Introduction

Given the suffering and ill-health endured by many bereaved persons, it is critical to understand processes underlying effective versus ineffective ways of coping with loss. Considerable effort was made across the decades of the 20th century to describe adaptive processes in coping, the concept of grief work (confronting, dealing with loss) becoming immensely important in theory-building and for the derivation of principles to guide intervention programs. It was widely understood that normal, “healthy” grieving entailed working through a loss, however painful this may be, and that persons with complicated forms of grief needed help with the normalization of their grief work [66].Footnote 1 In the scientific literature, grief work has been defined as a cognitive-emotional process involving confrontation with and recurring thoughts about a deceased person, the loss experience, and the changed world within which the bereaved person must now live (e.g. 20, 30, 37, 60, 66, 78]. In Bowlby’s [20] words, only if the bereaved person “…can tolerate the pining, the more or less conscious searching, the seemingly endless examination of how and why the loss occurred … can he come gradually to recognize and accept that the loss is in truth permanent and that his life must be shaped anew” (p. 93). Following this line of reasoning, grief work has been understood to be a confrontational, adaptive process, one that is related to good health outcomes. Experiencing the loss of the deceased in a repetitive and quite persistent manner as described by Bowlby is said to help bereaved persons to adjust to their loss.Footnote 2

However, there is a related phenomenon, namely rumination, which seems conceptually similar to grief work, but which—in contrast to grief work—has been both theoretically and empirically related to poor bereavement outcomes. Foremost research on rumination in the bereavement field has been conducted by Nolen-Hoeksema [48, 50], Nolen-Hoeksema et al. [46, 53], within the framework of her Response Styles Theory [43]. Rumination is characterized by persistent and repetitive, chronic and passive focus on the occurrence, causes, and consequences of negative (grief-related) emotions and symptoms and it is associated with poor adjustment [46]. This description of rumination seems close to the definition of grief work given above. Admittedly, definitions of grief work typically encompass confrontation of a wide range of concerns to do with a death, while rumination focuses more narrowly on confronting one’s own emotions and going over related problems (see definitions section below)—but still, there is apparent overlap if we follow the conceptualizations given so far. What, then, is the difference between grief work and rumination? More importantly, what psychological processes cause rumination to hinder good adjustment to loss? In considering these questions, it is necessary to acknowledge that “grief work” and “rumination” are theoretical concepts, their meaning being dependent on the way we define them. If we would simply define grief work as having positive outcomes and rumination negative ones, the reasoning would become circular and the explanation would lose power. However, the purpose of this paper is to try to identify psychological mechanisms underlying the cognitive phenomenon that is typically indicated as ‘rumination’ which may link it to poor adaptation (in contrast to grief work).

Looking back at research on psychological mechanisms in coping with bereavement, Nolen-Hoeksema’s research on rumination contrasted with previous approaches investigating good versus poor coping. As she noted [46], bereavement research into the phenomena and manifestations of complicated grief had previously concentrated on different processes, namely, on examining how denial and suppression processes in coping with negative emotions caused ill effects. Nolen-Hoeksema [46] focused on the “opposite form of coping” to denial / suppression (p. 545; our italics), namely, ruminative coping, referring to this as “the polar opposite of avoidance and denial” [50, p. 21]. Nolen-Hoeksema’s research thus marked an important advancement in the examination of processes underlying complications in grieving: from a focus on denial and suppression, or absence of grief work, to ruminative coping. It is important to note that Nolen-Hoeksema was more concerned to establish the causal role of rumination in poor adaptation to bereavement than to ascertain whether rumination was a confrontational or avoidant coping strategy. It is also important to note that she did accept that avoidance was also potentially problematic [51].

Nolen-Hoeksema’s interpretation has recently received support from other scholars in the bereavement field. Bonanno et al. [11], see also Bonanno et al. [9, 10, 12] contrasted rumination with avoidance as a process in coming to terms with loss of a loved one. Bonanno and colleagues described a so-called “grief work as rumination hypothesis”, in which extensive grief processing was viewed as “a form of rumination that may exacerbate rather than ameliorate distress” (p. 87), noting that “extensive grief processing may … be more akin to rumination than healthy working through” (p. 87). Thus, definitional equivalence of “grief work” and “rumination” was—unusually—assumed. But more importantly in the present context: here, again, the emphasis was on rumination as a focus on (i.e. confrontation with rather than avoidance of) grief-related thoughts. Similarly, Michael and Snyder [42] contrasted avoidance of thoughts and feelings relating to grief with cognitive processing which, in excess “may take on a distinctly ruminative character, with repetitive focusing on how awful it is that our loved one died and how bad it feels to grieve” (p. 451). Within the broader literature on coping with stressful life events, Tait and Silver [69] reviewed classic formulations of rumination, noting that ruminations are typically “…believed to play an integral role in the “working through” or processing of a negative life event whereby an individual gradually comes to terms with it…the occurrence of these ruminations, in alternation with periods of denial, allows the individual to come gradually to tolerate increasing doses of distressing aspects of the event” (p. 353; our italics). Again, the emphasis seems to be on intrusion and confrontation, not avoidance and suppression, of thoughts.

In this paper we examine the nature and functions of rumination, with the central concern to investigate whether rumination is indeed a confrontational strategy as suggested by the investigators cited above, or whether it could be an avoidant coping process. We consider these issues to be important: Theoretically, it might help to distinguish rumination from healthy grief work at a process level, and generally promote investigation of the nature and mechanisms underlying unhealthy versus healthy grieving. Clinically, it would imply different intervention strategies for persons with extremely ruminative coping styles (e.g. chronic grievers). For example, put simply, if rumination is too confrontational as a strategy (and assuming that the process referred to as rumination serves to maintain complicated grief), regular distraction might be considered one effective principle for designing therapy. However, if rumination is in some sense an avoidant strategy, then interventions aimed at confronting the person with painful features of the loss would be more appropriate.

In examining processes underlying rumination, we draw on research not only in the bereavement field but also from research on worry in the fields of anxiety and motivation [17]. First, we define the concepts of rumination and worry more precisely. Then we outline the content of the scales that are frequently used to measure the constructs, in order to illustrate the nature of the operationalizations. Next we summarize research that has examined the relationships of rumination and worry to adaptation. We then explore the functions and associated underlying processes of rumination and worry to try to establish how the links with maladaptation may occur. Finally we discuss implications of our analysis for intervention programming, and draw some general conclusions.

Defining rumination and worry

Nolen-Hoeksema defines rumination, or ruminative coping (the two are used interchangeably here) in slightly different ways in different publications, but along the same general lines. The focus is on repetitive negative thinking. For example, an early definition was in terms of engaging in thoughts and behaviors that maintain one’s focus on one’s negative emotions and on the possible causes and consequences of those emotions [44]. More recently, rumination was defined as “Passive focus on one’s symptoms of distress and the possible causes and consequences of these symptoms. The individual repeatedly goes over problems, and his or her feelings about the problems, without moving into problem solving” [47, our italics).

So according to Nolen-Hoeksema [46], ruminators think persistently and repetitively about how badly they feel (endorsing items such as: “I’m so sad,” “I’m so unmotivated.”) and about the causes and consequences of their feelings (“Will I ever get over this?”).Footnote 3 Such ruminative coping has the effect of worsening and lengthening negative emotions and of increasing the probability that a depressed mood will become a depressive disorder [46]. Following this approach, it would seem that bereaved persons who ruminate extremely will eventually “get stuck” in the grieving process and be prone to complications in their grieving process.

Nolen-Hoeksema’s definition of rumination contrasts with a much broader and similarly influential one of Martin and Tesser [40] (see also Lyubomirsky et al. [39], Scott and McIntosh [61]) the latter being familiar in the more general fields of cognition, motivation and anxiety [17]. According to Martin and Tesser [40], “Rumination is a class of conscious thoughts that revolve around a common instrumental theme and that recur in the absence of immediate environmental demands requiring the thoughts. Although the occurrence of these thoughts does not depend on direct cueing by the external environment, indirect cueing by the environment is likely, given the high accessibility of goal-related concepts. Although the external environment may maintain any thought through repeated cueing, the maintenance of ruminative thoughts is not dependent upon such cueing” (p. 7). Martin and Tesser [40, p. 42] included various modes of rumination, such as problem solving, day dreaming, and meaning analysis within their framework. Thus, their conceptualization was not limited to dysphoric thought, rather it included a broad range of instrumentally oriented recurring thoughts. In their view, goals were regarded as one of the primary mechanisms by which ruminative thought occurs: goal non-attainment and perceived lack of progress toward a goal were understood to lead to negative rumination. As such, negative thoughts would result from blocked goal pursuits.

Following Nolen-Hoeksema, a narrower definition focussing on dysphoric mood would seem appropriate in the present context (in our attempt to disentangle the underlying mechanisms through which rumination influences recovery from loss), given the nature of this life event and accompanying grief—and given the need to distinguish rumination from grief work. In a rejoinder to Martin and Tesser [40], Nolen-Hoeksema [45] made the case for distinguishing between problem solving and other forms of thought that they characterize as rumination.

Moving next to the worry domain: A separate tradition from that of rumination research has grown up around the concept of worrying. Rumination and worrying are close enough conceptually (e.g., repeated negatively laden thoughts are fundamental to both rumination and worry) that research and theorizing in the one area can usefully fuel understanding of phenomena in the other [21, 29, 62, 73, 74]. Indeed, Brosschot et al. [21] have taken rumination and worry to be functionally equivalent manifestations, suggesting the overriding term perseverative cognition to describe a core cognitive-emotional process and enable application to a wide range of emotional states and dispositions. Somewhat surprisingly, worry has not been a specific focus in bereavement research, although there are studies on anxiety in bereavement [54].

The concept of worry has been studied extensively over the past two decades. Research was fuelled by the addition of generalized anxiety disorder into DSM taxonomy, with chronic and pathological worry defined as its cardinal feature [1, 2]. Worrying is held to involve a predominance of negatively valenced verbal thought activity [18, p. 562]. Borkovec et al. [17, in 29] defined worry as: “a chain of thoughts and images, negatively affect-laden and relatively uncontrollable; it represents an attempt to engage in mental problem-solving on an issue whose outcome is uncertain but contains the possibility of one or more negative outcomes; consequently, worry relates closely to the fear process” (p. 10).

From the mid nineties onwards, attention has been devoted to worry and ‘meta-cognition’, that is, beliefs that individuals have about the role and function of worry. Interestingly, worriers tend to believe that worry helps to prevent negative future events and is functional in that it prepares for the worst [75, 76]. The same beliefs about worry being functional and adaptive were reported for adolescent worriers [31].

Although it will have become evident that rumination and worry are often defined in similar ways, some researchers have drawn a major distinction between them, namely, according to temporal orientation: rumination has been associated with regretting a past completed action [36, in 40]. Worrying has been described as repetitive thinking about a future difficulty [59]. In the case of bereavement, since regrets about past happenings and worries about future problems are both likely to be present, even to co-exist, both of these temporal orientations need to be included in investigations of dysfunctional cognitive processes.

In conclusion, conceptual differences/similarities between rumination and worry are somewhat arbitrary and dependent on definitions. For present purposes, it is not essential to take a stand on the issue concerning the temporal position of rumination/worry, nor is it necessary to define rumination only in terms of thinking about one’s own emotions, their causes and consequences (as in Nolen-Hoeksema’s early [44] definition). In reviewing the literature on the relationship to health outcomes in the next section, we work from a somewhat broader conceptualization of rumination. First, rumination includes not only repetitive thinking (dwelling on, pondering emotions, etc.) about the past, but also similar thoughts (worries) about the future. Second, although we exclude instrumentally oriented recurring thoughts (for reasons given above), we do focus on a broad range of dysphoric mood thoughts/emotions, such as dwelling on the meaning of the death now and for the future (in line with Nolen-Hoeksema’s [47] more recent definition). In this way, the bodies of research on rumination and worry can be drawn on for an analysis of dysfunctional processes of coping with bereavement.

Measuring rumination and worry

In the rumination domain, Nolen-Hoeksema and colleagues frequently use the Ruminative Responses Scale (RRS) of the Response Styles Questionnaire which assesses people’s tendencies to ruminate when distressed [3, 35, 44, 50].Footnote 4 This scale consists of 22 items (see Fresco et al., 2002, p. 183). These items are self-focused (e.g., I think “Why do I react this way?”), symptom-focused (e.g., I think about how hard it is to concentrate) and focused on the possible consequences and causes of mood (e.g., I think “I won’t be able to do my job if I don’t snap out of this”). Ratings are made on a 4-point scale on a scale from 1 “almost never” to 4 “almost always”—thus tapping frequency, though not necessarily repetitiveness). A psychometric analysis of this scale was conducted by Treynor et al. [71], revealing three emergent factors: reflection (“a purposeful turning inward to engage in cognitive problem solving”, p. 256), brooding (“a passive comparison of one’s current situation with some unachieved standard” p. 256), and depression-related (e.g. “Think about how sad you feel” p. 248) factors. The latter items, being confounded with depression, were excluded from reanalyses, the other two factors had differential associations with depression (see next section).

In the worry domain, research and assessment have been less restricted to using (mainly) a single scale. A frequently used measure is the Penn State Worry Questionnaire (PSWQ, see Fresco et al. [29, p. 183], Davey and Tallis [22]. This 16-item inventory assesses trait worry—its generality, excessiveness and uncontrollability—on such items as “My worries overwhelm me”; “I know I shouldn’t worry about things, but I just can’t help it”. Another measure that is frequently used in clinical settings is the Worry Domains Questionnaire (WDQ). The WDQ is a 25-item self-report instrument developed by Tallis et al. [70] that measures the tendency to worry across five general worry domains: (a) relationships, (b) self-confidence, (c) future, (d) work, and (e) finances. A third measure is the Reasons to Worry Questionnaire [16], which consists of six possible reasons to worry suggested by theory and by reports of former clients with Generalized Anxiety Disorder. These reasons are related to motivation (similar to items on the Scott and McIntosh [61], rumination scale), problem-solving, preparation, avoidance/prevention (e.g. of something bad happening), distraction from more emotional topics, and superstition (e.g. making it less likely that something bad will happen).

Fresco et al. [29] systematically examined the distinction and overlapping features of worry and rumination, using the RRS and PSWQ, finding 2 worry and 2 rumination factors, but high correlations between the factors. It is also important to note that, in the case of all the questionnaires, differences in the selections of items reflect the theoretical approach and definition of domain of the particular researchers [62].

Health outcomes of rumination and worry

Rumination has been linked to intensification of depressed mood (in both bereaved and non-bereaved subjects). In a series of studies, including laboratory investigations and field studies conducted longitudinally, Nolen-Hoeksema and her colleagues have shown that rumination precedes greater depression and enhances negative mood [38, 46]. For example, Nolen-Hoeksema [52] found a positive prospective association between ruminative coping and emotional distress in bereaved gay men. Other investigators have recently provided further evidence that rumination is related to poor bereavement outcome. Bonanno et al. [11] reported confirmation of their “grief work as rumination” hypothesis in the United States (though not in the People’s Republic of China). Grief processing was defined as rumination, as described earlier. Such processing was predictive of poor long-term adjustment among their sample of bereaved spouses and parents in the U.S. (as was deliberate grief avoidance). Further preliminary support for the notion that rumination may play a role in lack of recovery from loss was found by Boelen and colleagues in a cross-sectional study, in which—among cognitive behavioral variables—rumination was one of the most important correlates of complicated grief (CG) and depression [6]. Michael and Snyder [42] found greater rumination to be related to lower psychological well-being among bereaved students. In their cross-sectional study (across different lengths of bereavement), rumination emerged as a mediator in the relationship between making sense of the loss and well-being, suggesting that “the impact of making sense on improving psychological well-being is via the decreased levels of rumination that occur as a result of making sense” (p. 452). Thus, there is consistency across the studies so far with respect to a relationship between rumination and poor mental health.

However, the need for further refinement in both measurement and with respect to theoretical/conceptual clarity is suggested in a further study. Using data from a randomly selected community sample of adults, Treynor et al. [71] examined the two components of rumination described earlier, finding them differentially linked to outcome: reflective pondering was found to be adaptive and related to a lowering of depressive symptoms over time (although related to more depression concurrently), while brooding was maladaptive (both concurrently and in longitudinal analyses). In the present context, these results suggest the need to separate out a more active, elaborative aspect (reflection/pondering) from a more passive, repetitive (brooding) process, and that while the former may be helpful in the long run, the latter may not. This distinction is compatible with our analysis of cognitive processes (see below).

The relationship of rumination—more specifically—to CG (defined as similar to “chronic grief” and now called PGD, prolonged grief disorder, [56]) has been described in the following way by Prigerson et al. [57]:

“In contrast with bereaved survivors with uncomplicated grief, those with PGD are essentially stuck in a state of chronic mourning ... Much of the mental anguish stems from a psychological protest against the reality of the loss and a general reluctance to make adaptations to life in the absence of the loved one. … Such persons are often preoccupied by their sorrow and regrets concerning the loss. Their ruminations and inability to concentrate on things aside from their loss, the sense of feeling disconnected from people with whom they were close prior to the death exacerbates the sense of alienation and social isolation. Recurrent, intrusive and distressing thoughts about the absence of the deceased make it difficult for persons with PGD to move beyond an acute state of mourning and live in the present.” (p. 8)

In this description, extreme preoccupation with the deceased is seen as going hand in hand with—even deriving from—psychological protest against the reality of the loss. The question is whether protest against the reality of loss is similar to avoidance of this reality. If it is, then Prigerson et al. [57] might agree that avoidance processes are at play in (extreme) ruminative coping with bereavement. We return to the issue of rumination as an avoidance process later on.

It is important to emphasize that rumination is also part of normal grief [55, 64]. According to Nolen-Hoeksema [46] higher scores on rumination will be associated with greater distress, while only at extreme levels will rumination (and the associated avoidance, as we will argue) be a problem. In the case of bereavement, we would argue, extreme rumination could be expected to lead to complicated/chronic grief.

Like rumination, worry has also been associated with a variety of disadvantages and debilities including anxiety and depression ([13, 41]—both in [16, 18]), and including increasing distress more generally, and predisposition to physiological disruption and disease [62]. A recent review of the literature linking worry with poor health outcomes can be found in Brosschot et al. [21].

There may sometimes be advantages to worrying [23, in 16]. As Borkovec and Roemer [16] noted, identification of the maladaptive outcome of worrying has stemmed from research on chronically anxious groups, whereas emphasis on the possible positive effects of worrying have come from research on normal subjects. Whether there are similar advantages to rumination remains to be seen: Perhaps it could have positive effects in moderation; or if it is not detached/“worry-like”.

Mechanisms and functions of rumination and worry

We cited statements of Nolen-Hoeksema [46] and of other investigators at the beginning of this article, describing ruminative coping with bereavement as a confrontational strategy. The underlying mechanisms that Nolen-Hoeksema has described seem to endorse this: Rumination may enhance the effects of distressed mood on thinking, drawing people’s attention to the negative thoughts and memories made salient and accessible by negative mood. It may interfere with good problem-solving, because people are thinking so negatively about themselves and their lives. It may impair instrumental behaviors because, while ruminating, people are not engaging in the everyday activities that can increase their sense of control and lift their mood. So all of these mechanisms—drawing attention to negative mood; thinking negatively about the self; not engaging in distractive activities—seem to relate to confrontation processes.

However, an alternative conceptualization can be derived both from recent work in the field of bereavement (which has so far been mainly theoretical), and, more extensively, from the worry research area. In the bereavement area, Boelen et al. [7]—following the cognitive behavioural therapy (CBT) tradition—argued that the elicitation of certain thoughts and feelings (akin to rumination) may reflect anxious avoidance. According to these researchers “Anxious avoidance strategies occur when mourners believe that confronting the reality of the loss—that is, confronting feelings, thoughts or memories linked with it—will lead to “madness”, “loss of control” or otherwise “unbearable” consequences, and they consequently engage in attempts to avoid confrontation with this reality to ward off this threat.” [7, p. 115]. In line with such strategies, it was noted that mourners may engage in avoidance of situations (places the deceased used to visit), people (who might ask about the deceased), or objects (pictures of the deceased), all of which could elicit feelings or thoughts about the loss [32, 58]. Of particular importance to the argument that rumination is an avoidance strategy, these investigators postulated additional characteristics of anxious avoidant persons’ strategies: (1) they engage in counterproductive cognitive strategies to deflect from unwished feelings and thoughts; (2) there is anxious suppression of painful memories about the events leading up to the death; (3) they engage in continuous rumination about their own reactions or reasons why the loss occurred (to escape from having to admit to the loss and the emotions linked with it). Thus, according to this conceptualization, rumination can reflect avoidance of the reality of the loss and emotions that are associated with this reality, and, as such, be related to CG.

Further theoretical underpinnings were recently described by Boelen [4], who contrasted CG patients who anxiously avoid all reminders of the deceased/death-event, with those who continue to approach reminders of the lost person (e.g. dwelling on and yearning for the deceased, pouring over pictures and thinking back to the time he/she was still alive). Patients with the latter type of “preoccupied” responses do not appear to avoid at all. However, Boelen argued that these patients do refrain from confronting, elaborating, and adjusting to the reality of the loved one’s death and that they do indeed get stuck in their process of recovery. Although rumination was not explicitly referred to in this paper, the so-called preoccupied responses are quite similar to rumination.

What evidence is there for the theoretical propositions outlined above? A prospective study by Boelen and colleagues [5] provides some indication that rumination is an avoidant process in bereavement. Here they defined rumination as deliberate pondering on a narrow aspect of the loss, namely the events leading up to it. In line with their CBT approach, they reasoned that rumination can be an avoidance strategy because pondering on one particular feature of the loss can serve as an escape from admitting to the meanings and implications of the loss itself. In accordance with this notion, the authors found that items representing rumination (e.g. “I keep on pondering about who is to blame for the loss”) and those tapping behavioral avoidance (e.g. “I avoid places that remind me of the deceased”) loaded on a single factor, with factor loadings all being >0.60 [7], suggesting that these rumination and avoidance items are tapping similar concepts. Interestingly too, indirect support for the argument that rumination is an avoidant coping strategy comes from one of Nolen-Hoeksema’s own studies, which found rumination to be positively correlated with items on a scale measuring the tendency to avoid one’s mood through reckless behaviors such as excessive consumption of alcohol [see 49].

In this context it is also worth noting that in cognitive behavioural accounts of post-traumatic stress disorder (PTSD)—a disorder that bears resemblance to CG—rumination has also sometimes been conceptualized as avoidance. For instance, Ehlers and Steil [27] suggested that rumination may be one of the major strategies of cognitive avoidance in chronic PTSD, in that it focuses on experiences surrounding the traumatic event rather than the processing of the event itself [25]. In several studies, Ehlers et al. found evidence that rumination is a strong predictor of persistent PTSD [26]. In addition, the notion that rumination can be avoidance parallels the assumption of Jaycox and Foa [34] that a persistent focus on anger cognitions in trauma victims can interfere with the process of confronting and emotionally processing the traumatic event, thereby prolonging dysfunction.

Obviously, more empirical investigation of these concepts and their role in short- and long-term adjustment to bereavement is warranted. However, the above interpretation of rumination as avoidance receives strong support from the worry research domain. Research has more clearly and more unequivocally explained the relationship between worrying and negative outcomes (e.g. anxious experience, depressive affect) as due to avoidance or suppression processes [16, 18]. According to Borkovec et al. [18] “worry partly functions as a cognitive avoidance response to threatening stimuli…Worry distinctively involves a predominance of verbal thought whose function appears to be the cognitive avoidance of threat” [18, p. 573]. These investigators described things that people might be trying to avoid through engaging in worry, suggesting several functional levels of attempted avoidance and reviewing bodies of research that provided support for their lines of reasoning. The functions of worrying include: suppressing somatic anxiety; avoiding negative events or to prepare for the worst; distraction from more emotionally laden topics [16].

In one of their empirical examinations, Borkovec and Roemer [16] looked at the role of distraction in a study of the functions of worrying among persons with high (meeting GAD diagnostic criteria) versus low anxiety. Using the Reasons to Worry Questionnaire (described above), these investigators found that GAD respondents discriminated from normal respondents in their particularly high ratings of “distraction from more emotional topics” as the reason for their worry. The item for this function was “Worrying about most of the things I worry about is a way to distract myself from worrying about even more emotional things, things I don’t want to think about”. This, then, can be considered an avoidance item. The results support the view that worrying is done to stop an anxious person from worrying about even more emotional topics, but that this avoidance may maintain and strengthen anxiety. The notion that worry may serve as a distraction from more emotionally laden topics, seems particularly relevant for bereavement. It is in line with our idea that rumination, as a past-oriented “worrying” about particular aspects of the loss, may sometimes serve to keep attention away from things that are even more difficult to confront (such as thoughts about the irreversibility of the separation and feelings linked with this reality) and, as such, represents a form of avoidance.

Worry is seen as a predominantly conceptual, verbal-linguistic process, as opposed to an imagery process. In contrast to more imagery processes (which have strong efferent commands with emotions), such an abstract conceptual process does not have strong connections with the affective and physiological system. Thus, excessive thought in worry results in the avoidance of fearful images, and can thus be seen as avoidance of emotions and somatic activation. Several studies have supported this idea [14, 15]. For example, a negative correlation was found between percentages of thoughts and number of symptoms. When anxious, fewer somatic symptoms were experienced when worry was more composed of thinking. Furthermore, greatest suppression of heart rate response to phobic scenes, as an indicator of emotional processing, was found in participants trained to focus on thought during worry and in those participants where imagery during worry was most successfully reduced [14].

In other words: Since worry is a verbal process, which counteracts the activation of the affective system, the emotional processing of an event is suppressed and thus change in anxiety networks is precluded. Events do not become integrated in the relevant memory structures and processing does not occur. Worrying is a conceptual defensive process that seems to serve to maintain anxious concerns, especially so in GAD [15].

In conclusion: Boelen et al’s [5, 7] recent theoretical formulations are more in line with research conclusions in the worry area (i.e., that worry, as representing the future-oriented counterpart of rumination, is a predominantly verbal strategy that results in the avoidance of confrontation with unpleasant emotional material) than the earlier statements of Nolen-Hoeksema [46]. Researchers in the worry area clearly state and provide empirical evidence for the hypothesis that (extreme) worrying is an avoidant strategy [18]. Given the similarities between the concepts and the consequent relevance of the findings with respect to worry, there are good reasons to argue that high levels of rumination serve avoidant functions.

Implications: intervention for complicated grief

We have suggested that high levels of rumination are related to complicated grief. What, then, is to be advised to help ruminators? Nolen-Hoeksema [46] proposed a number of strategies for intervention, some of which differ from those emerging from our own perspective. She recommended positive distraction (rather than suppression), following the success of a distraction induction in reducing dysphoric mood and improving problem solving among ruminators, in her laboratory studies. This method “…provides a pleasant alternative to negative thoughts, without inducing anxiety over whether one is going to be successful at not thinking about one’s distress” (p. 555). Following our own analysis of rumination in terms of avoidance, our guidelines would place less emphasis on distraction, and more on the confrontation of those aspects to do with the loss that have not been dealt with, with the objective to raise not just cognitive, but also emotional awareness of the reality of irrevocable loss and to help the bereaved individual to (gradually) loosen the tie to the deceased person and to help them realize (in a deeply “gut-feeling” way) that they will not see, touch or smell their loved one again on this earth. In the context of cognitive behavioural treatment for people suffering CG, exposure interventions could be used to streamline confrontation with the loss, while cognitive restructuring could help to target cognitions underlying resistance to this confrontation [7]. Recent randomized clinical trials have found convincing evidence for the efficacy of the widely used CBT-techniques of exposure and cognitive restructuring for the alleviation of complicated grief [8, 63, 72].

Nolen-Hoeksema also suggested assistance in the creation of meaning in the loss (e.g., starting a memorial fund; dedicating something to the deceased), thereby providing a constructive alternative to rumination in retaining the tie to the deceased. From our perspective, such a strategy could assist in making the shift from unhealthy rumination to healthy grief work—which could take the form of adaptive cognitive processing. However, clinicians should be very alert to the possibility that, for example, starting a memorial fund, placing a memorial stone at the border of the highway where the loved one died, or frequently visiting the graveyard, may in fact function as another way of not fully accepting the reality of the loss. Confronting and elaborating the reality and implications of the loss and integrating this reality into pre-existing knowledge about the self and the world [7, 42] is, in our view, the core business of grief. Therefore ‘exposure’, in whatever way, is the core task of the therapist in helping a patient with complicated grief. Unlike rumination, such processing, would—as has been demonstrated empirically—facilitate recovery from loss [19, 24, 65]. Ruminators acknowledge searching for meaning, but are actually less likely than non-ruminators to say they can make sense of their loss [46]. However, in our view—and we elaborate below—deep-down acceptance that irrevocable loss has taken place is more fundamental to good adjustment than searching for meaning.

Another suggested strategy of intervention was guidance in engaging in problem-solving, to do with concrete difficulties in daily life that have emerged as a result of loss (in order to have less to ruminate about and opportunity to gain control). This is in line with Nolen-Hoeksema’s [46] finding that ruminators often say they focus on their emotions in order to try to understand and solve their problems, yet they do not actively engage in effective problem-solving behaviors, such as making plans of action which would likely help them to deal with their problems (ruminative coping and problem-solving coping were found to be negatively correlated). Thus, making action plans might usefully be incorporated in intervention, since, we would also argue, such action represents confrontation with problems associated with grief and grieving. However, following our own reasoning, ruminators would have more problems in dealing with the loss of the loved person per se, rather than in dealing with secondary stressors, so again, confrontation with the reality of loss (for example, by throwing or giving away the clothes of the loved one) would seem more critical than everyday problem solving.

Finally, Nolen-Hoeksema [46] suggested helping bereaved ruminators work more effectively with their social support network. Supportive others can serve a number of functions, including: “simply telling the story of his loss over and over to emotionally supportive others may help the ruminating parent to habituate to the story and shape the story to fit his belief system” (p. 558). All these strategies are aimed at helping the bereaved ruminator “find an adaptive middle ground between suppression and rumination” (p. 559), which were referred to as the “maladaptive extremes” (p. 559). As suggested by Nolen-Hoeksema, improved use of social support networks could indeed assist engagement in active problem solving (and regaining control), although, as mentioned above, problem solving may not be the main root of the problem in the case of excessive rumination. Furthermore, although using others for the repetitive retelling of the bereavement story may serve to habituate to the loss (particularly in cases of extremely distressing losses such as death of one’s child), this repetitive retelling could as well function as another manifestation of rumination. We would emphasize the greater need to curtail continuous rumination and to face up to realization and acceptance of the loss and emotions linked with it.

Conclusions

We have reasoned that rumination may function in an avoidant way: Sometimes, the bereaved ruminator may be avoiding what is simply too painful to confront. The reality of the loss may be potentially too negative and emotionally overwhelming to face up to. In other cases, ruminative coping may be caused by a desire to be loyal to the deceased. To acknowledge that one’s loved one will never return implies a need to move on without him/her, and this may feel like desertion, it cannot be faced up to either. To avoid this, the person may engage in counterfactual thinking (generating imagined alternatives to actual events) in “if only … then …” reasoning (e.g. “If only I had forced him to take his medicine, he would still be here”). This too implies an avoidant strategy, in the sense of avoiding acceptance of the loss and the need to live without the loved one. Rumination may act as a distraction from more emotionally laden topics and probably—at least in moderation—is a normal and quite functional strategy that is widely used in the first weeks/months after bereavement.

What are the implications of this reasoning for how grieving is conceptualized? The revision leads to fundamentally different ideas about the nature of healthy versus unhealthy grief work. The thrust of Nolen-Hoeksema’s research was that focus on negative grief-related emotions leads to poor bereavement outcome. Her empirical research provides strong support for this claim. She reported that ruminators neither do necessary things such as good problem-solving, nor engage in effective instrumental behaviors/healthy everyday activities/making plans of action/finding meaning in their loss [46]. She argued that it is the ruminative way of coping with grief that causes these sorts of problems, rather than the avoidance of dealing with such aspects that causes rumination and intensified negative emotions. In other words, not engaging in aspects such as problem solving or meaning making were the result of ruminative coping.

By contrast, in our view, it is the avoidance of facing up to the reality and pain of loss, that underlies rumination.Footnote 5 Healthy grief work involves good problem-solving, engaging in effective instrumental behaviors/healthy everyday activities / making plans of action/finding meaning in loss—all of which entail facing up to the reality of the loss. Confrontational ruminative coping is motivated by an avoidance of healthy but painful grieving, and these avoidance processes impact on adjustment. According to our formulation, ruminative coping is not an “opposite” form of coping to denial or suppression, as Nolen-Hoeksema claimed, but rather, denial and suppression are underlying and integral to the adoption of a ruminative coping style.Footnote 6 As such, ruminating about one’s feelings and associated problems can serve as an “excuse” to avoid the presumably even more painful work of admitting and adjusting to the loss of the loved one (grief work). Rumination—in the extreme—is symptomatic of an inability and/or fear-driven unwillingness to face up to the reality of loss, and it is this inability and unwillingness that causes complications, not (only) ruminative coping (per se).

Perhaps it would not be too difficult to reconcile Nolen-Hoeksema’s with our own position. There are indications that she understood rumination to reflect a too-narrow focus in terms of accepting the reality of the death: “…people who cling to their ruminations may be, in a way, clinging to the deceased loved one in the only way that is left to them” [50, p. 69]; “Some bereaved people may be reluctant to give up their ruminations … because it represents for them their last and final tie to the deceased” [50, p. 68]. Although there is no explicit mention of avoidance, these statements imply that there is assumed lack of acceptance of the reality of the death, which—as discussed previously—is a major problematic component in complicated grief. Furthermore, Nolen-Hoeksema contrasted rumination with reappraisal in a manner that is strongly reminiscent of the brooding/reflection distinction described earlier, that was identified by Treynor et al. [71]: “Whereas ruminating involves going over things you wish you had done differently and worries about how bad you feel and whether you will ever recover, reappraisal involves the search for a positive way of framing one’s experience” (Nolen-Hoeksema & Larson, p. 69).

We have examined the proposition that ruminators are avoiders theoretically, drawing on available evidence from different fields. However, the proposed avoidant functions of rumination specifically in bereavement need empirical exploration. Is our reasoning correct: Is it the avoidance of important aspects of grieving such as accepting deep down that the loved one has died and that one will never see or feel him/her in this life again, and consequent reappraisal, meaning making, problem solving—or any other things that a therapist might have to deal with in cases of chronic/complicated grief—that cause adjustment problems, rather than (or in addition to maybe) the fact that emotions and symptoms are focused on? Future theoretical and empirical research needs to explore (1) whether denial and suppression processes operate in ruminative coping with negative emotions in the ways suggested above, and if so (2) how these avoidance processes and rumination are actually linked and cause ill effects.