Different Ways of Being Emotional about the Past

According to Dorothea Debus (2007), all emotional aspects related to an act of remembering are pres-ent and new emotional responses to the remembered past event. This is a common conception of the nature of the emotional aspect of personal memories, if not explicitly defended then at least implicitly accepted in the literature. In this article, I first criticize Debus’ arguments and demonstrate that she does not give us valid reasons to believe that all the emotional aspects related to a memory are present and new emotional responses to that past event. I then criticize Debus’ thesis tout court for being a direct consequence of assuming a particular conceptualization of the nature of emotions: emotions as physiological changes. Finally, based on a different conceptualization of emotions that focuses on their relational nature, I propose an alternative framework for analyzing the different possible emotional aspects of our personal memories. This leads me to conclude, contrary to Debus, that some emotional aspects of our memories are not occurrent emotions but are better conceived as a sort of quasi-emotions.

It is common to think about a past event of our lives and cry. It probably happens at least once to all of us. It is also common to remember old times with friends and beloved ones and have fun and laugh. It probably also happens to all of us. Our memories of our past personal experiences, which in the psychological literature are known under the name of episodic memories (Tulving, 1972) and autobiographical memories (Rubin, 1988), sometimes bring on old emotions and sometimes make us feel new emotions. But other times, we do not feel any emotion at all. We just remember a past experience without crying, without laughing. Nonetheless, this memory does not always appear to us devoid of all feeling, of all sensation; we do not merely remember that night out with friends where we dance our hearts out, but we also remember how we felt during that night. And the memory of this feeling, of the way in which this past event positively affected us, does not come devoid of all feeling. How should we interpret this feeling? If it is not a real emotion, how could it be understood?
In order to provide an answer to this question, I proceed as follows. The first part is primarily critical. First, I present Debus' (2007) argument against the possibility that memories that do not elicit an emotion present nonetheless an emotional aspect. I then show that Debus' argument is a circular argument vitiated with mistaken analogies that finally do not prove or give valid reasons for her claim that all the emotional aspects of our memories are occurrent and present emotions. I argue that the main problem behind Debus' circular reasoning is an assumption about the nature of emotion that leads her to an erroneous starting point for her philosophical inquiry. The second part of this article is constructive: based on a currently more acceptable conceptualisation of the nature of emotions, I propose a different framework for analysing the different possible emotional aspects of our memories. Within this framework, there is room to claim for the existence of what I call "quasi-emotions" related to memory, that is, of an emotional kind of experience related to a memory that has some similarities to occurrent emotions but it is not experienced by the rememberer as a present emotional response. This framework leads me to reject Debus' thesis, that all the emotional aspects of our memories are occurrent and present emotions, and provides a novel concept that better describes the richness and complexities of our mental life.

Debus' reasoning
In 1911 the Swiss psychologist Édouard Claparède asked if emotions can be the object of our memories, that is, whether they can be remembered in a representational form that would have some properties of an emotion, rather than being remembered in a propositional form. In this non-propositional form, memories would have a sort of emotion-like character without eliciting at the same time an occurrent emotion. His answer was negative, in agreement with philosopher William James (1890) and 3/14 psychologist Edward B. Titchener (1895), and contrary to most of his French contemporaries, such as philosopher and psychologist Théodule Ribot (1894), who defended the existence of a specific kind of memory for emotions called "affective memory" " (Trakas, 2015(Trakas, , 2021aAthéa and Trakas, forthcoming). In a much more recent article, philosopher Dorothea Debus (2007) also defended the idea that emotions cannot be remembered other than in a propositional form. For Debus, all the emotional aspects related to a memory of a past event which we witnessed at the time it occurred are present, new emotional responses to that past event, similar in nature to the emotional responses we have to present stimuli.
In order to defend this idea, Debus (2007) proceeds as follows. First, she briefly considers and rules out with good justification what she calls the "universal-memory-claim". The universal-memory-claim states that all emotions directed towards an event that the subject experienced in the past are memories. The counterexample is simple and obvious: we can experience an emotion different from the past emotion, like feeling shame at a past amusement. So in this case the emotion felt is not a memory but a new and occurrent emotional response to the past event.
The second claim that Debus considers is a more restricted version of the universal-memory-claim, which states that similarity between the present and the past emotion guarantees that the emotion is a memory. But again she rules out this thesis: she argues that similarity by itself is not a sufficient warrant. Nonetheless, the counterexample she gives this time is quite counterintuitive: she invites us to think of the case of a very empathic subject A who could feel the same emotion as subject B felt in the past when B tells A about this past event and emotion. The emotion felt by A is similar to the emotion felt by A in the past, but it is not usually considered to be a memory.
This last counterexample is clearly not a case of remembering, so it would not rule out the restricted version of the universal-memory-claim, at least in principle. But it could be saved by adding a requirement to guarantee the identity between the experiencer and the rememberer. And this is broadly what Debus does: she proposes to consider a new restricted version of the universal-memory-claim (a third claim), which adds to the similarity condition a second condition: the requirement of a causal bond between the past and the present emotion. According to this new restricted version of the universal-memory-claim, the similarity between the past and the present emotion and the existence of a causal link between the two (in order to assure in a certain manner the identity of the rememberer) would be both sufficient conditions to claim that a past directed emotion is a memory.
Nonetheless, Debus argues against the sufficiency of these conditions for determining that an emotional experience counts as a memory. This argument allows her to rule out this third claim as well as more broadly any version (restricted or not) of the universal-memory-claim. Her argument can be summarised as follows. First, Debus states that the sufficiency of these conditions, especially the causal condition, seems to be not so much a matter of philosophical debate but a hypothesis that could probably only be ascertained by empirical science. But even if it is desirable that this hypothesis were true, the sufficiency of these two conditions have not been proven by science, and moreover, "it remains unclear whether relevant empirical research would confirm it" (Debus, 2007, p. 764). So she decides next to take as a principle of her inquiry the following: in characterizing the phenomenon we want to explain (i.e. memory experience) we should not be guided by any prior preferences for a particular account of it but we should aim to capture the phenomenon itself as accurately as possible (Debus, 2007, p. 765). Guided by this principle of neutrality, Debus concludes that the analysis of the experiential and representational level reveals that these two conditions -causality and similarity-are not sufficient for an experience to count as a memory. She argues that both of them can be fulfilled and the subject might not know that he is remembering: he might think that he is imagining, and on the basis of our everyday intuitions we would not say that he is remembering. This idea refers to some classic examples mentioned in the literature of a subject who has a visual image of x (Malcolm, 1963) or paints an image of x (Martin and Deutscher, 1966) that has been caused by a similar past perceptual experience of x, but who recognizes neither the visual image nor the painting as memory experiences. What is more: even in the case in which the subject asserts that this experience is a memory, if he does not use the information provided by this experience to form beliefs about the past, he is not in fact remembering. For example, if a subject asserts that he has a clear memory image of a blue car moving towards him but then, when choosing the colour sample of the car from a set of colour patches, he picks a green colour patch, it is counterintuitive to state that he remembers a blue car moving towards him.
From the analysis, Debus concludes that, besides causality and similarity, there is another condition, generally omitted in the literature, that has to be met for an experience to be considered a memory: the epistemic-relevance-condition. This third condition refers to the fact that the subject has to make epistemological use of the present experience when making judgments about the relevant past experience. If the subject does not assume, without relying on any inferential reasoning, that the present experience presents him with how he experienced things in the past, the experience does not count as a memory experience (Debus, 2010).
Once Debus has established the epistemic-relevance-condition as the essential condition that needs to be met for a mental experience to count as a memory, she examines whether emotions directed towards a past event meet this condition or not. For this purpose, she considers the following case: "when thinking about the wedding now, I presently feel happy and a little worried, and that's how I know that I felt happy about their commitment and a little worried about their future together at the time" (Debus, 2007, p. 768). Although in this hypothetical case the subject would be making use of his present experience to know how he felt about the past, Debus concludes that we would not say that he is being reasonable. The subject cannot rely on his present emotional experience to determine how he felt in the past; he necessarily has to make use of additional information, for example, that he had a huge smile on his face and dropped a tear during the ceremony. In fact, it is the consideration of this additional information and inferential reasoning that allow him to know that he felt happy at the time of that ceremony, and not the consideration of his present emotions towards that past event.
Therefore, Debus concludes that a reasonable subject never uses emotions directed towards a past event to judge how he felt in the past, and thus these emotions never meet the epistemic-relevance-condition. That is why emotions directed towards a past event never count as memory experiences, but are always present and new emotional responses to the past event. The universal-new-emotion-claim is then formulated by Debus as follows: […] whenever a subject experiences an APD-emotion [autobiographically past-directed emotions], the subject presently experiences a new emotion which is directed at a relevant past event (or situation). Thus, all APD-emotions are present, new emotional responses to the past events (or situations) towards which the relevant APD-emotions are directed (Debus, 2007, p. 772).

What is wrong with Debus' reasoning?
In the first instance, one could think that the weak point of Debus' argument is the "epistemic-relevance condition". If it can be true that some past-directed emotional states are not useful in knowing our past emotions, it seems that some others would allow us to know how we felt in the past without relying on any inferential reasoning. The most obvious cases are traumatic memories: the rememberer not only relives the traumatic event which seems to be happening in the present, but also the emotional state associated with it. In fact, according to some psychological theories (Boals and Rubin, 2011), traumatic memories are characterized by the inseparableness between the past event and the emotion associated with it in the representation of the past event. For example, if I remember the time when a burglar who broke into my house played Russian roulette with me, my heart pounds faster than usual, my stomach shuts down, I feel numbness in my fingers and feet, I start to sweat, my vision becomes blurry and I feel heavy as if I may faint. This occurrent emotional state of fear and anguish allows me and others to know that I experienced the same emotion (or at least an emotion of the same category) when the event occurred: fear. But traumatic memories are not the only case: (a) emotions related with memories of past stressful events that are relived because they are still emotionally open and have not reached closure for the rememberer, and (b) emotional states related with deliberate acts of remembering that aim to empathetically access the past, could also be considered as cases of occurrent past-directed emotional states that give us knowledge about our past emotional life. I will come back to these kinds of examples in the second part of this article.
But more broadly, we could argue against the "epistemic-relevance condition" by claiming that a memory state is not defined by its tendency to produce beliefs about the past but only by its causal connection. Debus explicitly disagrees with this position and that is why she introduces the "epistemic-relevance condition" (Debus, 2010). Nonetheless, other philosophers such as Martin and Deutscher (1966) and Bernecker (2010) have defended this view. For them, what determines the status of a memory experience is its objective conditions, that is, its causal link with a past representation, and not its subjective conditions. So if I paint what I think is an imaginary landscape but in fact I did not create it out of my imagination: it was a landscape I saw during my childhood, I am actually remembering and not imagining it despite believing that I am imagining the landscape. Another argument against the epistemic-relevance-condition comes from recent research on non-believed memories, which has shown that memory and beliefs about the past are better understood as two independent variables, because some memories do not produce in us a belief about the veridicality of their content (Mazzoni et al., 2010;Trakas, 2021b).
Nonetheless, the most important problem with Debus' general argument is its wrong starting point. This wrong starting point leads her into a circular argument and to use inappropriate examples and analogies to defend the idea that emotions directed towards a past event are always present and new emotional responses to that past event. Consider the examples and analogies that Debus uses to defend her view: when I remember my friend's wedding, my present feeling of happiness and worry allow me to know neither that I felt the same way nor the way I felt during the wedding. Therefore, because I do not make use of my present emotional experience to know how I felt in the past, my emotional response to the memory of the event is not a memory itself but a present and occurrent emotion. What is the logical structure that underlies this example? It can be formulated as follows: When thinking about a past event e, I feel x, and that is why I know that I felt x during the past event e. It is true that this statement goes against our everyday intuition, as Debus appropriately points out, but it is also true that the equivalent version for perceptual information is also counterintuitive. When thinking about a past event e, it is not because I see x now that I know that I saw x during the past event e; it is because I remember seeing x that I know I saw x during the past event e. When remembering my friend's wedding, it is not because I currently see the red dress the bride wore that I know that I saw the red dress she wore in the past; it is because I remember seeing that red dress that I now know I saw it during the wedding. And the same could be said about emotions: When thinking about a past event e, I remember feeling x, and that is why I know that I felt x in the past. This kind of reasoning would not be counterintuitive.
The upshot is that we can only retain her main claim and must omit her examples, and so establish that for some experience to be a memory the subject has to properly recognize it as a memory, which means that he has to make use of that present experience to judge about how he felt in the past. And because this does not happen with occurrent emotions, emotions cannot be remembered. But the problem is that this does not happen with occurrent perceptions either; it only happens with past perceptions that are experienced as visual memories. And it is precisely the existence of some analogue to visual memories, a sort of memory-like emotion or quasi-emotion that is experienced as such, that Debus should have considered in her reasoning in order to determine if something of this nature exists. But she did not. Her erroneous starting point led her then to construct a circular argument: if we want to analyse whether some experience that seems in a certain way emotional is a memory experience or not and we take as the starting point that this experience is an occurrent emotion, there is clearly no valid argument that can conclude something that is at odds with the first assumption.
Let's revise again the circularity of her argumentation. First, she herself recognizes at the beginning of the argument that autobiographically past-directed emotions "as discussed here are occurrent emotions, that is, experiences of certain emotions at particular moments of time" (Debus, 2007, p. 759). This does not mean something different from the conclusion at which she arrives: "all autobiographical past-directed emotions are present, new emotional responses to the past events" (Debus, 2007, p. 772). I do not see how an occurrent emotion could not be a present emotional response. Second, her arguments are based on analogies between cases that are dissimilar. These incorrect analogies originate in the same circularity that ensues from the wrong choice of the starting point for inquiry. Once past-directed emotions are not considered as occurrent emotions but as memory-like emotions that need to be further analysed, all dissimilarity with perceptual memories vanishes.
Therefore, the main problem with Debus' argument is an erroneous starting point, which does not permit the interrogation of the possible nature of phenomenal experiences that present themselves with some memory character and with some emotional character. Debus' starting point for inquiry eliminates thus from the beginning the possibility of questioning or denying her initial claim. A good way of carrying out this research would be to explore the nature of these different aspects, and not to start with some assumptions about its constitution. As I already mentioned, Debus' herself proposes that we should follow this methodology: "(…) we should aim to capture the phenomenon itself as accurately as possible. In characterizing the phenomenon which we want to account for, we should not be guided by any prior preferences for a particular account of the phenomenon" (Debus, 2007, p. 764). But it seems that nonetheless she let herself be guided by prior preferences about the nature of emotions and the nature of personal emotional memories, as I explain next.
The heart of Debus' circularity problem: emotions as physiological changes At the beginning of the article I mentioned that Debus' thesis was similar to that one defended by Claparède (1911), Titchener (1895) and James (1890): emotions cannot be remembered without losing their emotional character, so all emotional aspects related to past memories are occurrent present responses. As most of the readers probably know, James also defended a particular conception about emotions: Bodily changes follow directly the perception of the exciting fact, and our feeling of the same changes as they occur is the emotion (…) we feel sorry because we cry, angry because we strike, afraid because we tremble, and not that we cry, strike, or tremble, because we are sorry, angry or fearful. Without the bodily states following the perception, the latter would be purely cognitive in form, pale, colourless, destitute of emotional warmth (James, 1884, p. 189).
The Jamesian conceptualization of emotions as perceptions of bodily changes makes perfect sense and coheres with the idea that emotions cannot be remembered: because bodily changes and states are always experienced in the present as occurrent states, and because emotions are reduced to these bodily changes and states or to the experience of them, emotions cannot have a memory counterpart. As soon as my heart races, my muscles tense, my palms sweat, I am undergoing an emotional state. And in fact, it is exactly this notion of emotion that is assumed by Debus (2007) throughout her whole reasoning. This hidden assumption is the main cause of the circularity of her general argument: once the idea that emotions are physiological arousals is assumed, the conclusion that all emotions directed towards past events are occurrent and present emotions becomes a self-evident proposition.

7/14
Therefore, if the conceptualisation of emotions as physiological changes entails the conception that all emotional aspects related to memories are occurrent and present emotions directed towards the past, another conceptualisation of emotions would certainly entail a different conclusion. In the next section, I propose to explore the consequences of conceptualizing emotions as being essentially relational, that is, as appraisals.
Emotions as appraisals and the possibility of quasi-emotions Instead of conceiving emotions as simple bodily changes and feelings of those changes, I propose to adopt another conceptualization of the nature of emotions which is more compatible with current research on emotion in cognitive science. In spite of the specificities of different theories of emotions, all of them seem to agree that emotions are complex and dynamic processes that unfold over time and that involve different components, so they cannot be simply reduced to physiological changes and their consequent feelings (Russel and Barrett, 1999;Goldie, 2000;Lambie and Marcel, 2002;Barrett et al., 2007;Griffiths, 2013;Mulligan and Scherer, 2013). All these current theories implicitly or explicitly agree that the essence of affections and emotions consists in being relational, that is, that they are not only about feelings of internal changes of the body (as in pain), but they are about person-environment relationships that involve how one's concerns or one's self has been affected by some stimuli in terms of harms (for the negative emotions) and benefits (for the positive emotions) or in terms of morality and self-image. And this is precisely the idea that lies behind the concept of "appraisal" (Arnold, 1960;Lazarus and Smith,1988;Lazarus, 2001), despite the fact that in some literature this concept is often taken as signifying that emotions are merely judgments. Appraisals need not to be conceptual and disembodied as it is in general thought; they can also be conceived as automatic and unconscious " (Arnold, 1960;Lambie and Marcel, 2002), embedded in the experience (Arnold, 1960;Martin, 1992;Tye, 2008) and embodied in physiological changes (Prinz, 2004;Colombetti, 2013). As Moors et al. (2013) explain, the notion of appraisal and the notion of emotion are conceptually related, and although there is disagreement about the role that appraisals play in an emotion, appraisals are considered to be the driving force that lies behind an emotional episode. In fact, it is difficult to find a theory that denies that appraisals are in certain way essential to emotions (Mulligan and Scherer, 2013).
The idea that emotions are appraisals, that is, that they are about person-environment relationships that involve how one's concerns or one's self has been affected by some stimuli in terms of harms, benefits, morality or self-image, implies that emotions necessarily involve a self affected by an event in the world. So in order to understand the nature of an emotional experience, it is necessary to specify (a) the nature of the stimulus or event that affects one's concerns or one's self and (b) the nature of the self who is affected by the stimulus. The stimulus or event in the world that affects the self can be a present and occurrent event that is happening now, an imaginary or fictional event, or a past event that is being remembered. In the case of a present event, it is the present self who is affected and thus experiences an emotion. The case of the imaginary or fictional event is more complex and goes beyond the scope of this article, so I will not deal with it here. Let's go directly to the case under current consideration: the case where the self is emotionally affected by a remembered past event. Because of the intrinsic dichotomy present in personal memory between a self who remembers what a former self experienced, the self emotionally affected by a past event can be the present self who remembers or the past self who experienced what the present self is remembering. It seems thus that in a personal memory with some emotional aspect that-needs-to-be-defined, the self emotionally affected by the past event is in principle ambiguous: it can refer to the present self who remembers, or it can refer to the past self who directly experienced what is being remembered.
So taking into consideration the differences mentioned before, and in order to see if the conceptualisation of emotion as appraisal leads to a different picture of the nature of emotional mem-ories from the one introduced by Debus, I propose to consider the following possible three cases of personal memories which present some emotional aspect (for more details and other cases, see Trakas 2021c): a. The past event is still of concern for the present self in terms of harms and benefits, morality or self-image, and the appraisal made by the present self is the same as the one made by the past self.
b. The past event is still of concern for the present self in terms of harms and benefits, morality or self-image, but the appraisal made by the present self differs from the one made by the past self.
c. The past event is no longer of concern for the present self in terms of harms and benefits, morality or self-image. The present self is just a spectator of the way his past self was emotionally affected by the past event.
a. The past event is still of concern for the present self in terms of harms and benefits, morality or self-image and the appraisal made by the present self is the same as the one made by the past self.
In this case, there is an identification of the present self with the past self, and even an illusion of a single self present in both times. In narrative terms, the point of view of the present self corresponds to the first-person perspective, that is, the perspective of the actor, or of the audience that identifies with the characters to the point of experiencing their emotions. Borrowing Ribot's (1907) ideas, this identification can be weak, but it also can take a lively form and become hallucinatory: the past self is resurrected, relived for a short time in the present self and the past emotion is reexperienced, reenacted. These extreme cases are perfectly adequate descriptions of traumatic memories related to post-traumatic stress disorder (PTSD). PTSD is characterized by a strong sense of reliving that is reflected in a distortion of the sense of time: the traumatic event seems to be happening in the present rather than belonging to the past, as in the case of ordinary memories (Brewin and Holmes, 2003). In fact, individuals suffering from PTSD not only report that their flashback images seem real but also sometimes respond as if the traumatic event is happening again, by exhibiting signs of terror, autonomic symptoms, like sweating, and even invoking some behavior, such as ducking as if to avoid a blow (Holmes and Mathews, 2010). Whereas some theories consider that the cause of reliving in traumatic memories associated with PTSD is due to their low cognitive-emotional distinctiveness, that is, to the small degree to which the emotions associated with an event are separated in the representation of that event (Boals and Rubin, 2011), other theories, such as dual representation theory, emphasize the role of imagery in personal memory, and propose that the imagery that characterizes the PTSD may be responsible for the re-experience. This is not only because of its direct links with the amygdala but also because it may still convey a sense of immediate perceptual experience that makes the individual momentarily process them as real and as a real threat (Brewin and Holmes, 2003;Holmes and Matthews, 2010).
Nonetheless, traumatic memories associated with PTSD are not the only kind of memories that are still emotionally open to the rememberer and "not still behind him" (Beike et al., 2004). There are more common memories that are stressful for the rememberer that are also relived, even if with less intensity than PTSD. For illustrative purposes, I propose to consider the following example: the experience of seeing my partner entering a hotel with his colleague in a suspicious way, which produced in me a mixture of anger, sadness and humiliation. Each day after that episode, the memory of my husband going into the hotel with his colleague intrudes into my mind. So even if there is no actual external stimulus in my environment that produces in me anger and sadness, because I identify with my past self who witnessed the cheating scene, I re-experience the sadness, anger and humiliation in the same way as if the past event was occurrent and not past. This does not mean that my occurrent emotion is an exact replica of the emotion felt in the past; it just means that it belongs to the same 'category' of emotion and that the appraisal has not changed. In this sense, the emotional aspect of my memory is an occurrent emotion but it is not a new emotional response to the past event.

9/14
b. The past event is still of concern for the present self in terms of harms and benefits, morality or self-image, but the appraisal made by the present self is different from the one made by the past self.
In this second case, unlike the first one, there is no identification with a past self but a feeling of distance, of not-me (Libby and Eibach, 2002). There is no reexperiencing of the past emotion, but just a new and different emotion. A new appraisal is certainly the product of new knowledge, evaluations and feelings concerning the past event, what Goldie (2012) called the triple ironic gap. Coming back to my last example, let's suppose that after the "cheating scene" I discovered that in fact my husband and his colleague were going to the hotel for a meeting with some businessmen, and that the "suspicious" attitude I thought I saw was just a product of my mind, which had recently become caught up in jealousy and feelings of insecurity about my husband's love towards me. After this discovery, I no longer remember my husband entering a hotel with his colleague as the "cheating" scene, but as a simple meeting with businessmen; and I no longer feel anger and sadness but relief. Or imagine the following situation: after feeling anger and sadness when discovering that my husband cheated on me, I realized that in fact our marriage had been already broken a long time ago, that he did not love me anymore and I did not love him either, and that in fact seeing him entering a hotel with another woman was just a way of witnessing with my own eyes something that was already existent and was implicitly evident for both of us. Instead of reexperiencing anger and sadness, I feel frustrated and disappointed. I remember thus the scene of my husband entering a hotel with his colleague as the moment of revelation of the failure of our marriage.
These examples are still cases of emotionally open memories because the past experience is still of concern for the present self, even if in a different way. But the perspective adopted differs from the first cases: the subject is an observer and has adopted an external perspective, different from the original one. He is certainly not a detached observer, because he is still to some extent an actor: even if different from the past actor, he is still reinterpreting and resignifying what happened. That is why this kind of affective perspective cannot be considered as a third person perspective in the strict sense of the term. In fact, because of the emotional openness of these kinds of memories, the subject may be at the same time actor and observer of the past event, and thus, simultaneously experience the past emotion and the new emotion. Coming back to my example, after realizing that our marriage was broken, I can remember the cheating scene and feel frustrated and disappointed at the same time that I reexperience some of the anger and sadness felt in the past. The social psychologist Thomas J. Scheff (1981) has called this way of distancing aesthetic distance: "re-living the past is underdistanced, one is entirely a participant. Remembering the past is overdistanced, one is entirely an observer. Aesthetic distance corresponds to returning to the past; one is both participant and observer, simultaneously" (Scheff, 1981, p. 47). The possibility of adopting two simultaneous perspectives better accounts for the complexity of real life events which often carry several meanings and give rise at the same time to several appraisals, sometimes even to opposite appraisals (Fridja, 2013).
c. The past event is no longer of concern for the present self in terms of harms and benefits, morality or self-image. The present self is just a spectator of the way his past self was emotionally affected by the past event.
A third possibility appears when the past event is no longer of concern for the present self in terms of harms, benefits, morality or self-image, but there is nonetheless an emotional aspect in the memory. This is exactly the kind of phenomenon that Debus should have considered from the beginning of her inquiry, but she did not. In fact, this is precisely the kind of phenomenon that authorizes us to conclude that not all emotional experiences related to personal memories are necessarily occurrent and present emotions.
Let's first consider the characteristics of these kinds of memories. In these cases, the present self adopts a real observer and external perspective: he is just a spectator of his past experience, of the way his past self was emotionally affected by the past event, but certainly not an actor, because the present self does not appraise the past event anymore. In fact, highly emotional memories are rare; emotionally closed memories seem to be more the rule than the exception (Beike et al., 2004). Whereas in some cases the absence of a current emotional experience while remembering a recent emotional event can be the result of a pathological overdistanced perspective, most personal memories without an occurrent emotion refer to cases where the subject has already achieved an evaluative and emotional response about the past emotional event that he considers appropriate, which means that he has arrived at an optimal distance. In these cases, the emotion has faded and the past event is thus emotionally closed. Coming back to my example, when I come to a personal understanding about why our marriage was not working long before the cheating scene, the memory of the cheating scene would probably become an emotionally closed memory. Time may pass by, I may reconstruct my life and even have new relationships; so every time I remember this event, I will remember it as the moment of revelation of my failed marriage, without necessarily feeling frustrated or disappointed.
However, the fact that I do not feel an occurrent emotion does not mean that this memory necessarily lacks an emotional aspect. First, the past appraisal can be so intertwined with the representation of the event that it actually constitutes the intentional object of the memory of the event. In my example, when I visualize my ex-husband walking through the doors of the hotel, I do not simply remember that my husband entered a particular hotel at a specific time; I remember that my husband cheated on me and caused me harm. Second, this does not mean either that the emotional content of an emotionally closed memory can only be represented through language or visual imagery: action tendencies and interoceptive body sensations can also be part of the memory experience without their presence implying that the subject is undergoing an occurrent emotional experience. Why does the presence of bodily changes and feelings not imply that the subject is undergoing an occurrent emotional state? Because according to the conception of emotions assumed at the beginning of this analysis, an emotional state is not determined by the presence of physiological changes, as Debus (2007) implicitly assumes, but by the fact that some stimuli is of concern for the present self in terms of harms, benefits, morality or self-image. And in the example given before, it is clear that the event remembered is not of concern anymore for my present self, despite the fact that my body may suffer some physiological changes that I may even feel.
Apart from some philosophers who have defended the existence of an affective memory, a particular kind of memory of emotional events (see Trakas, 2015Trakas, , 2021aAthéa and Trakas, forthcoming), only a few cognitive scientists have explicitly acknowledged this emotional component that is present in some memories about the personal past but that does not correspond to an occurrent emotional response. Joseph LeDoux (1998LeDoux ( , 2008 has coined the term "emotional colouration" to refer to memories that do not lead to an emotional arousal but that are not emotionally flat. A similar terminology can be found in some experiments conducted by Hans Markowitsch and his team (Markowitsch et al. 2000(Markowitsch et al. , 2003Reinhold and Markowitsch, 2009). On the other hand, philosopher Richard Wollheim (1984) introduced the notion of "cogency" or "residual condition" to name the tendency to feel or reexperience the emotions and affections felt in the past while remembering a past emotional event. The emotional residue of memories would be the consequence of a sort of imprint left in the memory of the event due to the high emotional intensity of the past event and the subsequent rumination about what happened.
In fact, the notion of quasi-emotion may be appropriate to name this emotional aspect of memories that does not constitute an occurrent and present emotion for the rememberer. The notion of quasi-emotion has been introduced in the philosophical literature by philosopher Kendall Walton (1990) to refer to the emotional responses to works of fiction such as films, books and paintings, which are not considered to be genuine emotional responses. This notion can be useful to capture the particularities of the emotional aspect of some memories that presents some properties of an emotion without being at the same time an occurrent emotion. On the one hand, this emotional aspect shares with emotions some essential properties: (a) a stimuli that is of concern for a self in terms of harms, benefits, morality or self-image; (b) an appraisal made by a self; (c) physiological changes, interoceptive sensations, action tendencies. On the other hand, it cannot be considered as an occurrent emotion because one of these essential properties, i.e. appraisal made by the present self, is absent. In this reinterpretation of the notion of quasi-emotion, the relation between quasi-emotion and emotion correlates with the already established relation between visual memory and visual perception, and thus introduces a novel parallel and a novel concept for describing our inner mental world that does better justice to the richness of our emotional experiences.
One final remark: as Wollheim's (1984) term "residual condition" suggests, there is a tendency in some personal memories, even if they are emotionally closed, to make the rememberer feel and reexperience the emotions and affections felt in the past that are part of the memory. It is just a tendency, a residual condition, for which the best metaphor would probably be a comet's tail, and not an occurrent emotion. Nonetheless, this tendency or quasi-emotion can turn into an occurrent emotion experienced by the present self, that is, by the rememberer. In my example, this would happen if after time passed by and I reconstructed my life and even had other partners, while remembering the cheating scene I suddenly feel frustrated; or even more, I suddenly feel anger and hate towards my ex-husband. This could be possibly explained in two ways: (i) it is just a case of emotional contagion (Hatflied et al., 1993), and so the distinction between the past and the present selves is still kept, as well as the observer perspective; or (ii) this shows that probably because of particular circumstances and the mood of the present self (for example, a big fight with my current partner, or a new failed marriage, or a very pleasant encounter with my ex), the past event became emotionally opened again in a way that is actually appraised again by the present self, giving rise to a present emotional response. This last example, which questions the idea that the emotional closure of memories is a definite property that they can acquire, also accounts for the transitions and dynamisms between different perspectives as well as for the complex relations between our past and present emotional lives.

Different ways of being emotional about the past
As my present analysis shows, the adoption of a different starting point for inquiry, that is, of a different conception of the nature of emotions, led indeed to a different analysis of the nature of the emotional aspects of personal memory and thus to a different thesis from that one proposed by Debus (2007).
I agree with Debus and consequently reject (a) the universal-memory-claim, which states that all the emotional aspects of memories are memories; and (b) the restricted-memory-claim, which states that all emotional aspects of memories that are similar to the corresponding emotion felt in the past are memories. Nonetheless, I also reject (c) the universal-new-emotion-claim, the thesis defended by Debus, which states that all the emotional aspects of memories are present and new emotional responses to the past events remembered. Besides the problems intrinsic to Debus' argument that make her main claim unconvincing, in the second section I have explained in detail the two arguments that lead me to reject Debus thesis tout court. First, not all emotional aspects of memories that are present emotional responses are necessarily new emotional responses to a past event: cases of reliving and reexperiencing past emotions such as in PTSD need to be conceptually distinguished from cases where the rememberer appraises the past event in a different way and thus distances himself from his past self. Second, and more importantly, there are some emotional aspects of memories that are not occurrent and present emotional responses to a past event but that nevertheless share with occurrent and present emotional responses some essential properties. I proposed the term "quasi-memory" to refer to these particular emotional aspects of memory experiences that have the mark of the past.
It is thus possible to formulate the main thesis of this analysis in the following terms: Considering (a) the duality of every memory experience between a present self who remembers what a past self expe-