Enquiring into the Qualitative Nature of Anger

I was angry with my friend:

I told my wrath, my wrath did end.

I was angry with my foe:

I told it not, my wrath did grow…

—William Blake, A Poison Tree

Anger, “a strong feeling of displeasure and usually of antagonism” (Anger, n.d.) is typically experienced as an unwelcome and intrusive visitor that is difficult to channel or control. And yet, people have to come to terms with it – which is often an ambivalent and more or less successful endeavour, as is evident from the beginning of William Blake’s poem A Poison Tree (Blake 2008). In the current article we sought to enquire into the first-person phenomenology of anger in order to provide a thorough and systematic understanding of this emotion that is universally present in human life, but typically studied from the ‘outside’ view of third-person research.

People typically experience anger when their effort to reach a goal is obstructed by another party or context (Scherer 1997) or when physical or psychological restraint is experienced (Harmon-Jones and Harmon-Jones 2016). Among the theoretical explanations that have been proposed to explain the occurrence of anger, two are particularly prominent: Berkowitz‘s Cognitive-Neoassociationistic Model (Berkowitz 1990, 2012) and the Appraisal Approach of Anger Generation (Shiota and Kalat 2012). The Neoassociationistic Model is an extension of the frustration-aggression hypothesis and accounts for anger on the basis of learning experiences (Harmon-Jones and Harmon-Jones 2016). It assumes that anger is experienced when negative affect is associated with prior negative episodes (cf. fight-and-flight; Berkowitz 1989, 2012). In this model, external attribution of blameworthiness or even hostile intent is not a necessary feature of becoming angry (Berkowitz 2012). By contrast, appraisal models consider external attribution of blameworthiness as a necessary pre-requisite of becoming angry (Shiota and Kalat 2012). That is, when a situation is appraised as negative, several emotions can ensue. For an emotional response to be anger-related, the appraisal must be such that others can be blamed for an unjustified goal-blockage (Lazarus 1991). Some theorists incorporate a threatened self-esteem and a reduced capacity for coping as necessary components into such appraisal theories. In addition, there is discussion about the causality of anger-experience and appraisal, as it remains unclear which one causes the other (Harmon-Jones and Harmon-Jones 2016). Below follows a brief review of typical psychological and physiological facets of anger, triggers for anger events, and interindividual as well as contextual factors of anger.

Psychological and Physiological Facets of Anger

What are the psychological facets of anger? Generally, research has demonstrated that anger has a negative valence. Next to the experience itself, the trigger of the anger-experience (the anger-eliciting stimulus or event) is perceived negatively. At the same time, significant interindividual differences exist in the degree with which such negativity is experienced (Harmon-Jones and Harmon-Jones 2016). Anger frequently co-occurs with other adverse emotions and is associated with reduced happiness. In addition, it is also related to the subscale of positive activation in the PANAS which measures positivity and approach motivation (Harmon-Jones and Harmon-Jones 2010). As such, anger, in contrast to many other negative emotions, is an approach-related affect (Carver and Harmon-Jones 2009; Harmon-Jones and Harmon-Jones 2010, 2016; Plant and Devine 2003). Further support for the view that anger is approach related comes from research demonstrating a connection between trait anger (explained in more detail below) and other approach related traits such as bravery, self-assurance, and physical strength. Neuropsychological evidence links anger to left frontal cortical activity which is associated with approach tendencies. More specifically, the intensity of approach motivation while being angry determines left frontal activity (Harmon-Jones and Harmon-Jones 2016). Two competing theoretical explanations for the cause of the approach-related nature of anger have been that 1) anger occurs because an approached goal is blocked and this is experienced with anger and 2) that anger itself motivates approach behaviours, for example to reach goals (Harmon-Jones and Harmon-Jones 2016). Note, however, that anger can also be avoidance-related, especially when anxiety co-occurs with anger, for example in situations when anger expression would be punished. Accordingly, individuals who have high anger control show more right frontal activity which is associated with withdrawal tendencies (Hewig et al. 2004; Zinner et al. 2008).

On a practical level, the notion that anger is characterized by an approach motivation is plausible, since it serves adaptive functions. For instance, anger “organizes and regulates psychological processes, such as self-defense and mastery. It regulates social and interpersonal processes, and it organizes processes to assist with goal-directed action.” (Harmon-Jones and Harmon-Jones 2016, p. 774). Recalling that anger occurs as a response to goal-blockage, anger can be interpreted as a means to reach such goals (here Harmon-Jones and Harmon-Jones 2016 make an explicit link to reactance theory). On the downside, anger is also typically associated with dysfunctional patterns. Furthermore, in the context of social relations provoked aggression can have serious consequences such as violence and child maltreatment (Holtzworth-Munroe and Clements 2007). Moreover, the person experiencing anger can suffer from detrimental effects, such as coronary heart disease (Niaura et al. 2002). Our overview so far hints at a distinction between the experience and the expression of anger. Both dimensions are factor-analytically distinct from each other. Accordingly, anger can occur without aggression or violence (Fernandez and Johnson 2016).

Furthermore, anger is considered a moral emotion arising when one’s sense of autonomy is violated. Accordingly, taking revenge or seeking justice are two prevalent action tendencies when angry (Russell and Giner-Sorolla 2013). Fernandez (2008) characterizes anger on a range of dimensions, including the level of internalization vs. externalization, resistance vs. retaliation, and impulsivity vs. control of anger expressions. He also distinguishes different modes of expression and different functions of anger, that is verbal and physical expressions and restorative vs. punitive anger, respectively.

The affective state of anger can differ in intensity ranging from irritation to annoyance to fury. Furthermore, items from anger scales using such words load on a single factor. Though there is debate whether anger and rage are two distinct experiential states, the prototype view of emotion conceives the prototype of anger as “a variety of feelings such as irritation, annoyance, exasperation, disgust, and hate. Thus, various ‘nuances’ of anger experience may not be distinctly different affective states.” (Harmon-Jones and Harmon-Jones 2016, p. 777).

On a physiological level, anger leads to an increase in cardiovascular and hypothalamic activity, in breathing rate, and cortisol release (Shiota and Kalat 2012; Suls and Wan 1993). Thus, anger is connected to a certain bodily state of arousal (Russell and Giner-Sorolla 2013). On a perceptual level, the encoding and decoding of anger emotions is largely considered to be culturally independent (Ekman et al. 1987). Anger can be easily – and cross-culturally – identified from its facial expression: the characteristic frown, staring eyes, drawn back open lips, and a red face, though the intensity of the behavioural expression differs depending on how much it is controlled (Harmon-Jones and Harmon-Jones 2016; Russell and Giner-Sorolla 2013). A meta-analysis revealed activity to be mostly related to the left inferior frontal gyrus (Vytal and Hamann 2010). Nevertheless, based solely on physiological or neurological activity a differentiation of anger from other emotions is hardly possible. Thus a constructionistic approach to anger seems appropriate (Fernandez and Johnson 2016).

Many facets of subjective anger experiences and outwardly observable anger behaviour are tapped into quantitative anger scales, as was reviewed by Fernandez (2008): These include resentment, suspicion, assault, indirect and verbal hostility, irritability, negativism; control over aggressive impulses; intropunitiveness or ‘anger-in’ (delusional guilt, self-criticism, condemnation of anger) and extrapunitiveness or ‘anger-out’ (delusional hostility, criticism of others, hostile acting out); awareness of anger and expression of anger (physical, verbal, general); mistrust; anger regulation; frequency, intensity, and duration of anger; and hostile outlook (Buss and Durkee 1957; Caine et al. 1967; Cook and Medley 1954; Fernandez 2008; Megargee et al. 1967; Siegel 1986; Zelin et al. 1972). As with other personality variables, a distinction is made between state-anger and trait-anger. For instance, according to the State-Trait Anger Expression Inventory (Fernandez 2008; Spielberger 1988), state anger arises due to frustration and perceived maltreatment and refers to the “acute emotional-physiological reaction” (Veenstra et al. 2018, p. 98) described in the preceding paragraphs. Trait anger, by contrast, is a variable describing interindividual differences in how often, how intense, and how long a person experiences state anger and is therefore used to categorize anger-prone or hostile persons. Consistently, trait anger is useful in predicting state anger and aggression (Fernandez 2008; Veenstra et al. 2018). The distinction of state- vs. trait-anger corresponds to the conceptualization of anger as an emotion (short), mood (prolonged) or temperament (proneness to experience anger), as is evident in verbal expressions like “anger” and “rage” vs. “irritation” and “hostility”. Still, the demarcating line between them is not clear cut and both concepts can be conceived more of as a single continuum (Fernandez and Johnson 2016; Harmon-Jones and Harmon-Jones 2016).

Triggers of Anger

Anger can be elicited by many different causes. In a recent study Kashdan et al. (2016) identified five categories of triggers: other people, psychological and physical distress, intrapersonal demands, environmental aspects, and diffuse or undifferentiated triggers (relevance in that order). In this study, anger-related feelings and behaviours were stronger when the trigger was another person in comparison to the other triggers. Diffuse sources of anger produced the most intense anger feelings and anger reactions which were most difficult to control.

Recalling the appraisal approach above, a multifunctional view of emotion seems most appropriate in which different cognitive appraisals can elicit anger in order to fulfill an (evolutionary) purpose specific to a certain situation as such appraisals vary from physical discomfort to the feeling of unfairness and from goal blockage to the assumption of hostile intent. Such appraisals are accompanied by an interpretation of whether the feelings of anger are reasonable or not. As the appraisals that elicit anger differ, so the responses to these appraisals do, too. Hence, albeit different episodes of anger can contain very similar subjective feelings, behavioural and verbal expressions, as well as similar states of physiological arousal, the resulting action tendencies can be different (e.g., attack or reconciliation; Russell and Giner-Sorolla 2013).

Interindividual as well as Contextual Factors

In contrast to the above mentioned psychological and physiological facets, which specify more general dimensions of an anger experience (that is, how it is experienced), in the following we list reasons for why some people experience anger, while others do not; and why people experience and express it differently. Interindividual differences in experiencing and expressing anger have been traced to differences in genetic endowment (Rhee and Waldman 2002), gender differences (Bettencourt and Miller 1996), age and education (Schieman 2003), anger-prone temperament and hostile attribution style (Hazebroek et al. 2001), family environment (for example through model learning, Cadoret et al. 1996), left versus right hemispheric activation (Carver and Harmon-Jones 2009), levels of testosterone and serotonin (K. Anderson and Silver 1998; Carver and Harmon-Jones 2009), inhibition linked to activity in the prefrontal cortex (Shiota and Kalat 2012), and cognitive appraisal (Mauss et al. 2007; Memedovic et al. 2010). Contextual factors include heat, limited space, foul odors, or aggressive cues, among many other aversive stimuli (C. A. Anderson and Anderson 1998; Berkowitz and LePage 1967; Rotton et al. 1978).

People occasionally experience self-anger (a construct related to measures of anger-in, as mentioned above), which is generally accompanied by feelings of shame or guilt as self-anger normally does not imply aggression towards oneself (Ellsworth and Tong 2006). To clarify, anger-in vs. anger-out denotes the degree to which anger is openly expressed (Harmon-Jones and Harmon-Jones 2016). Finally, we would like to mention that anger has societal implications and is relevant in applied fields such as clinical psychology, where anger is prevalent among many psychopathological disorders and a key criterion for several types of diagnoses (Fernandez and Johnson 2016).

The above background provides an overview of the characteristics of anger to the extent that they are accessible to an outside observer’s (i.e., experimenter’s) point of view. Yet, we sense that the phenomenology of anger is even more complex and multi-faceted, and that important details about the qualitative nature of anger still needs to be addressed. Such further insight is not only of theoretical relevance – it is also of practical use, for instance in the applied clinical context. A considerable amount of research on interventions that would help people cope with anger-induced aggressive behaviour focuses on outwardly observable facets of anger measured with trait questionnaires (Del Vecchio and O'Leary 2004; Kashdan et al. 2016; Saini 2009). When developing interventions to cope with maladaptive types of anger one has to be mindful of the fact that different dimensions and phases of anger may require different coping strategies (Kashdan et al. 2016). Collecting first-person data on the qualitative phenomenology of anger could be a method for differentiating various anger dimensions in a more fine-grained way and to better tailor interventions to the individual needs. However, a systematic compilation of phenomenologically grounded first-person accounts of anger experiences does not yet exist. A first open question thus is: What are the different facets of the qualitative experience of anger?

Note that we are not calling for first-person methods to substitute third-person methods in the study of anger. Instead, we propose that first-person methods complement third-person research and that findings from both types of empirical investigation should be integrated in order to create a more complete and accurate picture of anger. To illustrate the need for complementing third-person research with first-person research, consider the following example: An experimenter is interested in identifying the physiological correlates of a particular facet of anger, for instance by measuring patterns of brain activity or the participant’s heart rate. Once we have established a sufficient number of correlations between this particular facet of anger and the corresponding heart-rate we are on safe ground. But how can we initially focus on the specific facet of anger in a reliable way – if not via an introspective exploration of the immediate phenomenology of anger? Or put the other way round: do we systematically limit our exploration of anger to the aspects that we already have ready-made mental or verbal categories for? Already in the phase of formulating research questions, as well as in the phase of conducting the study, both the experimenter and also the participants make use of superficial forms of introspection, which results in an understanding of the phenomenon of interest based on a limited and pre-conceived notion of what anger is.

If this pre-conceived understanding is obtained without any systematicity, it might very well consist of a biased, premature or shallow construct that lacks the more differentiated and subtle facets of the anger experience, which can only be obtained through a systematic form of introspection. Clearly, methods of obtaining, analyzing and interpreting the reports of behavioural and physiological measures of participants have been advanced significantly over the past decades. In contrast, the experiential spectrum from which these verbal, behavioural and physiological measures have been obtained has hardly been developed (Bitbol and Petitmengin 2017) in a such fine and rigorous way that it exceeds common human wisdom. Here, we suggest that the conceptual understanding of anger can be enriched above and beyond what an one-off, ad-hoc introspective account of a group of naïve subjects participating in an anger experiment can provide. In addition to enriching our conceptual understanding of the qualitative side of anger, we were thus also interested in reflecting on the method used to obtain such results.

While introspection has been viewed with suspicion ever since its inception in the early twentieth century (Danziger 1980; Jack and Roepstorff 2002) there has recently been a movement to advance introspective enquiry in a systematic manner and introduce it into psychology’s methodological repertoire in an effort to gain a more in-depth understanding of psychological phenomena (Clegg 2013; Overgaard 2006; Weger and Wagemann 2015b). Despite the critical stance towards introspection, it cannot be denied that it is crucial in unravelling human existence. Questions about consciousness or the constitution of the self have long been posed and elaborated by philosophers and scientists alike. Answering such questions requires both a focus on external circumstances as well as on internal aspects, including feelings, sensations and thoughts. Such an internal focus can be regarded as a kind of self-observation or introspection. Several philosophers such as René Descartes and John Locke have pointed out that the observation of internal states is essential in the human endeavor of self-understanding. The just mentioned philosophers, but also early experimental psychologists including William James, Wilhelm Wundt and Edward B. Titchener, utilized introspection and regarded it as a kind of perceptual process focusing on inner psychological states (Butler 2013).

Over the course of history, introspection has typically been neglected as an empirical method, but it did never fully vanish (Clegg 2013). Recently a revived discussion about introspection has taken place both in psychology (Weger and Wagemann 2015b) and philosophy (Schwitzgebel 2008). A more comprehensive account of the historical details about introspection within psychology is beyond the scope of this article, but the interested reader is referred to the relevant literature (e.g. Butler 2013).

Currently, new attempts have been made to integrate first-person methods in the study of psychological states (Bitbol and Petitmengin 2017; Jack and Roepstorff 2002; Varela and Shear 1999; Weger and Wagemann 2015b). The micro-phenomenological interview technique, for instance, which was developed by Petitmengin (2006) is well-established and used to assess subtle and procedural aspects of psychological experiences. Using the micro-phenomenological interview, it could be shown that a systematic use of introspection is successful in overcoming the subjective biases (Petitmengin et al. 2013) that have been reported in some of the studies most critical of introspection (Johansson et al. 2005; Nisbett and Wilson 1977). Moreover, the core pillars of psychological experimentation (systematic observation, use of adequate control conditions, replication) can also be implemented in an introspective research-context (Weger and Wagemann 2015b). While we would not deny that introspective enquiry is vulnerable to a host of potential flaws and inconsistencies, we also argue that it is no more so than other forms of enquiry (including experimentation – though the vulnerabilities occur on other points in the research context) and that there are also in principle no reasons why such flaws should not be addressed and redeemed (for in-depth reviews, see Bitbol and Petitmengin 2013; Weger and Wagemann 2015a).

Furthermore, we would like to emphasize that the introspective approach used in this study shares similarities with other phenomenologically based approaches such as autoethnography (Döring and Bortz 2016; Ellis and Bochner 2000), heuristic qualitative research (Moustakas, 1994), phenomenological inquiry (Giorgi et al. 2017) and dialogical introspection (Burkart 2010). The introspective approach used in this study aims to grasp more subtle aspects of subjective phenomena by suspending a split between research participants and the researcher. Other approaches such as autoethnography (Döring and Bortz 2016; Ellis and Bochner 2000; Hedican 2006) also combine researcher and research participant in one subject. However, while the main focus of the autoethnographic approach is to observe inner states induced by external circumstances more generally, the introspective approach used in this study aims to also capture how experiential phenomena unfold and how they are structured. The suspension of the split between the researcher and participant can also be advantageous, because it facilitates the researcher to immerse into the experiential aspects of the investigated phenomenon. This way the researcher herself becomes acquainted with the phenomenon under study and can use this knowledge to enrich the understanding of the phenomenon (Berger 2015).

Another, more general argument in favor of detailed phenomenological descriptions is, that they tap into the genuine aspects of psychological research constructs (i.e.,what a researched construct essentially is about). These genuine aspects refer to the constituting parts and boundary conditions to other related phenomena. Without such knowledge prior to undertaking third-person research and developing theoretical models, such models will be hollow with respect to the inherent valuable qualitative dimensions, which is, ultimately, what people experience.

On a more general account, introspection is suited for detecting aspects of phenomena which are more subtle and therefore ignored in non-systematic, everyday inner observation (Weger et al. 2018). In a related vein, this point was brought forward by Francisco Varela already in the 1990’s. He argued that a first-person approach is necessary for investigating mental and psychological phenomena. He emphasized using methods which sensitize awareness, such as contemplative and meditation techniques in order to find deeper insights of subjective experiences (Varela 1996; Varela and Shear 1999). In contrast to third-person methods, first-person approaches aiming at detecting such deeper or subtle information are scant. Hence, its ongoing development is important and also its combination with the prevailing third-person methods because only this makes a multi-perspective observation of phenomena possible (Hackert and Weger 2018; Lumma et al. in press; Pérez-Álvarez 2018; Weger et al. 2018).

As will become apparent in our methodological protocol, the introspective method we are advancing below is far from the stage of perfection that the experimental method has accumulated over the past centuries (cf. Kiene 2013). This is not surprising, given that it is only just in a revival phase. To our knowledge, only one study has yet explored anger introspectively (Burkart 2010), although it differed significantly from the way we understand introspection. During the course of the study, we became acutely aware of the need to not only use introspection as a method of studying a particular subjective experience, but that we also realized that we need to refine and develop this method further. In addition, to our enquiry into the emotion of anger, we hence added a separate section in which we outline the observed methodological challenges. We describe how we tried to address these challenges in the present study and propose ways how to deal with these in subsequent research, identifying a set of best-practice guidelines that we expect may beneficially broaden research on emotions in a first-person context in general.

Methods

We worked as a group of five German-speaking co-researchers, two female researchers and three male researchers, and three of them had been working in a team for 7–13 months and also collaborated on another introspective study, investigating a different phenomenon, prior to the start of the anger trial. The remaining two co-researchers joined the team 2–4 months prior to the start of the anger trial and one of the co-researchers worked on an introspective trial for the first time. In detail, the group consisted of one male psychology professor who has several years of training in meditation and who investigated several different phenomena with the proposed method of introspection outlined below. One male researcher with a doctoral degree in philosophy who has extensive experience in meditation and who investigates meditative states utilizing phenomenological approaches. One female researcher with a doctoral degree in cognitive sciences who had researched meditative states and who has some experience with meditation and one male researcher with a master’s degree in psychology who had approached the subject of introspection mostly via historical literature reviews. Finally, one female master student of neurocognitive psychology was part of the current research group who had no prior training in – or knowledge about – introspective enquiry. All participants except the student form a loose but permanent research group. There was no systematic approach to build the research team and hence there were no assigned roles to the different members of the group. Rather, the research topic grew out of an existing team and hence the approach was developing on the fly as far as the social set-up of the group is concerned. No researcher was an expert in the field of anger research. Outside the individuals’ varying level of expertise, there was no preliminary training in introspection before the trial. In the group meetings during the introspective trial, an open atmosphere was present and all researchers had the chance to speak freely and had as much time as they wanted to communicate their introspective experiences and to discuss relevant issues. The formal roles of the group members had no detectable influence on speaking time or possibility to talk about issues, although the researcher who is a professor could be seen as the informal leader of the group who convened the discussions and moderated them to some extent.

The methodological procedure of the current introspective trial was based on a protocol that we have used repeatedly in the past (e.g. Lumma et al. in press; Weger et al. 2018). Experiential data were collected over a period of six weeks and discussed in group meetings which were attended by at least three of the five co-researchers. The anger experiences were observed in the co-researcher’s everyday life and recorded in a diary as soon as a possibility for quiet note-taking emerged.

The time interval between the experience and the diary recording could range from only a few seconds or minutes after the experience up to several hours and longer. The frequency of the observations varied from person to person from at least one to several observations per week. The content of the diary entries contained descriptions of the observed experiences and did not follow any pre-defined structures. Diary entries could be either kept in paper form or digitally. Each participant created their diary entries in an individual manner without following any general procedure. The content of the diary entries were discussed in each group meeting and each participant had the chance to leave out private content.

As the instances of anger were observed as they occurred in everyday life, no deliberate steps could be taken to ensure a minimum number of experiences. Still, with ongoing practice more – and more subtle – instances were noticed. The group meetings allowed for a systematic discussion of observations and served as a basis for developing and refining questions to be addressed until the next meeting (an overview of the guiding questions can be found in Appendix 1).

Refinements of the research questions were made regarding content as well as method. The meetings included a presentation of the introspective data by each member and were discussed amongst the group with respect to clarity and detail. Discrepancies and common observations were noted and served as a basis for refining and generating new hypotheses. In addition, questions occasionally emerged in the context of an introspective trial. Such questions were also reported in the meetings, discussed and, if decided by the group, were kept for further introspective analysis.

In contrast to a prior investigation to study anger introspectively (Burkart 2010), our observations of anger lasted up to several weeks allowing us to detect more detailed and more fine-grained aspects of anger which are not reported by Burkart (2010). In addition, we had a stronger focus on cognitions and the origin of anger.

As is apparent from the preceding paragraph, our methodological protocol – though it is based on a set of existing publications – has still to be described as work in progress. Our joint background in experimental psychology also called us to be as conservative and careful as possible in our approach to anger through this new and developing form of methodical enquiry. From the second meeting onwards, we specifically discussed the methodological issues and decided to scrutinize the emerging challenges systematically in an effort to develop approaches that could deal with these challenges. We address these methodological issues in a separate section of our paper, since it was agreed that they deserved special attention. In a final meeting, we identified experiential dimensions of anger which the co-researchers identified in their diary-reports. We did not go by frequency but by attributed prominence. Our approach can thus be described as a form of Interpretative Phenomenological Analyses (IPA, Smith and Osborn 2008) in which the researcher does not try to bracket her expertise but uses this expertise in a reflected manner to make sense of the data. We consider such a method particularly useful in a context where the research participants and the researchers documenting the data are one and the same – as is the case in first-person research. It is a kind of participatory observation – but an internal participatory observation, not an external one. In the following we will first describe the observed dimensions of the anger experience (sections Cognitive, Affective, Somatic and Volitional Dimensions of Anger, Temporal Unfolding of the Anger Experience, Inhibition Versus Expression of Anger, Defining Different Types of Anger) and continue with the observed methodological challenges and suggestions on how to deal with these challenges (section Methodological Challenges and Opportunities of Introspective Research – Implications beyond the Anger Context).

Results

Cognitive, Affective, Somatic and Volitional Dimensions of Anger

The introspective trial revealed that the experience of anger can be differentiated into a cognitive, affective, somatic, and a volitional dimension and each dimension is characterized by different aspects that accompany the anger experience. The cognitive dimension refers to thoughts, which can occur both deliberately as well as incidentally. The affective dimension refers to subtle feelings and the somatic dimension refers to bodily sensations accompanying the anger experience. The volitional dimension refers to behaviours and behavioural motivations emerging before, during and after an anger experience.

Unlike the affective dimension, the somatic, volitional, and cognitive dimensions could be put into words relatively easily because they could either be identified in the body (attention to somatic experiences is well rehearsed and requires fine-tuned differentiation skills), observed through the outcome of an action, or noticed through one’s content of thoughts. It was more difficult to verbalize the affective dimension due to its subtlety and lack of experience in systematically delineating its qualitative nature. It could be best described as containing a twofold gesture: on the one hand there was a yearning or striving towards something that could not be accomplished according to one’s liking or understanding (often described as the frustration aspect; but this is only part of the story – it can also emerge, for instance, out of a fear of losing something that is appreciated); paired with this there was a tendency to offload and derail the inhibiting or counteracting circumstances. There was a prominent experience of a difference in interest between one’s own position and that of others. The subject-object distinction became highly salient and one felt stretched across this divide as one was realizing that the fate was unfolding in the other position’s favour. This tension uncovered a remarkable energy – calling up experiences on the somatic as well as the behavioural dimension. Experiences along the somatic dimension of anger comprised a variety of bodily feelings such as the rising of energy from the lower abdomen towards the head and arms, feelings of heat, muscular tension, cognitive alertness, vitality, and an accelerated heart rate. “I felt warmth in my body coming from beneath and going up to the chest. Further, I felt warmth in my hands, a little bit in my arms, too.” (individual observation of a co-researcher during the first round of observations). Further examples of raw observations can be found in Appendix 2. Experiences along the cognitive dimension included either the absence of thoughts, or escalating cycles of thoughts that were tunneled towards the inhibiting factors (cf. anger-rumination, Denson 2013; one was simulating how these inhibiting factors should be dealt with and how one could become involved in this).

Components of the volitional dimension were experienced as a wish to counteract or destroy something, to keep someone within their limits, or as an impulse to move one’s body in order to discharge the built-up energy. The remarkable energy that was suddenly available was stunning. However, it was experienced as a heteronomous energy and in drawing upon it, often enough we later experienced a sense of exhaustion and a feeling that one had been living beyond one’s resources. An individual experience during the first round of observation was the following: “Anger feels strenuous.”. The sense of heteronomy stretched across all dimensions of the anger experience, that is across the cognitive, affective and somatic dimension. Furthermore, anger can be directed towards others or towards oneself and discharging it can result in the wish of physically harming someone or destroying an object, or resulting in passive-aggressive behaviour, e.g. through violent communication or ignoring someone. For example, in one of the group meetings we described the accompanying gesture as “‘dashing forward’ [losstürmen] to the other” and as a “destructive intention” (aggregated observation during the second group discussion).

Temporal Unfolding of the Anger Experience

Our observations revealed that the experience of anger unfolds in different phases over time (see visualization in Fig. 1). The onset of the anger experience was first noticeable after being caused by a specific trigger (initiation phase). The different triggers noticeable in this early phase were either based on a feeling of tension and inconsistency regarding one’s relation to the world (comparable to the self-other divide described above), unfulfilled expectations, or the feeling that oneself, or something that is important to oneself is being unheard (e.g. the world is not how it should be). In addition, it was also possible to become aware of the onset of the anger experience through some of the somatic sensations described in section Cognitive, Affective, Somatic and Volitional Dimensions of Anger.

Fig. 1
figure 1

Visualization of the temporal unfolding of the anger experience in different phases and description of an exemplary situation eliciting an anger experience. Here, an example is depicted where anger reaches a high level. Such built-up of energy, however, can discontinue and turn into a suppression or metamorphosis of anger at any point prior to reaching the peak (not depicted in Figure 1)

Full size image

The subsequent internal unfolding phase of the anger experience began with the arising of thoughts related to injustice and frustration, which led to the intention to unwind from something unwanted. A swinging back and forth between these thoughts and moments of clarity could be noticed within the unfolding phase and was concurrently accompanied by a built-up of energy. At this stage it was still possible to interrupt and conquer the anger. The oscillatory phase stopped when it became apparent that there were no more possibilities to deal with the current circumstance and that led, again, to an increase in the already built-up energy. The further progression of the unfolding phase also included the directedness of the anger energy towards a target. In some cases, we observed that the target, which elicited the anger, was already familiar. In other cases, we observed that a deliberate seeking of a target towards which the anger could be directed was necessary. A target could be located externally and involve living entities and inanimate objects, but it could also be located internally and directed towards oneself.

In most cases the unfolding of the anger experience terminated with a release of the built-up energy towards the identified target, which was accompanied by feelings of satisfaction, but also feelings of fatigue and lack of energy (expression phase). The built-up energy could result in different types of releases towards the identified target including the generation of anger related thoughts, verbal expressions or physical actions. The release could either be based on an intention to direct someone towards a case of injustice or the intention to harm someone. We also observed situations in which the anger energy rebuilt after a first initial release. Such recurring waxing and waning of the anger energy was furthermore accompanied by a feeling of mental and physical exhaustion. In some circumstances it was observed that the built-up energy was not released immediately in the situation, which re-initiated and even amplified the anger experience. The absence of the anger expression resulted in a continued fixation on thoughts and feelings related to injustice and frustration and could also be experienced as bodily tension. Finally, some of the co-researchers had the experience that they could transform and constructively use the energy induced through the anger experience for different purposes. To do so, first it was necessary to have the cognitive capacity to notice the early and subtle stages of the anger experience. Next, it was necessary to exert cognitive control over anger related thoughts and direct them towards adaptive purposes. However, one of the co-researchers did not share the experience of constructively using the built-up energy of anger. Similar to modifications of parameters made within classical experimental paradigms, we further explored this particular finding in an additional round of observations.

Inhibition Versus Expression of Anger

It is a widely held view that anger should be inhibited in order to avoid aggressive behaviour and harming other people (Saini 2009). A striking observation in our introspective trial revealed that the bare act of introspectively observing the experience of anger resulted in a dampening of the intensity of the anger experience. “Anger becomes ‘cognitively’ controllable (it changes) when observed as an object under investigation.” (aggregated observation during the second group discussion). This observation is in line with a widely held concern that introspection alters the observed experience itself (Bitbol and Petitmengin 2013). Suggestions about how to deal with such a methodological challenge of the introspective method will be further elaborated in section Methodological Challenges and Opportunities of Introspective Research – Implications beyond the Anger Context. Interestingly, what is a challenge in methodological terms may become a fruitful resource in terms of intervention programs. Given that people can find a motivation to observe their anger response, this may lead to a natural reduction of it, which contrasts with cases where anger is suppressed through the exertion of willpower and effort. A deliberate inhibition of anger was generally effortful and requires the cognitive capacity to become aware of the early and more subtle phases of the anger experience. If it was possible to attend to the early unfolding phase of the anger experience, cognitive control strategies such as cognitive reappraisal of the situation could help to inhibit the escalation of anger-related thoughts and the built-up of the anger energy. If we first became aware of the anger experience at the somatic level, it was possible to control the built-up of energy through the regulation of one’s breathing. “There was a moment which could have had made me angry (bus just missed and a loud lawn-mower) but I kept relatively calm, via cognitive reflection, observation of the body and control of a steady breath” (individual observation during the fourth round of observations).

The control of the anger experience felt easier at the somatic level as compared to the cognitive level. We also observed that the inhibition of anger was particularly challenging in social conflicts because they often involve a flood of overwhelming emotions which incapacitate the ability to deal with the conflict. To summarize, these findings indicate that it would be timely to enquiry whether it is possible to train these cognitive resources and intervene at earlier phases of an anger experience through introspective training.

Defining Different Types of Anger

According to our observations, we suggest that it is meaningful to differentiate different types of anger based on the outcome of the anger experience. The most frequently experienced type of anger in our introspective trial was of a destructive type, encompassing the experience that an injustice has occurred to oneself or someone close to oneself and the ensuing intention to eliminate or harm the target object of anger (or a substitute object). This we call the sanctioning anger as it is primarily directed to dealing with and punishing the attributed trigger of our anger experience (e.g. another person). A more constructive type of anger was experienced less frequently and led the experiencer to the intention to restore justice by expressing the anger energy towards a target object in order to make the target object (either oneself or someone else) aware about the injustice. Here, the anger is directed not so much towards destruction but towards prompting change. We call this restorative anger. The distinction bears a certain resemblance to that between hostile vs. instrumental aggression in more general terms; but note that anger in itself is typically understood as a subcomponent of aggression and not to be confused with it as a whole.

A third type of anger could best be conceptualized as a wise type of anger – it was an experience of waking up that also catapulted attention towards an injustice. However, with this type of anger the experiencer did not let it escalate into a full-blown anger outburst. The acting out (either mentally or by action) did not occur. Hence, the usual masking of more subtle cognitive, affective, somatic, and volitional nuances did also not occur and the experiencer could become aware of their more subtle facets more easily. In particular, the anger did still fulfill its wake-up function. However, it did not outmaneuver or incapacitate the experiencer. We call this latter type of anger the alerting anger.

Methodological Challenges and Opportunities of Introspective Research – Implications beyond the Anger Context

Unlike in our earlier research, the members of the current research trial formed a rather heterogeneous group: one person was completely unexperienced in introspection and was also a newcomer to the team, one had worked on one introspective study before in the same research group; another was likewise a team-member and had also experience in meditation; one person had much experience in meditation but little experience in doing introspection in the way conducted in the current research team (which he had only just recently joined), and one had experience with both introspection and meditation. Thus, the level of introspection experience ranged from not existing to advanced within the current study-group. It is worth noting that experience with meditation could be helpful for introspective tasks as some of its core practices are also relevant to introspection, like being in a state of sustained attention and not getting distracted by irrelevant perceptions (Carter et al. 2005; Lutz et al. 2008; Malinowski 2013). This constellation prompted different questions and problems pertaining to the method – which did not come up with similar urgency amongst the more homogenous groups of earlier introspective trials. We turned what appeared initially to be a challenge into an opportunity and discussed the issues systematically in our meetings. Here, we will summarize and outline those methodical issues in an effort to develop best-practice guidelines for future introspective research.

Anger Experiences Change during Self-Observation

During our introspective enquiry we observed that some parameters of the anger experience appeared to change as a result of attending to them. For instance, the intensity of the anger experience decreased as we took a more reflective perspective. As a matter of fact, the core experience of anger waned and sometimes disappeared completely. As we were making this observation we first considered it to be a substantial challenge to introspective enquiry (cf. a similar observation of Franz Brentano as discussed in Meyer, Hackert and Weger 2018). However, the following points qualify our initial reservations.

To begin with, this problem is not unique to introspection, as third-person methods are confronted with the same problem. This can be illustrated on an epistemic and a pragmatic level. On the epistemic level, objects of investigation appear differently depending on the perspective of the researcher. For instance, electrophysiological brain activity is represented as colored images on a computer screen, subjective experiences are converted in numerical categories, and bodily rhythms are represented in mathematical diagrams. Any kind of qualitative information inevitably has to be translated into symbols so that these symbols can be comprehended by others. In this translation process, qualitative aspects of the experience are typically lost. It is true that the translation into a numeric level per se may not alter the experience in the participant – whereas such a change occurred in our self-observation trial. But note that the laboratory environment of making such third-person observations will likely also have an impact on the experience (cf. Barrett et al. 2011). After all, is it legitimate to claim that an fMRI setting would be less likely to intrude on the experiential quality compared to self-observation? Moreover, while the intensity of the emotion was changed, we did not notice any qualitative differences in the content of the emotion as a result of the introspective trial. If anything, introspection would hence weaken effects and thus require more conservative testing. On a pragmatic level, the Hawthorne-effect as well as other participant and experimenter effects (demand effects; expectancy effects) illustrate that this problem is not specific to introspection but rather a universal issue in psychological science and beyond (cf. McCambridge et al. 2014; Rosenthal 1994).

Also note that in any third-person experiment, when a participant is asked about her emotional experience she has to introspect on it – even if it is the most elementary type of introspection, for instance in a manipulation-check which is usually administered via a questionnaire, but which still requires an elementary form of introspection. Taken together, we treat observations from introspective enquiry as equal to observations from third-person enquiry with regard to uncovering the nature of a phenomenon.

Related to this general issue is the question whether introspection is a kind of real-time method or whether it is always retrospection. Retrospection bears the problem of possible memory distortion, since one has to recollect all information from a past experience. According to our understanding, introspection is always retrospection. This is obvious, as the real experience is over when the physiological parameters begin to fade. That is, they become low in intensity, or even stop altogether (but see Strawson 2011, who makes a reasonable point for simultaneous introspection). We would like to emphasize that the question whether introspection grants real-time access to psychological phenomena or not deserves further discussion that goes beyond the scope of this paper. Notably, already early introspectionists like Bühler or Wundt both stated that complex inner phenomena cannot be observed simultaneously; although the former claimed such observation are possible through retrospection, the latter rejected retrospection altogether as a sound method (Danziger 1980; Hackert and Weger 2018). Still, a simultaneous introspection can be approached asymptotically – that is, we aim to conduct the introspective enquiry in as close of a temporal proximity as possible. In most cases, this is certainly in closer temporal proximity than even in case of a questionnaire in which participants are asked about their past experiences dating back from days to sometimes even years.

Individual Experience Is Influenced by the Reports of Others

As introspective data continued to be reported and discussed in successive group meetings, one member noticed that she subsequently observed some aspect of anger she had not observed before (in this case: ‘feeling helpless’). Thus, the concern emerged that observations are triggered and reported by participants, which not occur reliably in their own experience, but rather through looking out for them (i.e., a form of priming effect).

As a preamble to dealing with this issue, note once more that this is not a problem specific to introspective research but to the research process in general, as is evident from the philosophy of science literature, at least since Ludwig Flecks paradigmatic work on thinking styles and thought collectives (Fleck 1993). Fleck – researching the development of the medical understanding of the syphilis disease and its transmission – observed that researchers enquired in directions that were consistent with what their peers considered promising and meaningful. Such an approach can be costly and time-consuming, especially if the early understanding of a problem is far from reality, as was the case in the syphilis example.

But beyond this general caveat, there are also other points to raise regarding the issue of intersubjective influence. Intersubjective influence is not a bad thing per se. It is a form of learning that is in fact highly important: we do not take note of – and cannot discover – things that we do not have concepts for or for which we are lacking at least some elementary form of mental representation. The challenge is not the influence per se; but the ability of each individual to follow her own compass in identifying what is reality and what is fiction. This is a universal issue, especially in the social sciences and social psychology context. And yet, the social influence literature provides clues on which people are particularly vulnerable to such forms of influence (e.g., newcomers and people without a reputation yet), allowing to provide for special protection of their observations in a social research context. Moreover, there are a number of precautions one can take to shield participants of a research process against a disproportionate influence of overly dominant opinions:

  1. a)

    Ensure a careful documentation of introspective observations including questions like “how sure and confident are you? “and “did you describe your experience fully and in detail?“ In addition, document these observations prior to sharing them with colleagues.

  2. b)

    Compare protocols (especially with a focus on consistency) illustrating pre- and post-observations of such aspects (i.e., the protocol illustrating the episode when helplessness was not noticed with the protocol illustrating the episode when it was noticed). Also conduct a comparison of protocols of different members.

  3. c)

    Formulate new hypotheses (e.g., contrasting hypotheses) and test again (reliability). This can be checked in subsequent introspective trials. In these, the introspecting researchers have to ask themselves whether the suggested aspect is really noticeable while introspecting or not. If so, it suggests new possibilities of experience (i.e., the experienced aspect is no artifact). If not, one’s own experience is disproved – indicating it was just an inducement.

  4. d)

    Prolonging the time intervals between group meetings to ensure that every group member collects a sufficient amount of introspective observations, which enable the group member to detect a multitude of different aspects of the phenomenon under investigation without help from others. Such safety standards are well established and documented for third-person methods and widely taught in undergraduate psychology textbooks (e.g., randomized, controlled, and blind designs). Likewise, safety standards could also be adapted to a first-person approach, as discussed above. In this regard, it is worth mentioning that the research member with the least amount of introspective training noticed an improvement of her introspective capacity over time. For example, she repeatedly noticed more relevant events and felt more confident with regard to whether she really experienced certain aspects of the event.

Rare Appearance of the Phenomenon of Investigation

In contrast to an earlier introspective investigation – that of empathy – all members mentioned that anger was not observed frequently. This seems realistic, since several questionnaire studies report only one or two incidences per week (e.g., Averill 1982; Schieman 2003). The problem of such relatively rare frequency is that it is more difficult to deliberately and systematically explore a phenomenon, draw comparisons across occurrences, and draw lawful inferences from this.

Once more, this problem is not unique to introspective research – ranging from all areas of natural science (extraterrestrial space experiments are usually conducted only once) to psychology (cf. single case studies, for instance in neuropsychological research) and much has been written about the legitimacy of such work. In addition, we experimented with mental simulation, in which we envisioned ourselves in the original situation of an experiential event as carefully as possible. This allowed us to re-scrutinze experiences and, hence, study them more systematically. Also, with progressive introspection the phenomenon was observed more regularly. This was possibly due to a better detection of more subtle states of anger or because of less (automatic) resistance to observe such aversive states.

Despite the mentioned shortcomings, bear in mind that the ultimate scale of judgment of an individual always remains her introspective insight, even when it comes in the most elementary form, such as translating numbers from a measurement device and translating them into a mental representation that we have to call up (Weger and Wagemann 2015b). Psychological occurrences (including data) have to be represented in consciousness, that is we have a subjective view on these occurrences, including data. We cannot advocate a methodology that tries to circumvent or eliminate any reliance on the human mind. An early pioneer in introspective psychology, Franz Brentano, made an important point in this direction, introducing the concept of evident experience (or evidence, Brentano 1874). When we have conscious experience, for example a perception (of something we see, hear, smell, or otherwise), then we can know that this experience is our reality. We know this when we have inner evidence of this experience, that is, an inner sense of truth about what we experience right now. It makes no sense to debate and question an inner sense of evidence – for then we would also have to question our skepticism. Inner evidence rather provides the basis of our experience (see also Meyer et al. 2018, citing Brentano 1874). As such, when there is inner evidence about anger episodes while introspecting on them they provide suitable data for investigation to begin with. Finally, remember that we are advocating an approach in which we propose a first-person approach to complement rather than substitute a third-person approach. We also realize that the different approaches can be useful at different stages of the research process: First-person research may prove particularly useful at the early stages of enquiring into a scientific problem, which can be regarded as the so-called context of discovery, where new ideas are proposed and little empirical evidence is yet available as to the appropriateness of one line of thinking over the other (Hoyningen-Huene 1987). By contrast, at a later stage, when ideas are empirically tested and scrutinized, third-person research may play a more dominant role. This later stage can be regarded as the so-called context of verification. In this way, both approaches can be used in tandem to advance a particular issue in psychology.

Conclusion

In this study, our aim was to elucidate the experiential side of anger. To date, research pertaining to anger – and to emotions in general – often lack a detailed view on how such emotions emerge and unfold in conscious awareness. Here we argue that such knowledge is an important aspect of psychological enquiry. It provides not only an insight into the first-person phenomenology of psychological experiences but also allows for a more holistic approach to psychological phenomena than the method of decomposition that seeks to explore the microstructural components (mediators, subcomponents etc.) of a psychological phenomenon.

The results from our introspective trial revealed several insights regarding the experiential nature of anger which are in line with findings from third-person research. We observed that the anger experiences consist of a cognitive, affective, somatic, and volitional component and these findings are consistent with data gathered through third-person research. In addition, to the generally more coarse-grained temporal profile of how state or trait-related anger develops we identified more subtle nuances and distinct phases of the temporal unfolding of the anger experience. Furthermore, from third-person research it is known that anger can be controlled to a certain extent, but it is not exactly clear at which specific phase an inhibition of anger is still possible. Our findings showed that anger wanes once it is introspectively observed, but that cognitive effort is required to deliberately control more subtle aspects of the early developmental stages of anger. These findings can be useful in a clinical domain and inform the development of interventions specifically targeted at improving cognitive resources to control early stages of anger. We also pointed out that introspection does not need to be an impromptu or ad-hoc mode of science but that it can be advanced methodologically, although we certainly encountered problems in this context. However, these problems do not seem to be of an apodictic nature that could not be addressed in principle. By contrast, they can be used to advance and refine the method to make them into a more suitable instrument of the research process and contribute to the diversity and complexity of phenomenologically based research methods (Levitt et al. 2018; Levitt et al. 2017). Finally, our findings provide first-person based phenomenological descriptions of anger experiences, which can be independently examined by other researchers and potentially add to a roadmap of experiential states (Weger and Wagemann 2015a, b).

On a more general note, several points need to be addressed in order to advance the study of more subtle aspects of psychological experiences using methods like introspection. Despite the revived interest in introspection, an elimination of the researcher-participant split is typically seen as being anachronistic. Future research and theoretical discussions are needed to clarify the advantages of this practice and how to best integrate it into the psychological repertoire. Furthermore, it might be helpful to re-consider the definition of introspection and first-person methods more generally. A constructive stance towards introspection might be to regard it as a complementary method that has both advantages and disadvantages like any scientific method does (Butler 2013). It should also be discussed whether a systematic training of the introspective method can be developed so that it can be embedded in research settings in a systematic and standardized way that allows for greater replicability and intersubjective validation.

Finally, instead of focusing on the shortcomings of the introspective method, an openness with regard to methodological pluralism is important. Introspective enquiry or other kinds of first-person methods might be appropriate to address specific research questions, while others require different methods. If a researcher is interested in studying qualitative aspects or the processual character of a psychological phenomenon, introspection might be a worthwhile method. Future research should also focus on how introspection can be combined with other methods and how such combinations of methods can inform complex research questions. Another consideration that should be taken seriously in the advancement of a psychological science which seeks to integrate first-person methods, is to enable a greater crosstalk between different disciplines. Especially in the early stages of developing novel methods it can be beneficial to learn from other perspectives and also consider potential overlaps with different perspectives. In the long run, a potential goal of psychology could be to create a map of psychological states that displays both experiential aspects of psychological states measured through appropriate first-person methods and behavioral aspects of psychological states measured through appropriate third-person methods.