Research Article, J Trauma Stress Disor Treat Vol: 7 Issue: 1

Living a Life That Should Not Be Lived: A Qualitative Analysis of the Experience of Survivor Guilt

Pethania Y^1*, Murray H² and Brown D¹

¹University of Surrey, United Kingdom

²South West London and St George’s Mental Health NHS Trust, United Kingdom

*Corresponding Author : Y. Pethania
University of Surrey, United Kingdom
E-mail: dryasminpethania@gmail.com

Received: February 19, 2018 Accepted: March 06, 2018 Published: March 13, 2018

Citation: Pethania Y, Murray H, Brown D (2018) Living a Life That Should Not Be Lived: A Qualitative Analysis of the Experience of Survivor Guilt. J Trauma Stress Disor Treat 7:1. doi: 10.4172/2324-8947.1000183

Abstract

Keywords: Psychopathology; Survivor Guilt; Shame; Trauma; Post-traumatic Stress Disorder

Introduction

“I am living a life that should not be lived. That’s where my guilt comes from because he took my place, because he stepped into my shoes and it should’ve been me, it would’ve been me...”

As illustrated in this quote from one participant in the study, excessive or irrational guilt feelings are commonly reported by survivors of traumatic events. Guilt has been associated with increased severity of post-traumatic stress disorder (PTSD) and other psychopathology [1,2]. Survivors may experience guilt in relation to various aspects of trauma, including actions they did or did not take, feelings they did or did not have, trauma-related beliefs or surviving when others did not, ‘survivor guilt’ [1,3-5]. The Diagnostic and Statistical Manual of Mental Disorders listed survivor guilt as an associated feature of PTSD, defined as ‘guilt about surviving when others have not or guilt about behaviour required for survival’ [6].

Survivor guilt has been observed in survivors of various traumatic events including the 9/11 terrorist attacks, industrial disasters and combat [7-9]. Studies have shown the prevalence of survivor guilt in combat veterans to range between 38% - 40% and between 36% - 61% in survivors of industrial or transport disasters [8-11]. Survivor guilt has been associated with more severe PTSD symptoms, persistence of guilt feelings and suicide attempts [8-11]. Despite the prevalence of survivor guilt and its association with more severe pathology, there has been little research considering how it is experienced, particularly in traumatised populations. Early psychoanalytic accounts of survivor guilt, such as Niederland’s work with Holocaust survivors, suggested that survivor guilt represents an unconscious betrayal of the dead, where surviving creates an ongoing conflict and constant source of guilt and anxiety [5,12]. Lifton studied survivors of the Hiroshima bombing and Buffalo Creek disaster and proposed that survivor guilt was related to feeling responsible for the death of a loved one and feeling their life was purchased at the cost of another [4]. These theories were based on observational data and have not been empirically tested; nor is it clear whether observations relating to specific historical traumatic events such as the Holocaust would transfer across time, populations and contexts. Survivor guilt also emerged as a major dimension in grief reactions in grandparents bereft of a grandchild [13]. Variables associated with survivor guilt were struggling with the notion of being punished by God; struggling with feelings of self-doubt and guilt for being alive; struggling with feelings that it was their turn to die not the grandchild’s and needing to apologise frequently to their family for causing them grief. Less specific guilt and shame reactions have been studied in traumatised populations [14,15]. Kubany and Watson suggested that trauma-related guilt arises from one of four types of beliefs: 1) perceived responsibility for causing the event; 2) perceived violation of standards of right and wrong; 3) perceived lack of justification for actions taken and; 4) beliefs about foreseeability and preventability [16]. Some beliefs related to survivor guilt conform to these categories, but other survivors report guilt simply for surviving, even knowing that they had no influence over the death of others.

Overall, the limited existing literature about survivor guilt lacks clarity and depth and does not sufficiently address key questions about survivor guilt: how it develops, how it is experienced, its underlying mechanisms and how or whether it is phenomenologically different from other types of guilt. This study used interpretative phenomenological analysis (IPA) to explore individual experiences of surviving a traumatic event where others have died, with the aim of understanding how individuals interpret and make sense of their experience of surviving and how/if survivor guilt is relevant to this experience.

Method

Participants and procedure

Participants were recruited from a traumatic stress service in London, UK. Clients in treatment at the service who had survived a traumatic event in which others had died, and spoke fluent English were given information about the study by their therapist. Six participants were approached, and all consented to the study.

The sample consisted of three men and three women. Participants ages ranged from 40-63 (mean = 48). They had been receiving trauma focused cognitive behavioural therapy for PTSD for 6-24 months. Therapy had not explicitly focused on survivor guilt in any of the cases, but may have been discussed as part of idiosyncratic formulations. Traumas experienced included a life threatening operation or health condition; military combat; genocide; and violent assaults. Four of the participants reported they had close relationships to those that died during the traumatic event, the other two participants reported those that died were strangers.

All participants were interviewed at the trauma service. As the nature of the subject was likely to be emotive and distressing, participants were given the opportunity to decline questions and were debriefed verbally at the end of the interview. Each interview lasted between 60 – 110 minutes and was audio-recorded. An interview schedule was developed using guidelines by Smith, Flowers and Larkin [17] and included the following questions: (1) Can you tell me about the traumatic event that you survived? (2) How do you feel about your survival? (3) What do you think about the people that did not survive? (4) Do you think surviving this traumatic event has played a role, if at all, in the way you view yourself? (5) How has surviving this traumatic event made you feel about life in general? (6) How do you think other people view your survival? The participant was positioned as the ‘experiential expert’ and encouraged to tell their own story. The researcher used prompts and supplementary questions to explore areas of interest that emerged during the interview [18].

Ethical considerations

Ethical approval for this study was granted by the UK National Research Ethics Service Committee, and University of Surrey ethics committee.

Data Analysis

Analysis followed the IPA procedural steps outlined by Smith and Osborn [18], and was conducted by the first author, with assistance from the other authors and an IPA research group. Each audiorecorded interview was transcribed verbatim, and read multiple times. A list of emergent themes relating to the nature, meaning and context of the participant experiences were identified. Patterns and connections between themes were identified, resulting in a list of super-ordinate themes and underlying sub-themes for each transcript. Patterns and connections across the transcripts were next identified, recognising themes that represented higher order concepts that the transcripts shared. A master list of themes was created for the group which illustrated how super-ordinate themes related to each participant.

Results

Three key themes emerged from the analysis, and the interacting relationships between them generated a theory relating to how survivor guilt is experienced. Overall, participants appeared to be engaging in an ongoing dynamic of trying to make sense of surviving many years after their traumatic event (over 20 years for four participants). This on-going dynamic related to three interacting key themes that were evident in all participant accounts: 1) persistent guilt and disentitlement to life; 2) trying to make sense and; 3) a need to repair.

Persistent guilt and a sense of disentitlement to life

All participants viewed their survival of the traumatic event as “unfair” or fundamentally “wrong” because others had died. Participants emphasised a strong sense of disentitlement to the life they were living since they perceived themselves as undeserving in comparison to the deceased or they felt they had not done anything to earn their survival. This was described alongside a persistent sense of guilt for surviving in the absence of any apparent wrongdoing, which had a different quality to other forms of guilt:

“It’s deeper, it’s stronger... it is a hard thing to describe, it’s a sickening feeling, it’s... if I smashed into somebody’s car and drove off I’d think, “yeah ok I shouldn’t have done that, I should’ve... but I’m gone, tough…”, this is guilt because a person’s life ended, a good friend, a brother…this guilt lasts and lasts, and it’s hard to shake, it’s hard to get rid of, you can’t just say “oh well, he’s dead” it comes back and smacks you in the face.”

Common to all participants was that they had “beat the odds” to survive which they perceived as unfair as they viewed themselves to be the “same” as the deceased. This left them with a strong conviction that they should not have survived and a sense of incomprehensibility and disbelief about surviving:

“We were all using the same guns, we all had the same magazines, we knew how throw around, roll right, round roll left, take cover but... as you were engaging the enemy, your comrade on your right is not moving, you feel, he’s gone, he’s gone... it doesn’t make sense...”

Being ‘in the same boat’ as the deceased, as opposed to having a close relationship, may be an important factor in the perception of inequity. The use of the metaphor “cheating death” by one participant indicates a sense of unfairness and having done something wrong in surviving. Several participants were also left “feeling responsible” for the death of others by surviving. A tendency to idealise the deceased was also noticeable within participant accounts, leading to some participants feeling less entitled or deserving of life in comparison:

“I kind of feel like it was my fault really, if he were operated on instead of me he might’ve survived... he would’ve... made it and he would’ve been a good member in society, contributing and doing as he should, that’s how I see it... and I’ve survived because I was put in front and now I’ve come out and everything’s so dark and heavy... I feel like I should be, going everywhere and finding everyone that’s needing a little bit of help and helping them...”

Disentitlement to life was indicated in participant descriptions of “existing” rather than “living” life with one participant stating “I’m just waiting for the moment when I will die”. Participants described conflicted feelings about surviving, feeling “happy one minute” and then “it’s just depression”. This fluctuation in emotions and thinking the deceased were “better off” was evident in all accounts. For some, it manifested as a physicality of “not feeling alive” and was described alongside an internal sense of “just existing”:

“...you cheated.., but you didn’t die... completely but you still died, physically because you’re not feeling normal, you’re not feeling alive, like the way you’re supposed to enjoy your life

Some participants described life as a constant struggle following the traumatic event. One participant conveyed the effort of keeping himself alive and getting through the day:”

“I’m in the process of trying to stave him [death] off..., instead of living; I’m just trying to keep death at bay.”

It appears feelings of guilt are reinforced by disentitlement to life i.e. external cues, such as the physicality of “not feeling alive” and internal cues such as appraisals such as “I should not have survived”. However disentitlement to life appears to be generated from feelings of guilt about surviving, suggesting a reciprocal relationship between the two.

Trying to make sense

Participants viewed their survival as incomprehensible. Perceiving they should not have survived the traumatic event left them with questions about why they did. Participants described a strong need to make sense of why they survived and described investing effort in trying to make sense but found it almost impossible. Existential questions such as “why did I survive?”, “why didn’t I die?” and issues relating to entitlement to life were dominant in participant accounts and central to sense-making.

Participants commonly referred to previously held beliefs or assumptions relating to how the world operates to help them make sense, however most participants found them unhelpful. One participant found he could no longer rely on previously held religious beliefs to make sense of his survival, generating existential questions:

“I was brought up in a religious upbringing so they talked about God and good and evil, those things. When I started looking at death in, it’s actual, vivid horrible, then nothing makes sense anymore.... what does all this mean?”

Being unable to make sense of existential issues resulted in ruminative thinking for most participants which made sense-making a frustrating and distressing process:

“…other people was the same..., situation but they died and I survived. Why did I survive? I have no idea. Why do I have to remember them every single time? I didn’t die. My question is why? Why this have to happen? Why do I have to go through and remember each and every one in my family? [Crying] It makes me feel really that I’m not supposed to be alive”

Rumination relating to existential questions also presented as powerful and alluring. The use of “void” below suggests a sense of emptiness. Making sense may be perceived as necessary to filling this “void”:

“There is that void of not understanding why it happened like that and why, why, why did I... how could I have survived? It’s a powerful void and it kind of sucks me so... what am I doing here? Why don’t I join them where they are?”

Ruminative thinking was mirrored during the interview process. Analysis of the linguistics of these extracts showed lack of coherence, hesitation, emphasis, repetition of the word ‘why’ and strings of questions. This could be indicative of participants being stuck in a current ruminative state and actively being in the process of ‘sense making’ during the interview.

Dissonance associated with struggling to accept logical explanations at an emotional level also related to being unable to make sense of survival, as highlighted below:

“...everybody’s told me, proved to me all the facts that I shouldn’t feel guilty, it wasn’t me that killed him, so many different facts but still there’s that in me that I feel it was wrong...”

“I was responsible but I’m not but you are and it’s a tug of emotions really within yourself, you’re trying to tell yourself, you’re trying to convince yourself that you didn’t cause it but you’re telling yourself you are responsible because you changed, you put him in that position ”

It appears attempts to make sense are fuelled by guilt and disentitlement to life, however, being unable to make sense also intensifies these feelings, generating a distressing cycle of rumination. Achieving sense was thought to reduce distress; one participant stated he would be “comfortable with the whole situation” if he could understand why he survived.

Two participants indicated some success with making sense of their survival and attributed their survival to chance or God. Rather than being in a present state of sense-making, these two participants spoke about the process of sense-making they had been through:

“...talking to my grandparents and their friends as well, people who were in the second world war, I think I came to the realisation that things just happen in war time, or things that happen in the military, it’s just, it happened...it was chance… I’ve had long meaningful discussion with... priests and pastors....”

“I pray a lot and, and I hear him sometimes, when I pray to God and tell him whatever is happening most of the time and say “why this have to happen like this?”. Many times he will answer me, many times... I will see something…”

These participants demonstrated less of a present ruminative state observable in others. However, despite having achieved some sensemaking, both continued to feel guilty and demonstrate a vulnerability to ruminative thinking and retreating into an internal world:

“Every time I see a picture of him, I feel like I wanna be physically sick. I get very depressed, I get so depressed... ... it’s funny in a respect that the level of guilt and depression and sickness and ... goes up and down. I tend to... withdraw into myself, I mean really withdraw”

A need to repair

Participants felt they did not deserve or earn their survival in any way, which drove a strong desire to repair. Making amends for surviving appeared to be an attempt to alleviate or ‘work off’ feelings of guilt. Participants reported a range of personally meaningful attempts of making amends or restoring balance. However, most participants were unable to engage in activities that represented repair, generating frustration and distress:

“I’m very down on myself in that, I don’t feel very useful and... to have survived... what I consider at the expense of another... I think do more with life... and it’s frustrating...”

Some participants felt repair was about restoring the balance through revenge or justice:

“I do feel guilty that nothing’s happened to him, they just got away with it, it is a bit coward not going to... not getting ‘em back, I mean, but that’s just a thought in your mind, you can’t really do something like that.”

Participant accounts highlighted that feelings of guilt and disentitlement about surviving generated a need to repair. However, being unable to repair appeared to intensify feelings of guilt and disentitlement to life, suggesting a reciprocal relationship between these themes. Life as “just existing” contradicts the desire to live a more meaningful life. The sense of disentitlement to life appeared to be a barrier to participants feeling they can engage with life and ‘doing’ things that represent repair.

The two participants that indicated they had some success with making sense of their survival also demonstrated some success with engaging in repair activities despite both feeling a sense of disentitlement to life. One said he was “making the most of” life in order to make amends as he felt it would be “disrespectful” to his deceased friend if he lived a life that was not “meaningful”. The other participant was completing a course relating to helping others:

“I think I have been looking for the people suffering with.., crying for help, you can listen to them... and of course I can’t change the world but you never know, in the future you can save one, two...”

These two participants demonstrated a stronger presence in their external world throughout their interviews by giving examples of what they were ‘doing’. In contrast, participants not engaging in repair activities demonstrated a strong presence in their internal world.

A possible theory of survivor guilt

The three main themes presented above each impacted on each other. A diagram of the theoretical model proposed is presented in Figure 1.

The model explained:

• Survivor guilt appears to relate to a persistent sense of guilt and disentitlement to life which is generated by a perception that surviving was unfair or wrong. The unique quality of the experience of survivor guilt, versus other types of guilt, is related to the reciprocal relationship between feelings of guilt and disentitlement to life which reinforce each other. Survivor guilt appears to drive the need to make sense of surviving and make amends or repair for having survived.

• Being unable to make sense is characterised by ruminative thinking which generates distress and reinforces guilt and disentitlement to life.

• Being unable to repair generates frustration, distress and reinforces feelings of guilt and disentitlement to life.

• Being able to make sense appears to be a gateway to ‘doing’ (engaging with life and activity) which enables repair activities to take place. However ‘doing’ is in conflict with a disentitlement to life, intensifying feelings of guilt and disentitlement to life and presenting a barrier to engaging in repair activities.

Discussion

The results revealed that participants found themselves in an ongoing dynamic of trying to make sense of why they survived when others died. Central to this dynamic is persistent guilt about surviving and disentitlement to life driving internal processes associated with sense-making and external processes associated with repair. The theoretical model derived from the interview data is an initial attempt at representing the phenomenology and underlying mechanisms of survivor guilt. Similar to the clinical descriptions of ‘survivor guilt’ by Neiderland [5] and Lifton [4], participant experiences centred on guilt for getting more than their share of life. However, in this study, guilt about surviving was demonstrated as interacting with other experiences and affects. The guilt experienced by the participants in this study was persistent, experienced in the absence of any wrong doing and experientially different to other types of guilt.

Some participant beliefs were similar to those suggested by Kubany and Manke [1] in relation to trauma-related guilt, including perceived responsibility for causing the death, or a violation of standards of right and wrong through perceiving survival as unfair. However, the model by Kubany and Manke [1] does not account for the sense of disentitlement to life indicated by negative self-evaluations or feeling less deserving of life in comparison to the deceased. These global negative self-evaluations could relate to Wallbott and Scherber [19] and Tangney’s [20] conceptualisation of shame, which is often thought to relate to depreciation of the entire self, whereas guilt involves depreciation of specific actions or behaviours. Survivor guilt may include elements of both guilt and shame. Participants in this study felt unworthy of life that was something wrong with them, and perceived themselves as having done something wrong by surviving. In this study, shame and guilt appeared to fuel and maintain each other i.e. ‘I should not have survived so I must have done something wrong; I did something wrong so I should not have survived’. Making sense of why they survived when others did not was of great importance to the participants in this study. Janoff-Bulman [21,22] described how traumatic events can shatter a person’s “assumptive world”, cognitive schemas related to the meaningfulness of the world and worthiness of the self. The violent loss and bereavement literature stresses the importance of sense-making in the restoration process by creating new meanings following loss [23,24]. Sense-making was attempted by all participants; however most participants were unable to develop helpful new meanings. Failed attempts at making sense of loss are suggested to result in rigid and recurrent cycles of thought which was demonstrated in participants’ rumination [25]. Participants described feeling responsible for those who died, by surviving. Downey, Silver and Wortman found that parents that had lost a child to Sudden Infant Death Syndrome were more distressed if they made attributions of causality than those that did not [26]. Self-attributions were more distressing than attributions to chance or God. The two participants in this study who did attribute cause to chance or God were also the most able to participate in their lives and the least ruminative. Other participants were unable to accept chance as an explanation for their survival. This might relate to participants’ struggle to reconcile ‘knowing’ they were not responsible, with ‘feeling’ or accepting this at an emotional level. Sense-making may therefore relate to seeking a personally meaningful truth rather than facts and literal truths [27].

Participants in this study described a strong desire to make amends for surviving. Individuals feel guilt when they believe they have done something contrary to their code of conduct and/or when their actions have injured others, resulting in attempts to repair [28,29]. Kubany et al. [30] theorised that survivors of traumatic events may perceive they have benefitted unfairly at the expense of others, leading to beliefs that they should do something to undo this inequity. In a situation where the person has died, restitution is blocked. Being unable to repair in participants’ experiences was related to frustration and distress as well as intensified guilt, disentitlement to life and rumination. The participants in this study who were able to engage in activities related to repair were the same participants who had experienced success with sense-making and seemed to have a greater focus on their external worlds. It may be that making sense frees up personal resources to engage in repair activities. However, engaging in attempts to repair did not alleviate feelings of guilt or disentitlement. Gilles and Neimeyer suggest that a change in distress may not an expected outcome of sense-making, given that it is an ongoing process throughout life [25]. Survivors perhaps need to find personally meaningful ways of living with survival. It may not be possible to ‘work off’ guilt by making amends for another’s death. The results of this study offer several insights to inform practice. First, it highlights the importance of directly exploring guilt reactions. Survivor guilt is a phenomenon that requires attention in the assessment and treatment of those who have survived fatal traumas. Second, making sense of their experiences was of utmost importance to participants in this study indicating this may be significant in the resolution of survivor guilt. In terms of limitations, the results and theoretical model proposed reflect the experiences of only six participants and definitive claims beyond these participants cannot be made. The analysis was led by one researcher and, while attempts were made to ensure validity, another researcher’s interpretation of the data could yield different results.

Research on survivor guilt is sparse and the results of this study offer some insights into the phenomenon and its underlying mechanisms. This theoretical model proposed is not being presented as the final version and components of the model require testing through further research. For example, further research could examine the links between guilt and repair in the context of trauma and loss.

References

1. Kubany ES, Manke FM (1995) Cognitive therapy for trauma-related guilt: Conceptual bases and treatment outlines. Cog Behav Practice 2: 27-61.
2. Lee DA, Scragg P, Turner S (2001) The role of shame and guilt in traumatic events: A clinical model of shame-based and guilt-based PTSD. British J Med Psycho 74: 451-466.
3. Janoff-Bulman R, Wortman CB (1977) Attributions of blame and coping in the “real world”: Severe accident victims react to their lot. J Pers Soc Psychol 35: 351-363.
4. Lifton RJ (1976) Death in life: Survivors of Hiroshima. Oxford, UK: Simon & Schuster.
5. Niederland WG (1968) Clinical observations on the 'survivor syndrome'. Int J Psychoanal 49: 313-315.
6. American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders (4th edtn.). Washington, USA.
7. Mallimson K E (2005) Survivor guilt and posttraumatic stress disorder 18 months after September 11 2001: Influences of prior trauma exposure to the event and bereavement. The Wright Institute.
8. Hull AM, Alexander DA, Klein S (2002) Survivors of the Piper Alpha oil platform disaster: Long-term follow-up study. Br J Psychiatry 181: 433-438.
9. Okulate GT, Jones OBE (2006) Post-traumatic stress disorder, survivor guilt and substance use: A study of hospitalized Nigerian army veterans. SAJP 12: 37-40.
10. Hendin H, Haas AP (1991) Suicide and guilt as manifestations of PTSD in Vietnam combat veterans. Am J Psychiatry 148: 586-591.
11. Joseph SA, Hodgkinson P, Yule W, Williams R (1993) Guilt and distress 30 months after the capsize of the Herald of Free Enterprise. Pers Individ Dif 14: 271-273.
12. Niederland WG (1981) The survivor syndrome: Further observations and dimensions. J Am Psychoanal Assoc 29: 413-425.
13. Fry PS (1997) Grandparent's reactions to the death of a grandchild: An exploratory factor analytic study. Omega: J Death Dying 35: 119-140.
14. Kugler K, Jones WH (1992) On conceptualizing and assessing guilt. J Pers Soc Psychol 62: 318-327.
15. Tangney JP (1995) Shame and guilt in interpersonal relationships. Guilford Press, New York, USA.
16. Kubany ES, Watson SB (2003) Guilt: Elaboration of a multidimensional model. Psychological Record 53: 51-90.
17. Smith JA, Flowers P, Larkin M (2009) Interpretative phenomenological analysis: Theory, method and research. Sage, London, UK.
18. Smith JA, Osborn M (2008) Interpretive phenomenological analysis. Sage Publications, London, UK.
19. Wallbott HG, Scherber KR (1995) Cultural determinants in experiencing shame and guilt. Guilford Press New York, USA.
20. Tangney JP (1997) Conceptual and methodological issues in the assessment of shame and guilt. Behav Res Ther 34: 741-754.
21. Janoff-Bulman R (1989) Assumptive worlds and the stress of traumatic events: Application of the schema construct. Soc Cogn 7: 113-136.
22. Janoff-Bulman R (1993) Shattered assumptions: Towards a new psychology of trauma. J Ner Mental Diseas 181: 208-209.
23. Currier JM, Holland JM, Neimeyer RA (2006) Sense-making, grief, and the experience of violent loss: Towards a mediational model. Death Studies 30: 403-428.
24. Davis CG, Nolen-Heoksema S, Larson J (1998) Making sense of loss and benefiting from experiences: Two construals of meaning. J Pers Soc Psychol 75: 561-574.
25. Gillies J, Neimeyer RA (2006) Loss, grief and the search for meaning: Toward a model of meaning reconstruction in bereavement. J Constr Psychol 19: 31-65.
26. Downey G, Silver RC, Wortman CB (1990) Reconsidering the attribution-adjustment relation following a major negative event: Coping with the loss of a child. J Pers Soc Psychol 59: 925-940.
27. Neimeyer RA (2008) Meaning-breaking, meaning-making: Grief therapy as narrative reconstruction. Conference presentation, University of Hertfordshire, UK.
28. Barrett KC (1995) A functional approach to shame and guilt. Guilford Press, New York, US.
29. Gilbert P (1997) The evolution of social attractiveness and its role in shame, humiliation, guilt and therapy. Brit. J Psychol 70: 113-147.
30. Kubany ES, Abueg FR, Owens JA, Brennan JM, Kaplan AS, et al. (1995) Initial examination of a multidimensional model of trauma-related guilt: Applications to combat veterans and battered women. J Psychopathol Behav Assess 17: 353-376.