Revista Científica Arbitrada de la Fundación MenteClara

1. Introduction

Epistemologically, John Bowlby's theory of attachment has roots in the Darwinism, and the dualism and structuralism inherited from the Psychoanalysis, specifically from the branch of the British Object Relations School. His theory stresses the role of adaptation of the organism in an environment whose attachment, as an instinctual system organised to improve the chances of survival, is close to Fairbairn's (1994) notion of drive as object seeking.

Thus, the attachment theory is grounded on notions such as object relations, on concepts from Ethology and on the cognitive dimensions of the mind -i.e. the capacity to internalize behavioural patterns and the formation of episodic and semantic memories of experience-. In this context, the theory of attachment explains the survival strategies of beings and is based on the principle that parents would support their offspring in situations of relational stress generated by events or environments experienced as vulnerable. In this way, attachment behaviours have been developed according to the possibility of maintaining the proximity of parental figures through indicators of interaction, aiming to attend their needs, and is modulated progressively by the social environment (Bowlby, 1969, 1973).

The fundamental basis of attachment behaviours is built on the representational capacities of the integration of information from the environment. This representational dynamic is referred to as the Internal Working Model – IWM (Bowlby, 1973, 1980).

According to Bowlby (Bowlby, 1969, 1973) and to empirical studies which verified this conceptual approach (Ainsworth, Blehar, Waters, and Wall, 1978) (Main, 1990) (Bretherton and Munholland, 1999), all the activations of this system articulate emotional and factual experiences that make up the internal models of the child -and then, of the adult-. For example, when a disruptive experience occurs, the individual experiences it as a state of vulnerability, his attachment system gets activated, triggering protection-seeking behaviours. If the caregiver responds appropriately and consistently, the individual will regulate his responses and return to a state of homeostasis and, due to regularity, will organize a pattern that will be internalized as "secure".

If, on the other hand, the caregiver does not respond or responds inadequately, the individual will either learn to inhibit his demands, internalizing an 'inhibited' –deactivated- model whose dominant representation is that the likelihood of maintaining tolerance and bonding with the attachment figure is stronger when demands decrease or stop; or activate them even more – here the dominant internalised representation is that exaggerating demands –hyperactivation- increases the likelihood of a favourable response from the figure of attachment. These are the secondary strategies. That is, implementing additional strategies to cope with the weakness or failure of his attachment figures to meet their needs. Between these two strategies -inhibition and hyperactivation-, there is another one where the individual does not know how to elaborate and choose a response in a coherent fashion. This strategy is called 'disorganised' and occurs when attempts to adapt to the attachment figure frequently fail. The individual may present as hyperactive and then shut down, becoming inhibited. This disorganisation takes place in a permanent internal conflict and in a constant state of internal tension. This is the case with abused children -they are afraid of the figure of attachment, but also need her-, or children whose parents have been traumatized or bereaved -they need her, but at the same time, they withdraw- (Main and Solomon, 1986).

Nevertheless, despite the fact that the IWM is constructed from the interactions experienced with caregivers, a filter is established and other information from the new experiences is inserted into this model, but according to this filter. From then on, the individual can alter his perceptions to bring them closer to the representations already formed. However, the IWM can also be influenced by the defensive models that the individual puts in place. If his defences do not allow for the integration of new information, then the IWM cannot be modified and, as a result, the individual will not adjust to present experiences. Bowlby (1969) called this process 'defensive exclusion'. The explicit or implicit messages transmitted by attachment figures may be inadequate to the constituted attachment system. For this reason, according to him, one can observe a repetition of attachment patterns in adult life, as well as a repetition of these patterns with their own children.

Based on these considerations, other studies have verified that the same patterns operate during different phases of life (Kobak and Hazan, 1991) (Kunce and Shaver, 1994) (Miljkovitch, 2003) (Cohin and Miljkovitch, 2007) (Cyrulnik, Delage, and Blein, 2007). In studies experiences of break-up of romantic relationships, attachment becomes an important variable to consider.

This article relates psychic processes – such as the organization of episodic and semantic memories, the description of the perceived stressful experience, and especially the perception of the behavioural operation to cope with it. In view of the extensive empirical production on this individual, there is a convergence on the assumption that a person's attachment experiences in a romantic relationship are associated with memory traces and/or memories related to their internalized attachment patterns from their relations with their invested external objects -figures of attachment-. Once the operating features of the experiences are determined by the articulation between affect and representation, it is necessary to account for the way in which such articulation occurs. The experience of relational stress can infer the emergence of a disruptive experience. It is argued that the disruption associated with the impact of relational stress experiences on the individual's psyche can be verified –inferred- from the investigation of the memory of attachment experiences in relation to their differentiated external objects, namely, their parental figures and their romantic partners.

This conceptual approach is based on the general assumption that a disruptive event impacts on the processing of experience and that this processing is mediated by the individual experiences with attachment figures, impacting on the ability to cope with relational stress.

The disruption is a term proposed by Benyakar (2006), which refers to a phenomenon generated by the impact of the factual -a situation that may be internal, somatic; or external, non-somatic- in the psyche. Such an impact can produce a disorganising effect that triggers an experience of discontinuity but also, in some cases, might be transformative. Thus, it is the way in which the individual metabolizes and/or transforms the encounter between the internal or external factual with the psychic that defines the experience.

According to this author, the experience can be characterised by the articulation between affect and representation, that is, a coherence between the memory traces of experiences and their representations, which signal a metabolized experience. For example, in the context of romantic relationships, during a relational stress situation -factual disruptive event-, what the individual experiences is perceived in congruence with what he represents of his relational experience with the other -the external object invested-, and the affect and behaviours performed to cope with this stress are congruent with the representation produced.

On the other hand, the experience may present itself as a non-pathological transient dysregulation, whose affect articulates to a representation, but incongruently. In the relational stress situation, for instance, what the individual experiences is perceived, at times, as not congruent with what he represents of his relational experience with the other, but associates to the retrieving of another factual event in his repertoire. This retrieving would trigger a similar response to this event as did in another one in the past. Here, despite the experience being sufficiently elaborated and metabolized, the behaviours performed to cope with the stressful situation are based on other affect and representations infiltrated by reminiscence of past experiences.

The experience can also be presented through a sustained and pathological dysregulation. In this case, the experience cannot be metabolized because the processing is blocked by a crystallised input from past experiences. Benyakar (2006) has conceptualized this input as an “incrustation”. This is, for example, the case when the individual still experiences stressful situations as not congruent with what he represents of his relational experience with the other, by recalling another factual event from the past, but re-experience it as a repetitive pattern. Here, the affect and representation are crystallised and not metabolized. The individual experience and responses are stuck, living in the present a disruptive experience from the past since the disruptive phenomenon remains in the psyche in its perceptual dimension as a repetitive recall.

The lived experience can also suffer from a disarticulation effect, which is marked by a rupture between affect and representation. Psychic processes try to construct meaning through experience, but in vain. Here, the experience is qualified as "traumatic". This is the case of individuals who have experienced, for instance, abuse by their significant objects, from whom they wanted to escape but, at the same time, relied on them to survive. Their behaviours are often expressive of a representational and behavioural disorganisation tainted by a traumatized psychic organisation.

In this sense, the relational stress experience can be conceptualized as a disruptive factual phenomenon of traumatic potential. However, if it triggers a transient dysregulation between affect and representation, the psyche can still mobilise itself to develop coping capacities. Thus, it can re-establish the regulation between affect and representation - which would not be the case if the experience was traumatic, as this would generate a disarticulation between affect and representation that would restrict the capacity for elaboration.

In this context, relational stress is superimposed as a state of vulnerability. For Zukerfeld and Zonis Zukerfeld (2016) vulnerability is the difficulty in adapting and developing mentalized responses to adverse conditions, which is mediated by the perception of support, the regulation of self-esteem, the stress itself, and, above all, the quality of bonding support –attachment-, because, as the authors state, it is the deployment of the experience of support that ignites the co-metabolic factors to cope with adversities.

This study was conceived as a posteriori analysis of the results of an investigation conducted as part of my Master’s thesis. The thesis was on the relationship between the experiences of romantic relationship break-up, internalized attachment representations, and the strategies put in place to cope with these experiences. While the original study sought to identify these relationships, the current study aims to verify, in light of disruptive phenomena, the idea proposed by Zukerfeld and Zonis Zukerfeld (2016) on how attachment representations mediate the experience of vulnerability associated with the break-up experiences vis-à-vis patterns of behaviours put in place to cope with it and those of grieving. The hypotheses of this study are as follows: 1) The more attachment representations, in relation to attachment figures, are internalized as secure, the more elaborated and accomplished is the experience of grief; 2) Individuals with inhibited grief do not describe a significant grief experience, unlike those who describe experiencing hyperactivated grief; 3) The more secure the relationship with attachment figures, the less the strategies of denial, detachment, self-blame -since they are limited strategies and do not relate to support from significant figures- were used to deal with the experiences of break-up.

2. Materials and methods

Forty-four participants were solicited on a voluntary basis, in the suburbs of two urban centres in north-eastern Brazil -Fortaleza and Teresina-. They all consented to the recording of the interviews. The criteria for participation were: aged between 18 and 64 years; having experienced a break-up of a romantic relationship which last, at least, one year; having, at least, one living parent; not having consulted mental health professionals in the last six months before the interview. Most participants were female (N=30, 68%). They were 29 years old on average (standard deviation = 9 years, minimum = 18 years and maximum = 63 years). While all participants had at least one romantic relationship, 28 (63%) had two, twelve (27%) had three and four (9%) had four partners. Most of these relationships were heterosexual (N = 34, 77%), six (13%) were same sex identified and four (9%) were bisexual identified.

The attachment experience covers two dimensions: sense of security and activation of secondary responses to maintain this sense of security. The feeling of security describes the representation regarding the greater or lesser confidence in the sensitivity of the attachment figure. The activation reactions refer to the representations of confidence or lack of confidence in the permanence of the bond and care.

These dimensions were measured using the Attachment Security and Secondary Strategies Interview (ASSSI, developed by Miljkovitch, 2009). The ASSSI is a semi-structured interview on attachment experiences with the parental figures and romantic partners. The ASSSI also makes it possible to evoke family dynamics and memories of these dynamics in circumstances of illness, danger and disputes, as well as around their availability to the individual. In the case of romantic relationships, it also evokes the dynamics of the relationship, as well as the circumstances of the break-up. This is done through detailed descriptions of their feelings and actions in relation to the different situations experienced with the attachment figures.

Each participant was given a security score and an activation score with each attachment figure based on those descriptions and were based on the ASSSI coding scale. The scales were conceived by the conceptual perspective that the attachment experience is a continuum. Therefore, each participant received an score for activation and deactivation. Two coders analyzed the interviews -the author and another Brazilian psychologist familiar of the instrument-. When the scores assigned did not differ by more than one point, the average of their two scores was kept. Inter-judge agreement was significantly correlated (r = .57, p < .001).

The Scale of Security and Activation of Attachment: The attachment security scale ranges from 0 to 8, scored in pairs to maintain sensitivity. A score of 0 describes no security, the individual feels totally rejected or unaccepted by the attachment figure even though in a parent/child relationship basic care is provided, or the individual fears for the security of the bond and that it may break. Conversely, a score of 8 describes total security, the individual feels that the attachment figure's responses are adapted to his needs, he describes a feeling of satisfaction with the relationship and feels free to express his affects or opinions in the relationship.

The activation of attachment scale ranges from – 8 to 8. Negative scores correspond to so-called inhibition strategies and positive scores correspond to hyperactivation strategies. A score of 0 corresponds to optimal activation -generally observed in secure participants-, with the individual not using any activation strategy. As with the previous scale, the scores range in pairs. The score – 8 corresponds to extreme inhibition, which is expressed both in behaviour -i.e., the individual completely withdraws from the relationship with the attachment figure- and in thoughts -i.e., such as denial of one's own affects and needs-. Score 8 corresponds to extreme hyperactivation, the individual develops behaviours of disproportionate demand for attention, to the extent that there can be no real threat in reality.

The Specific Experience of Grief scale: The Specific Experience of Grief (SEG) assesses the physical and emotional reactions to the experiences of the break-up with an attachment figure. It is a self-administered questionnaire. The questionnaire was translated into Portuguese and adapted from the Revised Grief Experience Inventory by Lev, Munro and McCorckle (1993). This questionnaire has 22 items. The scores range from 1 (= strong grief experience) to 6 (= very weak grief experience). The α-Cronbach's measure was .84, which is considered a high level of reliability.

Semi-structured Grieving Interview: In addition, the grieving experience after the break-up was also. Another dimension was identified: "Grief activation". It refers to the cognitive and emotional reactions related to the break-up. It was assessed on the basis of the ASSSI analysis model, using a pairwise scoring system. The scales were also conceived by the conceptual perspective that the grieving experience presents in a continuum. Therefore, each participant received an score for activation and deactivation Participants were scored according to characteristics of grief resolution or non-resolution, ranging from – 4 to 4. A score of zero corresponds to a resolved grief. Extreme scores correspond to unresolved grief. Participants with a score of – 4 showed extreme grief inhibition -i.e., characterized by denial and avoidance of grief-. Those with a score of 4 showed hyperactivation of grief -i.e., characterised by acceptance of the break-up and chronic suffering-. Here again, inter-judge agreement showed a significant correlation (r = 0.63, p < 0.001).

Carver’s Coping scale: Finally, coping strategies were assessed by another self-administered questionnaire. For this purpose, Carver's (1997) Brief COPE was translated into Portuguese. The questionnaire assesses 14 coping strategies. Participants rated each item using a Likert-type scale ranging from 1 (= I never did that) to 4 (= I did that a lot). The α-Cronbach's measure was .73 considered an acceptable level of reliability.

3. Results

Descriptive and inferential analyses were conducted to test the different hypotheses. Recall that there were 44 individuals, all of whom had experienced a relationship break-up and 36 of whom had previously had another partner. If considering the generally accepted threshold of p <. 05 to reject the null hypothesis, one is also interested in the significance threshold of p < .10, given the size of this population.

Since most of the variables are continuous variables, correlations have been made. The results are presented in the order of the assumptions made earlier. First, the results concerning the stability of the attachment models are presented, followed by those concerning the relationship between the grief experience and the attachment models. Finally, the last section looks at the links between coping strategies and the experience of grief.

Relationships between attachment patterns

In order to verify the relationship between the representation of attachment with the different attachment figures, correlations were calculated between the security scores and the activation strategies of the parent and partner figures.

See table 1.

Relationships between security scores according to the attachment figure.

The expected relationship between the security variables is linear since they have similar distributions. The results show that the correlation between the security score with the father and that with the mother is positive, strong and significant. The correlation between the security scores with the last partner and the security scores with the father is positive, but not significant. On the other hand, the correlation between the security scores with the mother is positive and significant (p < .05). The correlation between the security scores with the penultimate partner and with the parents is positive and the marginal level of significance (p < .10) could be explained by the small sample size. However, the correlation between the scores with the last partner and the security scores with the penultimate partner is very strong and the level of significance high (p < .0001).

Relationship between activation scores according to attachment figure.

Again, the results showed that the correlation between the activation score with the father and that with the mother is positive, strong and significant. The correlations between the activation scores with the last partner and those with the father and mother are positive, strong and very significant. Conversely, the correlations between the activation scores with the father and with the mother and those with the penultimate partner are not significant. On the other hand, the correlation of the activation score with the last partner is strong with that for the second last partner.

Relationships between security scores and activation scores.

Due to the different distribution of the scores for security (0 to 8) and those for activation (- 8 to +8), the expected relationship between the two is no longer linear. Indeed, a high score on the first variable (security) would be associated with a "median" score on the second (activation). This is why quadratic correlations have been calculated here. First of all, one can say that the correlations for each attachment figure between security and activation level are strong and significant. The results show that there is no correlation between the security scores with fathers and mothers, and the level of activation with the penultimate partner. Also, if in one hand there are positive and significant correlations with the level of activation for the last partner; on the other hand, there are correlations between the security scores and the activation levels for the last and penultimate partner.

Relationships between attachment patterns and grief experiences.

Relationships between the two measures of grief experience.

Henceforth, it no longer deals with the scores for the penultimate partner, as grief experiences are measured with the most recent break-up. Furthermore, the hypotheses focus on the relationship between grief reactions and the primary attachment figures -i.e. father and mother-.

Two variables referred to the experiences of grief during the loss of a partner for separation. These are specific experience of grief (SEG), measured by questionnaire, and grief activation, measured from the interview. The latter variable was designed as a continuous variable ranging from – 4 to + 4. However, it was chosen to transform it into a nominal variable. Thus, participants with a score between – 4 and – 1 were classified as having an inhibited type of grief. Those with a score of 0 are identified as having resolved grief. Finally, those with a score of + 1 to + 4 are classified as having a hyperactivated type of grief.

This classification was made because it was hypothesised that individuals with inhibited grief would not have the same characteristics, in general, as those with hyperactivated grief. Here, it was hypothesized that those with inhibited grief would not report a significant general grief experience, unlike those with hyperactive grief. If comparing the means obtained from the levels of grief through the SEG, self-questionnaire and the classifications of grief activation obtained through the scale with regard to the participants' accounts, using an analysis of variance (ANOVA), it is observed the following results:

See table 2.

In general, there was a tendency for the means between the groups to differ, F (2,41) = 2.12, p < .13. However, in line with stated expectations, it was assumed that it was mainly the hyperactive grieved who expressed a greater specific experience of grief. Therefore, a contrast analysis was conducted between this group and the other two, resolved and inhibited. This calculation confirms this hypothesis and shows that hyperactivated grievers present a stronger general experience of grief than the others, while inhibited grievers present an identical experience of grief compared to resolved grievers, t (41) = 2.02, p < .05.

Relationships between general grief experience and security and activation scores with different attachment figures.

Next, it was investigated whether the general grief experience is related to security and activation strategies of the attachment system with attachment figures. Linear correlations were conducted for the security scores and quadratic correlations for the activation scores.

See table 3.

Correlations of specific experience of grief (SEG) with security and activation scores for the representations of attachment.

The results showed that there is only one significant correlation, and marginally so, with the security score with the last partner. It appears that the overall grief experience of the loss of the most recent partner, for this sample, was relatively independent of attachment representations.

Relationships of grief activation with security and activation scores and different attachment figures.

This time, the aim was to test whether the particular grief experience of the last break-up, the way in which the grief is activated or not, is related to the security and activation strategies of the attachment system with the attachment figures. Linear correlations were sought for the security scores and for the activation scores. Here, it was kept the variable activation of grief in its continuous form, as no significant difference is expected between inhibited and hyperactivated grief in terms of security scores.

As for activation strategies, the relationship was assumed to be linear, otherwise when individuals tend to inhibit grief -negative scores- they would also tend not to activate attachment strategies -negative scores-. Similarly, those who tend to hyperactivate grief -positive score-) would also tend to hyperactivate attachment strategies -positive score-.

See table 4.

As for the correlations between security and activation scores with fathers and mothers, and grief activation, it can be said that they are not significant. In other words, there is no relationship between the experience of secure attachment, activation strategies with parents and the grief experienced during the break-up of the last romantic relationship. On the other hand, the correlations of security and activation scores with the last partners with grief activation are strong and significant. This indicates some influence of the secure or insecure model in couples on how grief is activated.

Coping strategies.

Relationships between different coping strategies

Before testing the hypotheses about the relationships between coping strategies, it was calculated the correlation between the different coping strategies. This allowed to see whether certain coping strategies tended to be found in the same individual. In synthesis, it was seen that participants who exhibit the coping strategy of self-distraction -i.e. TV, reading, going out- also tended to present a self-blame strategy (r = .46, p < .001). They also presented spiritual support seeking strategy (r = .42, p < .001) and a planning strategy (r = .36, p < .001). A few excerpts from the stories illustrate these reactions:

"I closed in on myself... I did a lot of things..., I reflected a lot, I meditated a lot, I practised yoga a lot... I took refuge in my spirituality a lot... I studied more... well, I decided to love myself more... And for me... but I confess that [silence] sometimes I felt frustrated..." (Patricia, 30)

"Honestly, I thought about suicide... but I have God in my heart... so I know I have my part of guilt and it was a difficult situation... so I put a lot of effort into my studies, I prayed... that's all" (Regina, 19)

There was also a correlation between drug -i.e. alcohol- use, and denial strategies (r = .35, p < .001), as well as the strategy of seeking spiritual support (r = .31, p < .001), as illustrated in the example below:

"I was drinking too much! My Lord Jesus Christ forgive me... I was praying to forget this demonic passion, but I was tempted a lot... and I was drinking, I was always drinking... [Silence] I sleep with one woman one day, the next day with another... it's temptation... my Lord Jesus knows how much I pray... to forget this woman... but I know I want to stay with her!" (Jose, 26)

On the other hand, there is also a significant correlation between drug use and the humour strategy (r = .32, p < .001). Here, another example:

"It's hard to live alone... but my belief teaches me to control anxiety states... so I try to live a normal single life [laughter and silence] ... but it's hard... because I stay anxious to meet another person... I live in expectation... so I always drink and alone... but I'm not an alcoholic eh! [Laughs]... I drink in my room!" (Joao, 31)

There was a strong correlation between the self-blame strategy and planning as well (r = .40, p < .001). Here, an example to demonstrate this correlation:

"I was too physically and emotionally fragile... [Silence]. I have no support... The break-up was very difficult... I didn't know what to say or what to do... I felt... I feel guilty... [Pause and laugh] ... actually... maybe... it was a good thing... an opportunity to grow for me... thanks to this guilt" (Livio, 36 years)

Secondly, the active coping strategy was strongly correlated with the planning strategy (r = .55, p < .001), the positive reformulation strategy (r = .68, p < .001), the spiritual support-seeking strategy (r = .61, p < .001) and moderately the instrumental support-seeking strategy (r = .34, p < .005). An example is as follows:

"I was trying not to remember! I was busy at work and sometimes I lost track of time! In fact, I got dizzy: I studied more, I went to church, I changed my look... you know? I changed my haircut. And... I decided to go out a bit more, to buy more clothes... to forget about the relationship! One good thing about it was the opportunity to change my life! I was depressed... it was horrible!" (Karim, 24)

Finally, participants who used the detachment strategy, also used break-up acceptance strategies (r = .42, p < .001). For example:

"Me, I was expecting it... I didn't feel anything special" (Valerie, 26)

"I worked more on my professional projects, in my studies, and I got closer to my friends... that's all" (Liz, 40)

"Totally relieved! Today I live happy and free!" (Ney, 47)

Relationships between coping strategies and the grief experience

In order to investigate the relationships between coping strategies and the type of grief, it was used the created groups -identifying participants as having resolved, hyperactivated or inhibited grief-. This nominal classification allowed to verify that coping strategies were not the same for participants with hyperactive and inhibited grief.

See table 5.

First of all, it can be noted that the three groups of participants do not differ in the use of humour (F (2,41) = 0.66, ns), positive reformulation, self-distraction (F (2,41) = 0.98, ns), and effusion (F (2,41) = 0.61, ns) strategies. Nor do they differ for the strategies of self-blame (F (2,41) = 0.54, ns), drug use (F (2,41) = 1.10, ns) and detachment (F (2,41) = 0.63, ns).

On the other hand, it was found three coping strategies which gave rise to differences between the groups, the significance of these differences being non-significant with the analysis of variance, but the calculation of contrasts between the groups made possible to reveal that one of the groups differs from the other two. These were active coping (F (2,41) = 1.75, ns), spiritual support seeking (F (2,41) = 1.49, ns) and instrumental support seeking (F (2,41) = 2.29, ns).

The contrast analysis verified that those with inhibited grief show less active coping strategies than those with resolved grief or hyperactive grief (t (41) = 1.85, p < .07). Similarly, they used fewer spiritual support-seeking (t (41) = 1.72, p < .09) and instrumental support-seeking strategies (t (41) = 2.02, p < .05) than the other two groups.

The results of the analysis of variance showed that there were significant (or a trend) differences between the groups in the use of denial (F (2,41) = 3.70, p < .05), planning (F (2,41) = 2.83, p < .07), acceptance (F (2,41) = 2.60, p < .09), and emotional support-seeking (F (2,41) = 5.19, p < .01) strategies.

Thus, it appears that those with inhibited grief used strategies of denial (t (41) = - 2.20, p < .05) and planning (t (41) = 2.05, p < .05) more than the other two groups. Those who presented a resolved grief showed more acceptance strategies than those with inhibited or hyperactive grief (t (41) = -2.20, p < .05). Finally, those with hyperactive grief sought emotional support more than those with inhibited or resolved grief (t (41) = 2.75, p < .01).

4. Discussion

To understand the relationship between attachment representations in relation to attachment figures and the grieving experience -first hypothesis-, it is important to comprehend how the representations of attachment are presented in this population. It was seen that there is no stability between the attachment experience with parents and romantic partners in terms of feelings of security; but that there is stability in terms of the level of activation reactions of the attachment system. Although a positive relationship between security scores with parents and last and penultimate partners was verified, it appears that this relationship is stronger with mothers. It can be assumed that the emotional memory of experiences with mothers may be more commonly retrieved than the memory of experiences with fathers.

A possible explanation for this difference could be the cultural context. The participants live in a region of Brazil where the patriarchal model is predominant and are part of a less privileged economically. In these microsystems predominates ideals of parental absolute authority, absence of emotional exchange and sharing of domestic activities, including care for younger children by older siblings (Montenegro, 2000). Even though the higher tendency to represent these relationships as insecure, the analysis of their answers about their attachment indicated there was a certain amount of trust in the relationship, as the participants were able to present some level of coherence of their episodic and semantic memories. This might suggest an internalization of a representational security. Therefore, when retrieving such memories, they did not repress them – or, as Bowlby (1969) postulated, engaged in defensive exclusion.

With regard to security with partners, it appears that participants were more interested in describing their love experience with the last partner than with the penultimate partner, probably because this relationship was perceived as the last “stable” relationship. The results showed that the security described with the last and penultimate partner was very similar, but much less so compared to the parents. It can be assumed that the assimilation of new concepts of love and self, favouring a more symmetrical relationship, could explain why the attachment representation towards partners, and particularly the last partner, appeared to be less influenced by the attachment representations associated to their parents. This is reinforced by the fact that there is an increase in security scores with the last partners compared to the penultimate ones.

The representations of attachment in relation to different attachment figures remains relatively stable when those a secure in relation to the parental figures. This is in line with the considerations of Kobak and Hazan (1991), Kunce and Shaver (1994), Miljkovitch (2003) among others. In that sense, the representations of attachment mediated the access to mnemonic information and attachment response. As asserted Bretherton (1990). That explains why attachment representations appear to be more stable and more easily retrieved than other types of stored memory. Therefore, if the individual had internalized an insecure experience of attachment with their parents, he will, most likely, present a tendency a tendency to present inhibited representations and deactivate his attachment behaviours with their last during vulnerable situations. On the other hand, if he had internalized a secure experience with their parents, there is a tendency to activate the behaviours during vulnerable situations, but return to homeostasis when security is re-established.

Among the participants of this study, since most of them presented lower security scores, it is possible that representations of 'insecure' relationships with parents were able to be retrieved through episodic memory and transmitted to their romantic relationships. On the other hand, one must remain circumspect with this assertion because of the absence of a correlations between the scores for security in relation to the fathers and the last partners, and the marginal correlations in relation to the penultimate partners.

The relationship between activation scores from the attachment figure.

Concerning the activation of attachment behaviours, it can be said that the more activated the relationship with the parents is, the more it can be activated with the last partners. Thus, the more inhibited the representations of attachment in relation to the parents is, the more inhibited in relation to the last partners as well. The same reasoning is observed between partners. In this study, the participants showed a tendency to experience romantic relationships through inhibited attachment behaviours. Despite this, there is some coherence between their episodic and semantic memories when talking about their romantic partners. According to the literature, it is possible that these adults have re-elaborated their insecure representations from past romantic experiences and tried to develop more secure relationships, although they still remain an internally insecure: this is known as representational security (Miljkovitch, 2009). It can also be explained by increasing symmetricity within the romantic relationship

Thus, among the participants of this study, if the relationship with the penultimate partner was represented as insecure, there was a tendency to repeat this insecure pattern in the last relationship. However, despite this tendency, it was noted that the level of inhibition towards the last partner decreased. In line with the current literature, this observation is explained by the idea that insecure attachment in relation to parents seems to influence experiences with romantic partners, (Miljkovitch, 2003) (Cohin and Miljkovitch, 2007) (Cyrulnik, Delage, and Blein, 2007) (Guedes and Moreira, 2020).

The relationship between security and activation scores from the attachment figure.

The results indicated an expected trend: the predominance of inhibited representations of attachment related to insecure attachment, which in the population studied highlighted the inhibition. However, such trend was stronger between last and before last partners. Similar patterns were observed in their experiences of grief and coping with the romantic separation. Again, this similitude can be explained by the fact that individuals who had internalized insecure attachment representations tend to repeat attachment patterns of behaviours in later relationships (Cohin and Miljkovitch, 2007) (Cyrulnik, Delage, and Blein, 2007) (Guedes and Moreira, 2020)

The relationship between attachment patterns and grief experiences.

Given that participants inhibited their attachment system with their parents and partners, and based on current literature about the discourse of individuals who internalized insecure attachment (Bretherton, 1990) (Pierrehumbert, et al, 1996) (Miljkovitch, 2001), it can be suggested that participants might be biased on their perceived experience. First, one can assume that, the experience itself and the appraisal of their own experience may have impacted the processing of the experience. For instance, that could indicate that the semantic memory of grief -the meaning of the loss experience, verified in their general experience- was less retrievable than the episodic memory of grief -the factual reminiscence associated with their loss, retrieved by their discourse about their experience- which might also suggest a splitting of experience, probably due to defensive mechanisms of denial and disavowal.

Second, participants showing more inhibited attachment representations showed less activated response of grief: this could be either because of an absence of grief by exclusion defense or because of one might present an omnipotent control by idealization of their experience and associating it to the absence of grief response or even have resolved the grief.

On the other hand, some participants expressed, precisely, that relief of distress because of the separation. One might contemplate that once the person has such experience, he no longer uses attachment strategies -whose aim is to maintain the bond-. Other than that, a break-up does not usually happen overnight and there may be a period of time when the person is torn between inhibition and hyperactivation, and, in the process, accept the imminent separation. That might explain why the participants’ attachment representations associates to the representations of the activation of the grief responses and did not in relation to the conscious perception of the experience of loss due to the separation.

This explanation goes hand in hand with the notion that the attachment is a factor on the resilience or vulnerability for an individual (Zukerfeld and Zonis Zukerfeld, 2016). This is also congruent with the variability of coping strategies verified in this study.

The relationship between the different coping strategies and the grief experience.

The final hypothesis was that participants presenting inhibited grief would use more denial and detachment and participants showing hyperactivated grief would use more emotional support seeking and self-blame. This hypothesis is confirmed. Nevertheless, one must be circumspect about this confirmation: When it is observed that participants showing resolved grief and those with inhibited grief had similar general experience of grief, this could mean that either the inhibited grief was a distorted perception through defense mechanisms or there was a desirability effect.

Also, from the correlations between the security and activation scores in relation to the attachment figures, in one hand was perceived that there was no direct association between attachment experiences with parents and the general grief experience of the break-up. On the other hand, a slight pattern of influence was observed regarding representations of security in relation to the last partner. It can be assumed that representations of security of attachment in relation to the last partner would have changed as a result of the break-up, but this question remains unanswered because no significant association between the general experience of grief and the activation of the attachment system in relation to this partner.

Regarding the relationship between specific strategies, it was perceived that some coping strategies tend to be found in the same individual such as self-distraction, self-blame, spiritual support seeking and planning strategy. This relationship shows that the participants who implemented these strategies tried to use all possible means to cope with and overcome the break-up situation. On the other hand, another relationship was well perceived, such as the relationship between drug use and the search for spiritual support, as a possibility to get away from the experience, but also to seek emotional support. It could also be a way to find a reinterpretation of the problem experienced. As much as drug use can mobilise the individual's entourage around these sufferings as a way of forgetting the break-up, spiritual support can, in the same way, mobilise others around this individual, as facilitating a state of victimization or self-blame. The other relationship encountered was the use of drugs and humour, which can be explained by the possibilities of distortion of the experience, as the effect of an incongruence between the experience of loss and manifestations of behaviour towards it.

Given that most participants inhibited their attachment system, the strong relationship between the detachment strategy and the acceptance of the break-up was observed, as expected, with the implementation of various strategies aimed at escaping responsibility for the break-up and asserting detachment.

In synthesis, regarding the relationship between coping strategies and the experience of grief, although participants who were classified as "resolved", "inhibited" and "hyperactive" did not differ in the use of humour strategies, positive rephrasing, self-distraction, venting, self-blame, drug use and detachment, there was a subtle difference in the use of the following strategies such as spiritual support-seeking, instrumental support-seeking and active coping. In accordance to theoretical expectations, those who were detected as presenting an inhibited grief seem to use fewer coping strategies and focus on denial and planning strategies -focusing on action to resolve the grief rather than on the emotion- compared to the other groups.

5. Conclusions

The first two hypotheses were partially verified: attachment representations in relation to secure attachment figures do not guarantee resolved grief. In addition, individuals with hyperactivated grief presented a stronger overall grief experience than others, whereas inhibited individuals described a grief experience identical to that of individuals with resolved grief. Already the third hypothesis stating that participants exhibiting inhibited grief would use more denial and detachment and participants exhibiting hyperactivated grief would use more emotional support seeking and self-blame was verified.

The associations highlighted between attachment representations and the break-up experience allows to verify that the reactions, at the time of the break-up, seemed to be little articulated to the attachment model experienced with the parents. Nevertheless, the attachment representations tended to repeat in later relationships when the attachment with parents was represented as insecure.

Discrepancies and lack of association might be related to defensive exclusion of certain representations as a way of avoiding being in contact with the sources of anxiety associated to relational stress. It might also be the biased result of social pressures that impel individuals to self-reliance and less introspection about those experiences.

Overall, this study verified that poor socio-emotional network increases the experience of vulnerability and the ability for effectively coping. However, it presents some limitation: it was built based on an small sample, did not differentiate among genders, was not able to access unconscious projections about their object relations, it was possibly impacted by the desirability effect, and could not follow up with questionnaires and interviews to verify consistency.

Nonetheless, this exploratory study sustain the idea proposed by Zukerfeld and Zonis Zukerfeld (2016) that attachment representations mediate the experience of vulnerability associated with the break-up experiences vis-à-vis patterns of behaviours put in place to cope with it and those of grieving. It illustrates how an empirical investigation can address the impact of the disruption phenomena through the triangulated analysis of narratives through semi-structured interviews about the disruptive experience and self-perception reports of behavioural tendencies. Overall, it illustrates how, within the Psychoanalytical field, an investigation can involve quantitative and qualitative methods and articulate the subjective experiences of individuals in their complexity intersubjective, intersubjective, and psychosocial.

It proposes studies involving projective materials, especially those focused on object relations, such as the Object Relations Test of Phillips and other projective test access representations at the figurability level, as a graphic test. It also suggests an study in a more diversified and amplified sample, comparing experiences of loss, for instance, those who experience separations from those who are at risk of losing their partners, to verify if the quality of the disruption phenomena is captured in the unconscious, preconscious and conscious experience, and whether can be differentiated.