Managing shame and guilt in addiction: A pathway to recovery

A dominant view of guilt and shame is that they have opposing action tendencies: guilt-prone people are more likely to avoid or overcome dysfunctional patterns of behaviour, making amends for past misdoings, whereas shame-prone people are more likely to persist in dysfunctional patterns of behaviour, avoiding responsibility for past misdoings and/or lashing out in defensive aggression. Some have suggested that addiction treatment should make use of these insights, tailoring therapy according to people ’ s degree of guilt-proneness versus shame-proneness. In this paper, we challenge this dominant view, reviewing empirical findings from others as well as our own to question (1) whether shame and guilt can be so easily disentangled in the experience of people with addiction, and (2) whether shame and guilt have the opposing action tendencies standardly attributed to them. We recommend a shift in theoretical perspective that explains our main finding that both emotions can be either destructive or constructive for recovery, depending on how these emotions are managed. We argue such management depends in turn on a person ’ s quality of self-blame (retributive or ‘scaffolding ’ ), impacting upon their attitude towards their own agency as someone with fixed and unchanging dispositions (shame and guilt destructive for recovery) or as someone capable of changing themselves (shame and guilt productive for recovery). With an eye to therapeutic intervention, we then explore how this shift in attitude towards the self can be accomplished. Specifically, we discuss empathy-driven affective and narratively-driven cognitive components of a process that allow individuals to move away from the register of retributive self-blame into a register of scaffolding ‘reproach ’ , thereby enabling them to manage their experiences of both shame and guilt in a more generative way.


Introduction
Guilty baby I'm guilty And I'll be guilty the rest of my life How come I never do what I'm supposed to do How come nothin' that I try to do ever turns out right? -Randy Newman Addictive behavior is notoriously hard to change. Although in the recent decades there has been an increased focus on the neurobiological mechanisms underlying addiction, treatment often still focusses on the psycho-social dimensions of addictive behaviour encompassing motivation and other mental states, both cognitive and affective (Lewis, 2015).
Since the 1990s a stream of literature has suggested that the emotions people experience in the wake of their (mis)doings are importantly related to outcomesin particular, that feelings of guilt can play a supportive role in changing behavior in a constructive way, while feelings of shame can be detrimental to such development (Leith & Baumeister, 1998;Baumeister, Stillwell, & Heatherton, 1994;Lewis, 1971). A recent revival of these theories on shame and guilt examines the relevance of these emotions for treatment of substance dependent people (Dearing, Stuewig, & Tangney, 2005;Treeby, Rice, Cocker, Peacock, & Bruno, 2018).
In this article, we will elaborate on the role of feelings of shame and guilt in maintaining and breaking with addictive behavior. We begin by outlining the dominant theoretical perspective according to which guilt and shame play oppositional roles, one productive and the other counter-productive, in moral motivation, self-control, and recovery more general. However, this dominant perspective has not gone unchallenged; and we review some methodological considerations that ground a reasonable scepticism regarding the clean separation of shame and guilt with respect to these purported roles. Finding support for this scepticism in our own qualitative longitudinal study among 69 opioid and alcohol dependent people in Australia, we present some self-report data that reveal a more complicated picture of how these emotions relate to addiction and recovery. In order to make sense of these results, we next introduce an alternative theoretical framework for understanding shame and guilt. This framework rejects the simple opposition proposed for how these emotions operate in people's moral psychology, arguing instead for a critical distinction between retributive and scaffolding experiences of each of these emotions sourced in an individual's understanding of their own agency and relationship to themselves (McGeer, 2020). This theoretical shift in perspective suggests that a more constructive response to these emotions does not involve eradicating shame in favour of guilt, but rather developing agential resources for managing both of these emotions in a positive 'scaffolding' way. We conclude the paper by drawing out some implications of this alternative approach for addressing the problem of recovery from addiction.

The dominant theoretical perspective on guilt and shame
It is widely accepted that guilt and shame can be distinguished along two critical dimensions: first, with regard to their intentional focusfeatures of the situation made salient to the individual experiencing these emotions; and, secondly, with regard to their motivational profile (Dearing et al., 2005;Leith & Baumeister, 1998;Lewis, 1971).
With respect to their intentional focus, the general consensus is that guilt highlights features of the individual's action, whereas shame highlights features of the self. (Lewis, 1971). A person feeling guilt who misses a meeting because of a hangover might think 'I feel bad for drinking too much and causing inconvenience to other people', while a person feeling shame might think 'I am unreliable'. Moreover, while both emotions are triggered by transgressions, feelings of guilt seem to arise in connection with a perceived violation of socially endorsed norms or principles, whereas feelings of shame are more closely associated with a perceived violation of one's own values or ideals (Flanagan, 2013). Additionally, guilt is associated with other-directed feelings of empathy, whereas shame seems to interfere with empathy (Leith & Baumeister, 1998;Baumeister et al., 1994;Tangney, 1991Tangney, , 1995b. Hence, guilt is claimed to be a more social emotion, reflecting a person's concern with the impact of their behavior on others, while shame is regarded as a more self-centered emotion reflecting a person's concern with how others view them. Related to this difference in intentional focus, guilt and shame appear to have quite different motivational profiles. Research suggests that shame funds two types of anti-social reaction, which may occur separately or together: (1) a desire to withdraw from others and/or avoid anything related to shame-inducing events; and (2) increased aggression towards others in a defensive attempt to shift blame for such events. Guilt, by contrast, appears to fund more pro-social behaviours, including the desire to confess, apologize, and take reparative and corrective action (Leith & Baumeister, 1998;Lewis, 1971;Lindsay-Hartz, 1984;Tangney, Wagner, Fletcher, & Gramzow, 1992;Tangney, Miller, Flicker, & Barlow, 1996;Tangney, Wagner, Hill-Barlow, Marschall, & Gramzow, 1996;Wicker, Payne, & Morgan, 1983;Tangney, 1998). Hence, feelings of shame are argued to be debilitating and selfdefeating, whereas feelings of guilt are seen to provide a positive impetus for constructive change (Dearing et al., 2005).
Theorists argue that the differences between these two emotions are only magnified if they become dispositionally entrenched. Research suggests there are individual differences in affective style such that, in similar situations, some people are more likely to experience shame, whereas others are more likely to experience guilt (Harder & Lewis, 1987;Harder, 1995;Tangney, 1991;Tangney, Burggraf, & Wagner, 1995). This is claimed to have knock-on psychological and behaviour effects. As shame-prone individuals are dispositionally inclined to a negative view of themselves, they are more likely to experience on-going difficulties with self-esteem, anger-management, and self-regulatory strategies in general. By contrast, guilt-prone individuals, with their dispositionally-entrenched focus on the causes and consequences of their actions, are less vulnerable to the negative effects of low self-esteem and are more successful in developing strategies for affective and behavioural self-regulation, potentially mediated by their increased empathy for others (Tangney, 1995a;Tangney, Wagner, et al., 1996). 1 Pertinent to the problem of addiction, and in keeping with this theoretical outlook, research suggests that proneness to shame is a riskfactor for developing a substance abuse problem (as well as engaging in other risky behaviours), whereas proneness to guilt is protective (Dearing et al., 2005;Meehan et al., 1996;O'Connor, Berry, Inaba, Weiss, & Morrison, 1994;Stuewig et al., 2015;Treeby et al., 2018). In a questionnaire study among the general population (n = 281) in Australia, Treeby and colleagues (2018) found that guilt-proneness is associated with effective strategies to regulate alcohol use and avoid alcohol related harms, while shame-proneness is not. More speculatively, the research suggests that guilt-prone individuals with substance dependency have better prospects for recovery that those who are shame-prone (Treeby et al., 2018).
If this research is on the right track, it opens up new possibilities for treatment. It argues for focusing clinical attention on differentiating between guilt-prone and shame-prone individuals and tailoring therapeutic interventions accordingly. For instance, the moralization of addictive behavior might have a protective effect on guilt-prone people because it evokes their feelings of guilt and, hence, their motivation to change. However, moralization could be detrimental to shame-prone people in so far as it reinforces their feelings of shame and hence their self-defeating behavior (Quiles, Kinnunen, & Bybee, 2002). Shameprone people, so this research suggests, would be better served by therapeutic interventions aimed at down-regulating their feelings of shame in relation to their substance use, perhaps by replacing these with more 'constructive' feelings of guilt directed towards the impact of their addictive behavior on others as well as themselves (Treeby et al., 2018).

Sceptical challenges to the dominant view
While these therapeutic recommendations may seem theoretically sound, there are reasons to question the underlying perspective that gives rise to them, beginning with some methodological concerns arising in connection with the relevant research. Much of the empirical work that identifies guilt as a constructive emotion and shame as a destructive one measures these concepts using the TOSCA instrument (Test of Self-Conscious Affect) or its precedent, the Self-Conscious Affect and Attribution Inventory (SCAAI), both developed by Tangney and colleagues (Tangney, Wagner, & Gramzow, 1989). However, these instruments have been criticized for only measuring adaptive forms of guilt and maladaptive forms of shame (Luyten, Fontaine, & Corveleyn, 2002; 1 This is not to say guilt-prone individuals fail to have, or draw upon, a conception of themselves (something the dominant view of shame and guilt may tend to overlook). As one of us (Kennett) has argued elsewhere, while there are different ways of exercising self-control, one critical way is precisely by drawing on a (generative) conception of the self. More precisely, by seeing oneself as a certain kind of person, one can be motivated to shape one's thoughts, feelings and behaviour in a certain way (Kennett & Matthews, 2008;Kennett, 2013;Tice, 1992). However, if someone's self-image is eroded by feelings of shame, then it stands to reason that focusing on the self (with all its liabilities) will hardly support this mode of self-control. We return to this point below. Sabini & Silver, 1997;Harder, 1995). To illustrate, the TOSCA consists of scenarios describing experiences of shame and guilt, followed by a potential reaction. An example of a scenario is: "You make plans to meet a friend for lunch. At 5 o'clock, you realize you stood him up." Potential reactions include: 1) You would think, "I'm inconsiderate", 2) You would think, "I should make it up to him as soon as possible", 3) You would think, "My boss distracted me just before lunch", 4) You would think: "Well, they'll understand." Participants then have to rate on a Likert scale whether they would react in the manner described. However, the guilt items in the TOSCA only measure appropriate constructive reactions, rather than more pathological guilt responses. The shame items, by contrast, focus strongly on maladaptive aspects of shame, and neglect more adaptive aspects of this emotion (Harder, 1995;Sabini & Silver, 1997;Luyten, Fontaine, & Corveleyn, 2002).
This methodological concern gives rise to two lines of scepticism challenging the dominant view, and we discuss each of these in turn. The first questions whether shame and guilt can be usefully disentangled in real world situations; and the second questions whether shame and guilt have separate and opposing action tendencies, with different implications for recovery.

Questioning the clinical relevance of distinguishing between feelings of guilt and shame
As reviewed above, the dominant view argues for distinguishing between guilt-prone and shame-prone individuals who are substancedependent for the purpose of identifying helpful therapeutic interventions (Dearing et al., 2005;Treeby et al., 2018). However, further research shows that these emotions often occur simultaneously in the context of substance abuse, questioning the utility of making such a distinction. (Harris, 2003). Harris (2003) presented 720 drink-driving offenders with a questionnaire after they attended court or a restorative justice meeting. This questionnaire contained several statements reflecting the phenomenology of guilt, shame, and embarrassment. The results did not show a distinction between guilt-prone and shame-prone people; rather it showed that both emotions are often deeply intertwined, and often occur 'as a single response' (Harris, 2003). Harris (2003) and Blum (2008) consequently suggest that the much-heralded distinction between shame and guilt is an artificial one that does not hold up in real life situations. Echoing this, Sabini and Silver argue that it is artificial to distinguish guilt and shame as Tangney and colleagues do. Character and action cannot be separated that neatly: "Acts for which we are guilty (and don't just in some weak sense feel guilty), do involve the self. And because they involve the self they involve shame." (Sabini & Silver, 1997, p.8) Examining the studies of Dearing et al. (2005) and Treeby et al. (2018) more closely, it seems that they also struggle with this intertwinement. Using the TOSCA, these studies distinguish not so much between shame and guilt, but between shame and shame-free-guilt, allowing for the possibility of a third category: people who experience both guilt and shame (Dearing et al., 2005;Treeby et al., 2018). But as these studies do not specify how many people experienced only shame, rather than a mixture of both shame and guilt, no firm conclusions can be drawn about the extent of their potential entanglement. Other studies suggest it is substantial. At the very least, as Harder (1995) notes, the TOSCA does not sufficiently reflect the relationship between guiltproneness and negative self-evaluation.

Questioning the negative action tendencies associated with shame
Even if these emotions are often intertwined, the dominant view still holds that feelings of shame are detrimental to recovery, while feelings of guilt herald positive change. A second line of criticism challenges this claim.
In the literature focussed more narrowly on addiction, both Flanagan (2013) and Sawer, Davis, and Gleeson (2020) have contested the claim that feelings of shame are detrimental to recovery. They both resist the claim that a negative evaluation of the self necessarily leads to selfdefeating behaviour, arguing that it can also promote attempts to redeem oneself. Flanagan argues that shame can play an important role in making people reflect on their normative failures: their failure to control their behavior and their failure to live up to the standard of a good life that they set for themselves (Flanagan, 2013). Such shame can play an important role in recovery if it is not accompanied by moralization; the idea that addiction is a moral failure. Following Pickard's concept of responsibility without blame (Pickard, 2017), Flanagan proposes a shame without blame (we discuss Pickard's approach in more detail in Section 5 below). In their study of the narratives of people in recovery Sawer et al. (2020) likewise found that shame can play an important role in recovery when people found a safe place to talk about it and make sense of their feelings of shame.
Miller (2003) shows how a negative evaluation of the self can lead to a positive change of behavior. He describes the following true story: A heavy smoker is about to pick up his children from school. It is raining heavily. When he is almost there, he notices that he is out of cigarettes, and although he sees his children standing in the rain waiting for him, he changes direction to buy cigarettes. But just at that moment, he is struck by the following thought: 'I don't want to be the kind of person who leaves his children standing in the rain to buy cigarettes'; and he subsequently quits smoking (Miller, 2003;Premack, 1970). Although in this example the smoker doesn't say he is ashamed, he certainly evaluates himself negatively, presumptively with a strong degree of negative affect: 'I am the type of person that leaves his children in the rain to buy cigarettes'. However, instead of dwelling on what these thoughts and feelings indicate about his character, he decides that he does not want to be this type of person and successfully changes his behavior.

Empirical findings from our own study: questioning the constructive role of guilt
Findings from our own qualitative study support these sceptical challenges to the dominant view. But in this section we go further in questioning the positive motivational force the dominant view associates with guilt in particular.
Two of us (AS & JK) did a qualitative longitudinal study on how people with opioid and alcohol dependency in Sydney, Australia, saw their self-control and agency (n = 69). During the first interview, we used a timeline to support the respondents in narrating their life stories (Adriansen, 2012). At the end of the interview, we asked them to elaborate on their goals for the next year and the future in general. We did follow-up interviews four times over three years. Each interview, we recapitulated their goals from the former year, and asked them how their year has been, and what got in the way of reaching their goals, whether their goals changed and why. All interviews were recorded, transcribed ad verbatim, and analysed in Nvivo. Respondents were mostly from low socio-economic backgrounds.
The focus of this study was addiction, self-control, and recovery in general, with results extensively published and discussed elsewhere (Kennett, Matthews, & Snoek, 2013;Snoek, Levy, & Kennett, 2016;Snoek, 2017;McConnell & Snoek, 2018). But while the study was not specifically designed to examine feelings of guilt and shame, we were struck by the role self-reported feelings of guilt in particular played in maintaining addictive behavior. We present these findings here.
Feelings of shame and guilt are often reported by people with substance dependency (O'Connor et al., 1994;Meehan et al., 1996;Flanagan, 2013;Ehrmin, 2001). These feelings were also extensively reported in our study. Respondents regarded shame as detrimental to recovery, which seemingly supports the dominant view discussed above: I think the biggest thing probably would be to try and cut out the shame around it and to you know seek support or … one thing I found really helpful is the internet and Tumbler and different forums and just reading about people … people's own experiences with any kind of addiction or … yeah, anything like that just to normalise it a bit more and cut out that isolation and shame, I think that's … that's hugely important or has been for me. (R17C) Try not to be ashamed, I would say. Because you've got … they've got enough problems as it is, just if you can cut the shame out, that stops … I think it [feelings of shame] stopped me from getting better for a long time because I couldn't address it properly because I felt so guilty. (R22) However, contrary to the dominant view, self-reported feelings of guilt did not help them change their behaviour for the better. On the contrary. Based on our coding system, one of the nodes that appeared in several interviews was 'guilt reinforces use'. Thus, feelings of guilt, so far from igniting behavioural change, appeared to result in avoidance by reintoxicating oneself, in order to not feel guilt:

I don't want to feel, I don't want to think because when I think and feel I'm that guilty now, the guilt and shame of the last four or five years, I haven't been there for my kids, I haven't even spoke to them, I don't know what's going on, I haven't seen a picture, there's so much guilt there and shame and anger and all that that I just think, stuff it I just want to get off my face so I don't have to feel that pain. (R58) I just beat myself up because I'd stuffed up so many times with things. That's why I drank as well, it wasn't to self-harm myself, it was just to, like I say, get drunk and stop thinking about what I'd done wrong and where I went wrong. (R12) It's not like: 'am I going to do it, I'm not going to do it', it's like: 'no I'm just doing it' and then I wake up in the morning and go 'oh what have I done, oh I'll just have another drink.' (R10)
In the quotes above, the respondents explicitly target their guilt feelings, as these are associated with their actions (not trying to connect with their children, stuffing up many times); they are primarily focussed on what they have done. In the quotes below it is less clear whether the respondents refer to guilt or shame as discrete emotions or, in support of the entanglement thesis, some mixture of the two. However, judging from these self-reports, both emotions are clearly experienced as reinforcing use: For these respondents, substance use seems to be a way to selfmedicate against feelings of shame and guilt alike, rather than motivating them to make amends. In sum, contrary to the dominant view, our respondents did not give a central place to guilt feeling in their recovery. Given that our study used self-report, it might be objected that people often use the words guilt and shame interchangeably in everyday language (Leith & Baumeister, 1998). So it is unclear whether our respondents are guilt-prone or shame-prone. However, when reviewing qualitative studies, both Tangney (1995a) and Harris (2003) found that shame narratives were often followed by references to the self, and guilt narratives by descriptions of actions (Lindsay-Hartz, 1984;Niedenthal, Tangney, & Gavanski, 1994). This is evident in the quotes from our study as well. So, once again, our evidence indicates that feelings of guilt can be just as counter-productive for recovery as feelings of shame; we see no evidence that feelings of guilt are particularly ameliorative in this sample.
In sum, we think our own empirical findings coupled with the sceptical material we reviewed in Section 3 are sufficient to question the theoretical validity of the dominant view and its utility in guiding therapeutic interventions aimed at supporting recovery from addiction. The relevant challenges from these two critical sections can be summarized as follows: (1) guilt and shame often constitute an intertwined emotional response to perceived self-failure, casting doubt on the feasibility of eliminating the 'shame' strand of this response; (2) feelings of shame are not always, and so not necessarily, detrimental to recovery; and, finally, (3) feelings of guilt are not always, and so not necessarily, conducive to recovery. But these challenges also suggest that we need a new theoretical framework for understanding how both of these emotions can have either negative or positive effects in a person's moralpsychology. In the remainder of this paper, we offer such a framework and discuss its implications for addiction and recovery.

Two styles of 'self-blame': retributive and scaffolding
One of us (McGeer) has proposed an alternative framework for understanding the complex working of shame and guilt in people's moral psychologyin particular, why these emotions can sometimes exert a positive motivational force supporting self-regulation and recovery, but nonetheless often fail to do so (McGeer, 2020). We outline this framework in the next two sections, using important contributions from Brandenburg (2019) and Brandenburg and Stribjos (2020) to emphasize and enhance its central features.
In keeping with previous work on shame and guilt, we agree that these emotions can be differentiated in terms of their foci of attention: in feeling guilt, subjects are more directly focussed on problematic aspects of their behavior, whereas in feeling shame, they are more directly focussed on problematic aspects of themselves (their attitudes and dispositions). But we disagree that focusing attention in these different ways is in itself responsible for the opposing motivational profiles theorists have come to associate with these emotions. We do not deny that feelings of shame can motivate a range of anti-social and self-defeating attitudes and behaviours (avoidance, withdrawal, denial, deflecting blame on to others, and defensive aggression), but so too may guilt. And, likewise, we find that shame and guilt can both be implicated in motivating more pro-social and self-supporting attitudes and behaviours (apology, making amends, committing to do better, learning to trust oneself and others, developing compassion, etc.). Hence, we conclude another factor is at work in sustaining an anti-social versus pro-social motivational profile.
That factor, we think, is blame. Shame and guilt are natural emotional responses to being the target of another's blameto be judged by them and to be found wanting. But we don't need others to be the target of someone's blame. We can blame ourselvesand when we do, we endorse the view that we are at fault, intensifying the shame and guilt we feel in consequence. But the intensity of these emotions is not the thing that really matters in terms of how well we manage them. What matters more is the manner or mode in which we blame ourselves.
But what exactly do we mean by 'blame', whether directed towards the self or directed towards others? Blame is generally understood to involve a critical stance of holding responsible. But its manifestations are various, leading philosophers and psychologists into vigorous disputes about its essential, or at least canonical nature, with these disputes funding substantial differences in how the term is used. 2 For instance, one sharp point of disagreement concerns whether blame essentially involves having (or expressing) hostile feelings, thought and attitudes towards the person blamed. Hanna Pickard (2011) thinks that it does, and argues for a more supportive form of holding responsible that does not involve blame. In her view, adhering to a practice of 'responsibility without blame' is particularly important in therapeutic contexts, where hostile responses to persons who struggle with addiction or personality disorder, reinforce stigma and undermine their recovery. But, as Pickard argues in collaboration with Nicola Lacey, a more general practice of 'responsibility without blame' would be a very good thing since more detached (i.e. non-hostile) ways of holding people accountable to norms is a better, more empowering way of encouraging norm-compliance (Lacey & Pickard, 2013). Brandenburg and Stribjos (2020) critically expand on the framework of responsibility without blame, adding that our practices of holding accountable need not be detached in order to be conducive of recovery. Indeed, the types of retributive blame that Pickard describes fail to facilitate recovery, but other types of reproachful engagement can be helpful. 'Reproach' is here introduced as a term of art to capture nonhostile forms of negative affectivity often associated with blame; forms that seek to engage rather than exclude or punish. 3 There are a number of reasons for promoting moderate responses of this type towards the moral failures of others. One reason is that within engaged human relationships, it is hard or even impossible to avoid being negatively affected by perceived moral mistakes or failures. Hiding such feelings can be misleading and confusing to others. A controlled and moderate expression of these responses signals that you stand in an honest and transparent relationship to the person (Rogers, 1957). At the same time, having such responses is often indicative of the fact that one respects the other person and cares about the wrong that occurs (Strawson, 2008;McGeer, 2011). Therefore, a controlled and constructive reproachful response may engage others in a way that displays your respect for them and concern about what they do (Brandenburg & Strijbos, 2020).
Another important reason for engaging in constructive reproachful interactions is setting an example for how to deal with moral failures, whether these are the failures of others or our own. Holding responsible is a social activity; but it is an activity we internalize from our wider culture and, importantly for this paper, come to direct towards ourselves. A practice of constructive reproachful engagement, not only with others, but with ourselves shows how it is possible to reflect on, and respond to, our own moral failures in a way that constructively manages the negative self-directed emotions such failures typically bring in their wake (we return to the topic of 'self-blame' below) (See Brandenburg & Strijbos, 2020).
One further point deserves emphasis. This reproachful style of response need not presume that a failing agent is 'blameworthy' for their behavior in the standard sense (Brandenburg, 2019). The question whether the person's transgression was entirely voluntarily, expressed ill will towards others, or whether they could (in some sense of the word) have done otherwise, is bracketed and can be set aside when one engages in these more forward-looking reproachful responses to the person. The agency that is implied in such engagement is explicitly developmentali.e. it is the ability to grasp and constructively respond to these forms of reproachful engagement (see too: McGeer, 2011McGeer, , 2018b. In response to Pickard it should be mentioned that a number of philosophers have proposed and defended accounts of blame that are consistent with the type of reproach described here, McGeer among them (see, too, Fricker, 2016, McKenna, 2012, Macnamara, 2015. Drawing on a range of empirical findings, McGeer and colleagues have argued that affectively-laden blame can occur in one of two modes: 'retributive' and 'scaffolding' (Funk, McGeer, & Gollwitzer, 2014;McGeer, 2013McGeer, , 2018aMcGeer & Funk, 2015;McGeer, 2020). Retributive blame focusses on simply condemning people for their faults. It is exclusively backward-looking (concerned with past doings); it is punitive in intent; and it is non-dialogical in structurei.e., it invites no response from the person blamed other than accepting condemnation as their due. Scaffolding blame, by contrast, is focused on calling people to account for their faults or failures. While it takes account of past behavior, it is primarily forward-looking (concerned with potential future doings); it is transformative in intent; and it is essentially dialogical in structurei.e., it calls on the person to make amends where amends are due; it calls for apology, self-examination, self-adjustment and reform. Further, because scaffolding blame insists upon these things, it conveys a hopeful message to the person blamed despite its critical stanceviz., that they have what it takes to understand and amend their behavior going forward; that they are agents of potential and not just agents with unchanging and unchangeable faults. To indicate the lack of punitive, hostile intent associated with this form of affectively-laden blame, we shall henceforth refer to it as 'scaffolding reproach '. 4 Retributive blame and scaffolding reproach are very different in character, affecting how the person held responsible experiences guilt and/or shame in consequence: as indicative of standing faults, whether of behavior or character, that are fixed in time; or as indicative of faults or failures that can be understood, addressed and changedagain, whether these are faults and failures of behavior or character. Unsurprisingly, blame in the retributive mode is counter-productive in fostering positive change. Indeed, it is fundamentally disempowering, since 'taking responsibility' from the retributive point of view primarily involves the passive acceptance of deserved condemnation and punishment for past deeds; it does not call for or support any pro-active efforts on the target's part to engage in serious self-examination and lifestyle improvements going forward. So why blame in the retributive mode?
Unfortunately, this style of blame comes naturally to human beings. In fact, both styles of blame come naturally to us, often co-occurring in our psychology (McGeer & Funk, 2015;McGeer, 2013). So the problem is one of transforming the urge to engage in retributive blame into a more hopeful, forward-looking determination to engage in scaffolding reproach: moral criticism that takes the business of transformation and development seriously. This is challenging enough when a person's blame is directed towards others. But it is particularly challenging when 2 Indeed, the authors of this paper are not of one mind on this topic, so what follows is an attempt to emphasize the important themes on which we largely agree (with notes indicating minor points of disagreement) 3 Tellingly the word 'reproach' is thought to derive from the Latin repropiare one meaning of which is to bring close or into dialogue. 4 As a minor point of disagreement, one of us (McGeer) does not think strong negative affect is necessarily hostile or, better, retributive (understood as a concern with making the target suffer), so is inclined to downplay the need for some milder affective response that Brandenburg and colleagues associate with 'reproach'. Still, what matters for our purposes and where we are in full agreement, is that one can have negative emotional responses to one's own or other's faults and failures that are not retributive or hostile, but facilitate the forward-looking aims associated with responsibility: self-examination, adjustment, and reform or development. Importantly, as we will shortly argue, these are the sorts of responses that, when directed towards the self, explain how guilt and shame can play a positive role in self-improvement and recovery. a person engages in self-blame; and all the more so when they view their faults as serious and persistent (Stichter, 2020). Here the path of least resistance is to embrace a retributive, condemnatory style of blame that requires nothing more than wallowing in feelings of guilt and shame, accepting bad things that happen as one's due or even actively engaging in self-destructive behaviours (such as more substance use) as a form of deserved punishment. As Stichter explains, the cognition of one's own moral failure creates distress, which often quite naturally prompts responses like avoidance or defensiveness. These responses stand in the way of self-improvement and need to be actively moderated and managed in order to allow for the development of moral skills (Stichter, 2020). The findings from our study make sense in terms of this shift in theoretical thinking. While it might be argued that people who engage in retributive blame are more likely to feel shame than guilt, we saw no evidence of this. As noted above, respondents in our study tended to report a mixture of these feelings, with guilt figuring prominently in their experience of persistent blameworthiness and self-condemnation. Nor did we see evidence of guilt-feelings simply replacing those of shame, as individuals managed to shift into a more productive mode of self-blame. What did become evident is a reduction in the torturous experience of these emotions, as respondents learned to manage them better in the context of taking a more pro-active, hopeful, and forwardlooking stance towards their own agential capacitiesthe very hallmark of scaffolding reproach. As one respondent stated:

.. I don't feel guilt and shame anymore because that was just a waste of time 'cause that's just beating myself up and it doesn't get me anywhere it just makes me worse. So there are times when I feel, 'oh, I'm a shit daughter sometimes', but I think, 'well, making myself think that and believing it is just making it worse, there's no point …' it's what it is and what I've done is what I've done and I can't go back so there's no point. So, I think that's where my attitude's changed that … even though I did used to feel guilty and I always used to think, 'oh my God, I've wasted so much time and all that', it's not getting me anywhere by thinking or dwelling on it. I just need to move forward. (…) if you stare in the mirror and go, 'I am guilty, it's not getting me anywhere.' (…). It's not going to get me anywhere except to the bottle shop. (…) But I wrote down a list of things that I felt guilty for and then I wrote next to it another list, why did I feel guilty about it and then I wrote where's the evidence that I should feel guilty about it and then I did another list and … what can I do to change it. (R36C)
Important to this respondent is not the kind of emotion she experiencesguilt versus shame; rather what matters is how she experiences these emotions. And this we argue is shaped by her attitude towards herself: not as someone defined and fixed entirely by her past deeds, hence continuously punished by her shame and guilt; but rather, as someone capable of change and recovery, using these emotions constructively as a guide to the work she needs to do. Such work naturally leads to their replacement with more positive, normatively fitting self-directed emotions such as hope-for-oneself, self-trust and selfforgiveness (Flanagan, 2013;Sawer et al., 2020). As our respondent notes, "I don't feel shame and guilt anymore".

Scaffolding reproach: cognitive and affective dimensions
Thus far, we have claimed a more fruitful theoretical approach to understanding how people experience guilt and shame is through the lens of understanding their mode of self-blame, and in particular how this mode of self-blame impacts upon their attitudes towards their own agency. If they feel their entire identity is bound to their dysfunctional past and is thereby appropriately condemned, it matters little if they are guiltily pre-occupied with a series of bad deeds or shamefully preoccupied with their bad dispositions and bad character. Neither emotion is likely to motivate self-supporting attitudes or behaviour.
What is needed instead is a strategy to shift them out of the backwardlooking register of retributive self-blame and into the forward-looking register of scaffolding self-reproach.
We noted above that a key cognitive element in making this shift involves a changed attitude towards the self: in particular, to viewing oneself as more than the sum of dispositions and traits that led to past dysfunctional behaviour. It involves envisaging a different kind of selfnot just one with different traits and dispositions, but one that has the power to bring that self into being. In the following section, we will have more to say about how this shift in attitude is cognitively supported. But in this section, we focus more directly on the affective elements involved. However, contrary to the standard preoccupation with shame versus guilt, we highlight a distinctive affective process that is perhaps more fundamental. Specifically, we focus on the importance of empathy in supporting this shift in attitude, emphasizing in particular that it is not just empathy for others that matters; equally important is developing empathy for oneself (McGeer, 2020).
Philosophers and psychologists have long emphasized the importance of empathy in developing a properly attuned capacity for moral agencyempathy for what others experience in their situation. Of particular importance is having an empathetic response to other people's suffering, especially when we are the cause of that suffering. In fact, Baumeister et al. (1994) see this as the source of guilt feelings, noting that proneness to guilt is empirically correlated with empathy (Tangney, 1990(Tangney, , 1991. We think that shame, too, can result from an empathetic awareness of others' distress. But what it more interesting is how both of these emotionsshame and guiltcan actually block a properly-tuned empathetic response to others (and self) when they are experienced in an excessively self-punishing mode. We therefore suggest that empathetic responsiveness to others depends upon developing some empathetically mediated concern for one's own distress in experiencing shame and guilt, and a newly discovered agential power to manage these emotions more productively.
In the next section, we report a case study from our empirical work demonstrating how empathy for others and empathy for the self are codependent. In particular, we focus on how this respondent is able to develop a proper concern for others by relating to himself in a new way, no longer seeing himself as simply a victim of self-punishing guilt and shame, but as someone who can take responsibility for his emotions out of concern for how they are impacting upon himself and upon those around him.

Case-study Tom: from bashing oneself up to future-oriented self-empowerment
Tom has been addicted to heroin since he was 16 years old. His father was a violent man, and when Tom was 14, his parents split up. Tom started acting outsmoking pot, stealing a carand his father threw him out of the house. After living on the street for some time, he moved in with some heroin dependent friends and became addicted as well. In his thirties, he briefly gave up heroin and he got married and had four children, but relapsed again in his late thirties. At the age of 45, he regained some stability: found housing, got on methadone, remarried. But his substance use was still not really under control. We first interviewed him at the age of 48. A few weeks prior to our meeting he almost died of an overdose. But directly following this, though still using, he reported a shift in thinking, specifically highlighting a "moment of clarity": "It sort of slammed home that somebody else was involved and I was hurting people around me, whereas before that, it never concerned me." Expanding on this moment of clarity, he reports realizing that his current behaviour is very selfish towards his loved ones, mainly his wife and his mother. In fact, he comes to see himself through his mother's eyes, as 'selfish': "My mother's been a major influence. She's still in the picture, and it was actually her that said quite a few years … probably several years ago -'you're selfish, how can you do this, you're so selfish to yourself and other people'. That actually rang home, and the penny dropped, yeah, four weeks ago. It was, '

wow, (…) I left four kids behind, I left all sorts of things, and all sorts of great opportunities, and yeah.' But that's not said in a guilt trip, that's said in realization that I just can't continue that way anymore, it's just ridiculous, yeah."
What instigates this moment of clarity in which Tom sees himself in a new way, as much as he sees others in a new way? Notably, this shift in perspective involves a dramatic shift in "concern", but not just in his concern for others. Importantly, he shows a distinctive shift in concern for himself, as someone who's been operating in a 'ridiculous' and unnecessary way. And it is this shift in concern for himself, as someone who 'deserves' something better, that allows him to handle his feeling of guilt and shame in a forward looking self-reproaching way that avoids selfpunishment and opens up new possibilities for himself going forward: "Part of my use was to punish myself and when I did use I would punish myself more by … because I would be regretful and thinking 'oh why did I do that?' and then it would become a prolonged punishment and 'oh well you don't deserve any better' and all this sort of stuff you know thinking to myself (…) I would bash myself up, something chronic, guilt-wise.

"All these things come up in your life and you can either give up and run and jump into your old life behaviours or what I'm choosing to do is stand up (…) So I guess yeah sitting back assessing the situation for what it is
and not collapsing and falling over and giving up. "I knew I was an intelligent, smart person at the start but I acted pretty darn … I acted pretty selfishly and stupidly by reaching out to drugs. But I don't condemn myself for that anymore, that's an experience I went through but now with my faculties I'm going to use that experience not as a waste but something … as a strength for other people. "I don't want to suffer that anymore, Anke. The guilt, the shame and the self-condemnation." Tom clearly shows a shift in his attitudes toward himself, becoming more generatively future-orientated so far as he feels empowered to bring that future about. When we ask him whether he has found his purpose in life, he replies:

"No well I wouldn't say I've found it, I knew it. I've unlocked it and I've allowed it just yeah, absolutely just become the forefront, the future and it … like the present and the future of my life is my purpose, yeah."
While this is a single case study, other respondents in recovery, similarly tortured by feelings of guilt, shame and self-condemnation, report that an important step forward involved developing a concern for themselves coupled with the cognitive realization that they are not simply victims of these emotions, but can take responsibility for how they experience themno longer in a "self-bashing" way, but in a way that gives them valuable future-oriented insights into how they should manage their lives and their relationships with other people.
Below is a quote from the respondent we cited before. Her coping strategy with guilt is to make lists of the things she feels guilty for, and determine whether there is anything she can do to make amends, or whether it is unproductive guilt.
"I thought there has to be a way to get over guilt, because it doesn't serve anyone any purpose. (…) I just went, oh my goodness, put that behind me, today's a new day. I need to start over as long as I'm committed and, you know. Because otherwise if I live in the past I'll never go forward, yeah. (R36C) 8. Narratively reclaiming one's past, present and future in the service of scaffolding a new self-identity In the preceding section, we stressed that developing a new attitude towards the self is an essential part of recovery, one that is proactively scaffolding as opposed to reactively self-punishing. And we noted that this has both cognitive and affective dimensions.
With regard to the affective dimension, we argued that a key element here is not simply replacing feelings of shame with feelings of guilt, as the standard theory suggests. Rather, it involves developing some empathetically mediated concern for the pain and distress one is causing oneself, not only by acting in dysfunctional ways, but by experiencing the shame and guilt this induces in a self-condemnatory mode. As Tom says, "I don't want to suffer that anymore…The guilt, the shame and the self-condemnation." But experiencing shame and guilt is not the same as suffering it, as Tom himself comes to see. It is rather how he handles these emotions in light of the memories, thoughts and events that invariably stimulate them. As he says, "All these things come up in your life and you can either give up and run and jump into your old life behaviours or what I'm choosing to do is stand up (…) assessing the situation for what it is and not collapsing and falling over and giving up." In short, Tom has found a way to live with these emotions, even put them to constructive use. This gibes with a growing recognition amongst researchers that shame, and not just guilt, can be productive in recovery (Flanagan, 2013;Sawer et al., 2020).
In this final section, we focus again on the cognitive dimension of developing a new attitude to the self, exploring in more depth the elements involved. As emphasized above, a key aspect of this shift involves investing oneself with the power to effect lasting changes in one's life, not just in behavior, but more deeply in one's attitudes and dispositions. It involves seeing one's dysfunctional attitudes and dispositions as malleable, not fixed in stone, not the inevitable result of past circumstances and events, not the defining feature of one's identity. But, more deeply than this, it involves re-awakening one's own sense of agency as instrumental in effecting such changes, in part by understanding one's own agential role in maintaining a dysfunctional lifestyle via persistent condemnatory self-blame. In effect, it involves coming to see oneself as principal author of the "condemnation script" one is no longer doomed to carry out (Maruna, 2001), but instead can use one's agential power to rewrite.
But writing a new script is no easy task. One of us (Kennett) has emphasized the difficulties faced by persons with long term addictions in imagining a changed future for themselves (Kennett, 2013;Kennett & Wolfendale, 2019). The problem is in part one of identifying with a future sober self. The agent who feels guilt and shame in the retributive mode usually has a very negative self-image. They have internalized a view of themselves as both incapable and undeserving of anything better. So, though they may wish for, and see the value of a life without drugs, they have trouble projecting themselves into that future. If they cannot imagine and identify with a sober version of themselves, it does not feel like their future. In addition, the agent has to see the desired future (perhaps involving a return to study, a job, and the repair of family relations) as achievable by her own actions. It has to be possible. If the envisioned future does not seem open in these ways it will not motivate or sustain change. As one of our study respondents remarked about their struggle to stop using: "It was hard, like before I wanted it, I wanted you know to get my family back and to have my health and all that but it was … it just seemed so far away or that I just didn't want it because there was no point. Like I felt sort of hopeless a bit, hopelessness, a bit of hopelessness" (R42C).
If retributive self-blame contributes, as we have argued, to persisting feelings of hopelessness and agential disempowerment with regard to making a better future, how is this cycle to be broken? How is such a person to develop a more positive vision of themselves as someone valuable and ultimately forgivablea vision that can guide and motivate them as they struggle to revise long-standing attitudes and dispositions and resist falling back into dysfunctional patterns of behaviour? How is this positive vision of the self to be stabilized and secured in order to underpin the scaffolding work of self-reproach?
Here we think narrative therapy may have something substantial to offertherapy that encourages individuals to tell a better story of their lives and supports them in that endeavour. Perhaps surprisingly, research indicates such therapy works best if it is not solely focussed on the future, on what recovering individuals could be or become (Maruna & Ramsden, 2004;Maruna, 2001;O'Reilly, 1997;Rotenberg, 1987;Vaughan, 2006). Rather, individuals do better when they are encouraged to 'reclaim' or 'rebiographize' their past (Rotenberg, 1987), both focussing on, and enhancing, remembered moments of happiness, particularly as these are linked to positive character traits or successful moments of coping (McConnell & Snoek, 2018). In effect, individuals are better able to project a newly emerging positive image of themselves into the future by first projecting that image into their past, finding in remembered events evidence that the person they want to be is in fact the person they have always been, empowered now to address those dysfunctional features of themselves by drawing on the lived experience of their own positive agency. In keeping with Kennett's insights, a better future becomes imaginatively possible for these individuals in so far as they are no longer trying to create someone new out of thin air. In short, as forensic psychologist Mordechai Rotenberg observes, stability and security in the new self-vision depends on maintaining "psychological continuity and cognitive congruity" with one's (re-biographized) past, a way of anchoring that self-vision in a previously lived reality, however reformatted that reality needs to be (Rotenberg, 1987).
We return to the case study of Tom to illustrate these themes. After our initial interview with Tom, we followed him for three years interviewing him on a yearly basis. In these years, we witnessed him undertaking hard work of recovery, an important element of which involved reclaiming his identity as someone who is essentially helpful, rather than someone who is essentially selfish. With the help of others, he develops narrative resources to elaborate and verify this alternative vision of himself now secured in his memory of past events.
In his interviews with us, there is one story about his childhood that he repeatedly references. When he was young, one of the neighborhood kids used to wet his bed and had to go to school in his soiled clothes. The boy got picked on for that reason, and Tom took pity on him and gave him his spare school uniform and made sure he was not picked on anymore. He also took in stray dogs and offered children from a boys' home nearby to come and live at his place.
In the course of relating these events, Tom describes himself as a helper. "I want to help others. I want to contribute to others' lives who are disadvantaged and stuff." But, importantly, the helping events he now recalls are not just from his pre-using past. The helpful persona persists even through periods of dysfunctionhence, becomes the underlying truth about himself on which he can depend. In evidence of this, Tom recounts an incident during his using life in which he stood up for his substance dependent friends. When attending church one of Tom's friends is asked to leave and not come back to sermons anymore, because some people in church are afraid he will steal their wallets. Tom defends this person, and his friend is welcomed back in the church: "I thought every Sunday you get up and say we're here for these people and now you've just shut the doors on one of the very people you say you're here for. (…) The people, the very people you are supposed to be supporting, you know the very people you say Jesus came for." True to this reclaimed identity, Tom has become an advocate for people who are homeless or struggling with addiction. He is involved with various homeless organizations, where he negotiates the rights and needs of homeless people with policy makers. His church has provided other opportunities for him to develop his identity as a helper. A pastor friend of his has erected a church in Papua New Guinea, and Tom is eager to join him in doing community work there. In Australia, he is already involved in supporting local Aboriginal communities. Of late, he has also been offered a paid position as a mentor for homeless people.
Thus we see Tom anchoring his reclaimed identity, not just in narrative recountings of past events, but alsoand criticallyin core current and future projects to which he is committed in an ongoing way. And while we see this process of reclaiming his identity as central to his recovery, it is not a magic bullet. Tom has faced a number of challenges in the years that we have followed him, both practical and emotional. For instance: now that he has a fixed residential address, he has had to deal with court cases re-emerging from the past. In addition, he suffered a brief physical relapse after trying to come off methadone, with this relapse costing him his marriage. More deeply, he has faced scepticism regarding the stability of his new identitysome of it coming from others, but some of it funded from his own fear of continuing stigma in the eyes of others. When we asked him if it was hard to make new friends, he replied that the biggest challenge was relating to others in a way that took their acceptance of his reformed persona for granted: Only hard because of my own views, like you think people you know, know about you (…) mistrust you but that's only … (…) that was from my own personal … that was from myself, not because of the way I was being treated or anything in any way, so yeah but I've become a bit easier about it that people are accepting and you know they see change and they're willing to encourage you in that change. I think you know those battles within yourself about yourself, honesty, your truths, and I think that's where a lot of other addicts really, really struggle 'will people accept me for who I really am or is it easier to be seen as the user or the struggling ex-user?'

Conclusion
Tom's experiences show how hard it can be for recovering and recovered individuals to reclaim a positive identity for themselves. The fight to overcome one's own self-condemning, self-stigmatizing shame and guilt is a hard enough mountain to climb. What we have only noted in passing in this paper is how other peopleindeed, society at largecan materially contribute to these difficulties through punitive, stigmatizing, and identity-fixing attitudes and practices. We close with a brief consideration of this critical issue.
As one of us (Snoek) has argued elsewhere, stigma from society can result in internalizing this negative identity-fixing attitude towards the self, resulting in a pre-emptive practice of self-stigma that hinders recovery (Matthews, Dwyer, & Snoek, 2017). Kennett and Wolfendale (2019) have also argued that agency can be externally undermined through social and political narratives that deny people important elements of moral recognition (see also Haslanger, 2012). Social narratives around addiction which portray persons with addiction as worthless and interpret all their actions through the lens of their addiction greatly increase the difficulty of individuals' developing and sustaining narratives that are genuinely supportive of change. If being an addict is taken to wholly define who someone is, how is that person able to maintain hope for a better future? Hostile social and structural environmentsincluding stigma, poverty, and lack of access to services allact as external barriers to recovery and undermine an addicted person's attempts to exercise their agency in rebuilding their identity. As someone with a long-term addiction put it: …I'm totally unemployable. I'm over the hill, got no references, no appreciable skills, patchy work history at best, … I mean [participating in treatment] is not the answer to all my problems. Recovery is not going to make my problems go away (Weinberg & Kogel, 1995, p. 217).
In this paper, we have outlined some steps that we believe are instrumental in promoting recovery and that are morally valuable in themselves. Primarily, we have been concerned with people's attitudes towards themselves. But, as we noted earlier, such attitudes are generally learned from the surrounding culture and can be especially toxic in relation to blame. Thus, we here want to re-emphasize a point we made earlier that supporting people in recovery requires foregoing attitudes and practices of retributive blame directed towards them in relation to their addiction. This recognizes the fact that others can play an essential role in modelling more constructive practices of holding responsible that we associate with scaffolding reproach. We see a good example of this in one of our studies, with one of our respondents describing how his son's manner of reproach became a turning point for him: "Dad, if you drink or not I don't care. I will still come and visit you, because I know who you are, you are more than your drink. You are a positive person, and I will always love you. Only, if you continue drinking, you will have a smaller role in my life. Because, if you stay drinking, you only have a year or three left in life." (R6) This heartening example of reproach with encouragement shows how it is possible to provide external scaffolding to people with addiction, supporting their efforts to manage their guilt and shame by holding out a positive vision of who they area person that attracts, and is capable of attracting, love and admiration. All going well, a person who is externally scaffolded in this way will come to internalize this skill, becoming an empathetic self-scaffolder aided by narrative resources devoted to sustaining a positive vision of who they are and what they can be. Still, we don't want to minimize the amount of care, attention and social support involved in co-constructing a recovering person's path to recovery. The future that we hold out to people with addictions and encourage them to work towards must be one that is meaningfully available and valuable to them. As Tom's case shows, successful selfscaffolding often requires social recognition and social opportunities to make our revised stories come true.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.