause A to experience emotional pain in the short term. Responding with tactfulness will probably cause A a small amount temporary peace, but will probably be bad for him in long run. According to Slote, in such a scenario, acting according to honesty or kindness would be ethically problematic for B. As Slote says, ‘these two qualities are paired opposites, and in some situations where they clash [such as this one], acting on either one of them will be ethically less than ideal’ (ibid., 31).

Now if these two virtues, or any other pair, are naturally opposed to one another, such that there exist possible scenarios in which acting kindly would entail being dishonest, or vice versa, then it seems as though no agent (even God) could possess both virtues perfectly. Assuming impeccability entails a perfectly virtuous character, and assuming that heavenly agents are also impeccable agents, then the impossibility of a perfectly virtuous character would necessarily lead to the incoherence of the notion of a heavenly agent. We can state the problem more formally as follows:

Premises (1) and (2) appear to be uncontroversial, for as I mentioned in the introduction, it seems a widely embraced feature of Christian orthodoxy that the agents in heaven will be impeccable; further, the property impeccableness implies (if not entails) that an agent with the property cannot sin (deontic condition) and will have a fairly virtuous character. I have chosen to leave the term ‘heavenly character’ vague for premise (2), for the purpose of keeping (2) uncontroversial. That is, I want to leave the question of the degree or kind of ‘virtuousness’ of character for an impeccable agent open until premise (3). Premises (3)–(5) are those I take to be more controversial. Up to this point, I have not really argued for (3) beyond mentioning that other theists have argued that heavenly agents should be understood as perfectly virtuous. Together, (4) and (5) are Slote's basic thesis, a thesis that is far from obvious at this point.

As far as I can see, the theist who wants to affirm the impeccability of the saints in heaven has three options for responding to the previous argument. First, one might attempt to undercut premise (5) by arguing that Slote has not sufficiently demonstrated that the relevant virtues for the notion of perfection are actually partial virtues. Second, one may argue that (4) is false by showing that a perfectly virtuous character is possible even if some of the relevant virtues conflict. One way to do this would be to acknowledge that such scenarios as Slote envisions are possible, but to deny that an agent in such a scenario would be required to act in a way that confers moral blame or sin. Finally, one might take the more extreme route and argue that (3) is false, that a heavenly character does not entail a perfectly virtuous character. This final option amounts to the admission that scenarios such as Slote's are possible, that partial virtues may exist, and that an agent in such a circumstance might be required to act in a way that confers moral blame on him or her, but deny that the possibility of such scenarios (and the possibility of moral blame that accompanies it) causes any problems for the notion of impeccability or the heavenly character of the saints in heaven. I will discuss each of these options in turn.

Option 1

The first option open to the theist is to deny that partial virtues exist by arguing that Slote has not been fully successful at showing that circumstances in which virtues necessarily conflict are possible. Slote himself acknowledges that most Aristotelian virtue theorists would reject his claim that the tact/truth scenario involves more than one appropriate response to A. According to Slote, ‘Aristotelians want to say that whenever there is a choice between tact and frankness, there is a right choice in the matter, a choice not open to moral or ethical criticism. On their view (and speaking rather roughly), frankness and tact never clash as virtues’ (Slote (Reference Slote2011), 41). One way to do this, for the Aristotelian, is to argue that in scenarios in which apparent conflict occurs among virtues, there is always one virtue that takes priority over the other; thus, acting from one virtue over the other would not produce any moral reproach, blame, or criticism. So, for instance, if there is a scenario in which acting mercifully and acting justly cannot both occur, acting justly may trump showing mercy. If so, then acting justly does not show that one is not perfectly merciful, and it does not indicate one is acting contrary to mercy; acting justly simply means showing mercy is not appropriate in that scenario (ibid., 42).

Slote's response to this point is that ‘putative virtues like tact and frankness are not as well behaved as the Aristotelian picture of the virtues assumes’ (ibid., 41) and that the scenario involving an apparent conflict between justice and mercy is not like the scenario he describes between tact and honesty. So Slote is willing to acknowledge that in a situation involving other virtues, say justice and mercy, claiming that justice ‘trumps’ mercy is not the same as claiming that acting justly requires acting unmercifully, and thus there is no real conflict between the two virtues. In the scenario with tact and honesty, however, Slote claims there is no one right response, no one path required by virtue; thus, necessarily, whichever path is chosen, the agent will come away ‘morally compromised’.

In his response to Aristotelians, it seems that Slote assumes that there is a fundamental difference in quality or value between the virtues of mercy and justice on the one hand, and honesty and kindness on the other. The distinction seems evident based on his faith that scenarios like that with honesty and kindness are possibly not duplicable for the virtues of mercy and justice.Footnote ¹⁰ Perhaps it is not the quality or value between mercy/justice and honesty/kindness that Slote believes produces the significant difference, but the nature of the relation between mercy and justice and the relation between honesty and kindness. Either way, it does seem that Slote needs to demonstrate why the virtues of justice and mercy, or the relation between them, are so different from the virtues of honesty and kindness, or the relation between them. If Slote cannot showcase such a meaningful and relevant distinction, then it seems an Aristotelian can simply ignore his objection as unproven. The Aristotelian can do this by arguing that, just as there is a trump card in conflicting situations between justice and mercy (sometimes it is justice, sometimes it is mercy), perhaps seeming conflicts with honesty and kindness function in the same way.

It is possible, however, that the Aristotelian will feel differently here about who has the burden of proof in demonstrating meaningful distinctions about the relationship between the mentioned virtues. It may be that Slote would simply respond that his example of the tact/honesty scenario is sufficient to show a meaningful difference, and that any claim otherwise needs to be followed with an argument. Though I disagree that Slote's objector has the burden of proof here, let us assume that Slote is right, and thus the theist needs to look elsewhere for a response to the earlier argument; perhaps Option 2 will provide a stronger reply to Slote.

Option 2

It is important at this point to remember the implications of partial virtues; they are a duet of virtues whose natural opposition to each other leads to the possibility of circumstances in which acting from both virtues is appropriate (or called for) but impossible, and that acting contrary to one of the virtues in such a circumstance produces moral blame for the agent. With this in mind, the next option for the theist is to acknowledge that Slote is correct, and that there is an actual conflict and no one right option available to the agent in question, but deny that acting contrary to honesty or kindness (or any other virtue) should confer moral blame on the agent. According to A. D. M. Walker, it may be true that scenarios like Slote's, which involve apparent conflict between tact and honesty, actually are occasions on which there is not only one appropriate response. Perhaps there are scenarios that call for truthful and tactful responses, and one must choose which virtue to respond with, or act according to whichever virtue is dominant. In cases such as these, Walker (Reference Walker2006) believes that the tactful can act kindly while regretting the evasion of being completely honest, and the frank agent can be truthful while regretting causing pain, and both will be morally admirable and not criticizable.

Slote, however, is not willing to admit the emotional response of regret into the inventory of appropriate responses for a perfectly virtuous agent. Or as he says, ‘the fact that Walker thinks the tactful person has reason to regret not having been frank or open and the open person reason to regret having to cause distress should give us pause with Walker's conclusions’ (Slote (Reference Slote2011), 58). Perhaps the occurrence of regret, in and of itself, is not reason to qualify an agent as being less than admirable; however, Slote's point is that if the occurrence of regret ‘rises to an explicitly ethical level’, then there is reason to think the agent experiencing such regret has acted less than ethically optimal. Slote believes his example of tact/truthfulness is such an example.

Perhaps this feeling of regret is sufficient to indicate that an agent is less than perfectly virtuous, but is the feeling of regret an appropriate emotional response? A theist might respond to both Slote and Walker that an agent who has to choose tactfulness over honesty may have nothing to regret. Robert Adams speaks about the importance of discretion, tact, and even keeping secrets as being an ‘unrivaled value in human communication’; because such things as discretion and secret-keeping are so important for healthy social living, it may be appropriate and perhaps morally obligatory to tell lies or withhold facts at times. Adams feels like this is a commonly accepted feature of human experience that few would contradict. ‘Secrets are important. Most of us believe it is occasionally right even to tell a lie, and often imperative to avoid mentioning secrets or facts whose utterance would give offense or affect inappropriately the social dynamics of the situation’ (Adams (Reference Adams2008), 129). If there are occasions in which discretion and withholding of the ‘complete’ truth are appropriate, it seems strange that we would hold someone morally suspect in such situations. In such scenarios, we might even say that such an individual was acting admirably.

So, to catch up, Walker does not seem to be denying that such scenarios which Slote describes are possible, that there are possible scenarios in which virtues conflict; further, Walker agrees that such scenarios may require the agent to act in a way that is regrettable. However, Walker does not agree that such scenarios require the virtuous agent to act in a way that confers moral blame when they are forced to act contrary to a given virtue, even though the agent does have regret for acting contrary to the particular virtue. So Walker does not think such scenarios demonstrate that an agent is acting in a less than perfectly virtuous way. Slote, on the other hand, who seems to think regret implies blame, believes that since such scenarios show that an appropriate response to acting contrary to virtue (when the virtue is required) is a moral remorse or regret, then such remorse indicates that the agent is blameworthy and thus less than perfectly virtuous. Adams's response is directed more to Walker than Slote, in that he claims that such situations as Slote envisions should not necessarily evoke remorse or regret in the agent who has to sacrifice acting according to one virtue in order to act according to the other.

I think we should agree with Slote that, if the agent feels regret in his or her choice to perform a given action, then the agent is morally blameworthy or criticizable in some relevant way. However, it is important to point out that the emotional response or attitude of regret, by itself, is not really a moral attitude (Williams (Reference Williams2006), 196). It is possible to regret a host of occurrences without any of them having much to do with the moral notion of obligation. I can regret the temperature level, the truth of the current president, or the fact that I've had few opportunities to increase the aesthetics of my home; however, none of these have much moral significance for particular actions I ought to have performed. Given this point, I think we can agree with Slote as long as he understands ‘regret’ to have risen to the ethical level of ‘remorse’ or ‘self-reproach’ (Williams (Reference Williams2006), 196). In other words, the agent who feels regret, must also have remorse for failing to perform an obligatory act; he must ‘blame’ himself for his failure.Footnote ¹¹

What I am less confident about is deciding one way or the other concerning the appropriateness of the response of remorse. If remorse and guilt are so closely aligned, and the feeling of guilt implies an action or motive (in our case it is both) that is less than moral, then feeling remorse implies acting (with the relevant motivations) immorally. However, if the agent does not believe he or she acted wrongly in having to be dishonest or unkind (as Adams believes), then there should be no remorse. It seems that Slote's only response to this would be to argue that the agent has acted immorally, and thus the agent should feel remorse. So the disagreement seems to be based on whether it is ever morally permissible to act contrary to a given virtue when that virtue is appropriate for the situation the agent finds himself or herself in. I imagine the robust Aristotelian conception of the virtues that Slote is assuming, and attacking, doesn't allow for such permissibility, though Adams's conception of the virtues might. While I am by no means convinced the Aristotelian position that Slote is reacting to is the more consistent (coherent, accurate, etc.) position on the nature of the virtues, or that Slote is right that a necessary feature of this position is that agents should experience moral remorse in the context of partial virtues conflicting, let us assume at this point that the Aristotelian picture he provides is the more coherent, and that Slote is right that the appropriate emotional response to the failure of acting according to a virtue that is called for in a situation is a remorse that rises to an ethical level. Again, it is not obvious that Slote is correct, but I'd like to assume that he is in order to propose an additional strategy for the theist in responding to Slote. We turn to this point next.

Option 3

The last option for the theist who wants to affirm the doctrine of impeccability is to concede defeat to Slote, to acknowledge the existence of partial virtues and the possibilities of difficult scenarios such virtues produce, but to argue that the existence of such virtues causes no problem for the doctrine of impeccability. To pursue this option, a theist might say that rather than entailing that an agent is perfectly virtuous, impeccability may only entail being significantly virtuous. Let us say an agent can be significantly virtuous in one of two ways:

or

So for the significantly virtuous agent under SV1, a trait, say courage, will be applicable across all possible scenarios, those that the agent is familiar with, and those he isn't. However, in situations in which courageous action is appropriate, it is not impossible that the agent could fail to act courageously; it is only probable that he will act courageously in each scenario in which such actions are appropriate. For significant virtue under SV2, an agent with the virtue courage can be guaranteed to act courageously in each situation that calls for courageous action. However, the ‘situations’ that the agent can be guaranteed to act courageously in should not be understood as all possible situations; rather, the situations in which courageous action can be guaranteed are only those which the agent is already conditioned to handle (situations not far beyond his normal experiences).

To posit that impeccability merely implies an agent be significantly virtuous appears to solve, or at least aid, two of the theist's largest problems. First, Slote's arguments appear irrelevant to the claim that significant virtue is possible. If partial virtues exist, the scenarios they make possible (in which two virtues conflict) would cause no problem for either sense of significant virtue. For SV1, the theist can claim that there is no guarantee that the veracious and kind agent will always act honestly and tactfully in all possible scenarios; the situation Slote envisions may simply be one of those situations in which the virtuous agent fails to act truthfully (or kindly). For SV2, the theist can claim that while the veracious and kind agent will always act honestly and kindly in each situation in which such virtues are called for, the extent in which virtuous action can be guaranteed only extends within a limited sphere. Thus the scenario Slote envisions may simply be a situation outside the sphere the virtuous agent is capable of handling.

Second, the conditions for impeccability are not damaged if impeccability merely implies significant virtue. Note that my conception of impeccability entails two features: the agent cannot perform sinful or morally wrong actions, and the agent must have a heavenly character; it is the ‘heavenly character’ that I was intentionally vague about in my introduction, and it is this feature that the options of ‘perfect virtue’ and ‘significant virtue’ have tried to describe. The perfect virtue option explained the first feature of impeccability (deontic condition) by positing the second feature: the impeccable agent will never sin because their perfectly virtuous character makes such acts impossible (it is this form that Slote's argument caused a problem for). The SV2 option is similar to the perfect virtue option here in that it claims the impeccable agent will never sin because their significantly virtuous character makes such acts impossible.Footnote ¹² Thus the theist who wants to affirm that impeccability merely implies significant virtue, like the theist who wants to affirm that impeccability entails perfect virtue, can claim that it is the impeccable agent's character that entails the inability to perform sinful actions. The only difference between the two positions is that the perfect virtue advocate will argue the impeccable agent's character needs to entail the inability of sinning in all possible situations; the significant virtue advocate will argue that the impeccable agent's character merely needs to entail the inability of sinning with a limited sphere of possible of experiences, and this limited sphere of experiences is simply those possible experiences that make up heaven. He'll add that a character that precludes sinful actions in all possible situations is unnecessary for the heavenly agent; it is sufficient that the heavenly character merely preclude sin while in heaven.

The SV1 option is a little different in this regard from the SV2 and perfect virtue options. While still affirming that the impeccable agent cannot sin, it does not affirm that the impossibility of sin is to be solely explained by the agent's virtuous character; the impeccable agent's character merely makes it unlikely that the agent will sin, or fail to act virtuously in a situation in which virtuous action is appropriate. Thus this form of significant virtue has the problem of finding some other explanation for why the impeccable agent is necessarily unable to sin. This apparent hurdle, however, has not historically been a problem for theists; many have argued that the explanatory factor that prevents heavenly agents from sinning is simply God himself.Footnote ¹³

Now even if the issues raised by Slote's argument are solved for the theist by the claim that a heavenly character merely implies a significantly virtuous character, the idea that heavenly agents are merely significantly virtuous does have its own problems. Someone might object, for instance, that what I call a significantly virtuous agent really doesn't look all that different from a really virtuous agent outside heaven, and that what I am calling a ‘heavenly character’ is not really distinct from a ‘really virtuous non-heavenly character’. History has recognized a considerable number of really virtuous and saintly men and women; surely some of these moral individuals qualify for SV1 or SV2.Footnote ¹⁴ Others may simply object based on the intuition that heavenly agents are supposed to be morally superior to non-heavenly agents; if non-heavenly agents can have heavenly characters, it appears that this intuition is misplaced. Thus these objections can be summed up as follows. First, the title ‘heavenly character’ appears to be a misnomer; agents outside heaven seem to be able to have heavenly characters, and thus such descriptions are ultimately unhelpful. Second, the intuition that those agents fitted for heaven are morally superior to agents outside heaven appears to be false; whatever relevant distinctions exist between heavenly and non-heavenly agents, moral character does not appear to be among them.

In response to the first problem, I am not sure why it is problematic if a minority of agents outside heaven have possibly had heavenly characters. If there are agents outside heaven who have characters indicative of SV1 or SV2, I think it is wrong to say they have ‘really virtuous non-heavenly characters’; rather, it seems more helpful, and more accurate, to say they are just non-heavenly agents with ‘heavenly characters’. I do not think it is a misnomer to refer to the characters of such agents as ‘heavenly’ since that term would correctly describe the entirety of characters for those human agents in heaven. Whereas I think it would be incorrect to say either that some human agents in heaven have less than ‘heavenly characters’, or that the majority of human agents outside heaven have or had ‘heavenly characters’; if either of these were true then I think it would probably be incorrect or unhelpful to use the term ‘heavenly characters’ for agents in or out of heaven.

Also, I am sympathetic to the intuition that heavenly agents are morally superior, in some way, to non-heavenly agents. But I do not think much is lost to say that heavenly agents are morally superior to the majority of non-heavenly agents, or perhaps to say that a minority of non-heavenly agents are morally superior to the majority of non-heavenly agents, such that the minority is sufficiently fit for heaven.

Some may not be satisfied with this response. Some may think there is something about both forms of significant virtue that implies, perhaps subtly, that the line separating heavenly characters from non-heavenly characters is way too thin. Take the example from Tim Pawl and Kevin Timpe about an especially vicious husband who is somewhat quarantined by his wife in order to preclude opportunities for sin:

An objector may say that this example highlights an important problem for both forms of significant virtue, in that the story seems to indicate a large amount of commonality between someone with an especially vicious character (Smith) and someone with an allegedly heavenly character. In the example, the reason for Smith's inability to sin appears to have more to do with the situation or circumstance Smith finds himself in (his wife's supervision), rather than with the actual virtuous character that Smith has (or doesn't have). In other words, Smith's inability to sin is better explained by his situation than his actual character. An objector might say that this appears to be true for the heavenly agent under SV1 and SV2 also.

For the SV1 agent whose dispositions for virtue are broad-based, and thus cross-situational, the lack of sinful behaviour is based on probability rather than necessity; thus, if the SV1 agent is guaranteed not to sin in heaven, an at least partial explanation for the lack of sin must have something to do with events or circumstances external to the actual agent. This also seems to be the case for Smith. For the SV2 agent in heaven, his dispositions for virtue are narrow-based, indicating that the agent is never guaranteed not to sin whenever he is outside his normal array of circumstances. An objector might claim the same could be said of Smith. Let Smith wander outside his wife's supervision, or let the SV2 agent outside heaven, then there is no guarantee that either won't sin.

I think it is important to point out here, in response, that even if a significant explanatory factor for the absence of sin in heaven is the situation or circumstance which is somewhat external to the agent, there still appears to be an important difference between Smith and the SV1 or SV2 agent: virtue! To qualify for an impeccable or heavenly character, the agent must have all the relevant virtues, whether those virtues are understood broadly or narrowly, and whether they are dependently or independently related. For both SV1 and SV2, the agent is assumed to have a sufficiently adequate virtuous character, at least as to qualify for traditional virtue ethics.Footnote ¹⁵ This cannot be said of Smith.

Also, I am not sure why it is such a problem to say that, in addition to the SV1 or SV2 agent's character, a significant explanatory factor for the absence of sin in heaven is the situation in heaven.Footnote ¹⁶ This does not appear to be a completely foreign position in Christian tradition; most Christians would not only acknowledge, but emphatically affirm, that there are certain properties or facts about God that are in some way causally relevant to the disposition, and therefore behaviour, of the heavenly agent. Take the following passage from Jerry Walls as an example; here Walls is reiterating a point made by Augustine about the very nature of the heaven:

Here we see the idea that a proper relationship with God, in a proper environment, allows one to respond to various features of God (his nature, his presence, his acts of power and grace, etc.) with fascination, love, and gratitude. Obviously a certain type of character is necessary to be properly related to God (certain attitudes, desires, and beliefs), but I see no reason why the agent's character has to be the sufficient explanation for the lack of sin. Surely being in God's presence and being so aware of his divine majesty, glory, and love will be a factor causally relevant enough so as to preclude the possibility of ever turning away from or rejecting such a being (i.e. sin).Footnote ¹⁷

Perhaps someone may feel that if a position like SV1 or SV2 is adopted, all notions of libertarian freedom and moral responsibility are lost for those in heaven. Under SV1, since the agent's character merely makes it probable that the agent won't sin, there is still the need for God, like Smith's wife, to ensure that the heavenly agents are never in situations in which their character is insufficiently strong enough to withstand sin. Under SV2, it seems as though, even if it is one's character that precludes the possibility of sin, it still could be said that a necessary condition for the state of one's character are external features from his or her environment (i.e. God's presence or nature). Thus under both SV1 and SV2, it appears as though something external to the agents (God) plays a causally significant enough role in the actions of the agents to engender doubt that they are actually free and responsible in a libertarian sense.

Though I understand why someone may be motivated to make this objection, I do not think it is very strong. If one is concerned about maintaining a libertarian stance, an adoption of a source-incompatibilist position of agency provides resources to ensure that both the SV1 and SV2 agent in heaven are free and responsible. According to source incompatibilism, to be free and responsible for a given action, two things must be the case: (a) determinism must be false, and (b) nothing external to the agent in the causal history of a given act can be the sufficient cause of the act (i.e. the agent must be the ultimate source of the act).Footnote ¹⁸ Both of these conditions can be met for both SV1 and SV2. A theist can simply affirm that causal determinism is false, and that there is nothing external to the agent prior to heaven that sufficiently caused the agent to trust in Christ or cultivate a virtuous character; the theist could even say that, prior to heaven, the agent had access to alternative possibilities when he or she made a choice to go to heaven (i.e. trust in Christ) and also to cultivate a virtuous character. If a necessary condition for entering heaven is that the agent must trust in Christ and cultivate a virtuous character, and do both according to libertarian conditions, and these conditions are met, then it is somewhat irrelevant what causal features are in place in heaven that constrain the agent's actions (whether it is the agent's character or God himself).Footnote ¹⁹ Even if the agent in heaven is constrained in their acts by something external (God) to them, they are still rightly deemed free and responsible for the actions there given that the causal history of the acts in heaven can be traced back to a point prior to heaven in which nothing external to the agent was the sufficient cause of the choices and actions that led to the acts the agent performs in heaven (Sennett (Reference Sennett1999); Walls (Reference Walls2002), 61; Pawl & Timpe (Reference Pawl and Timpe2009) ).