Introduction

Emotion research, including cross-cultural studies, accepts six basic emotions—happy, sad, fear, anger, surprise, and disgust. Most of the empirical studies in psychology have focused on them. The Sanskrit word for emotion is ‘bhava’ which means ‘something like a state of mind that becomes or that is one that movements flow from’ (Shweder & Haidt, 2000). Unlike the six basic emotions referred in the Western literature the third century AD classic Indian text, Natya Shastra, by Sage Bharat proposed the Rasa theory wherein he listed eight rasas, namely shringar (the mood of eros), vira (knightly mood), karuna (the mood of pathos), raudra (angry mood), bhayanaka (the mood of terror), bibhasta (the mood of revulsion), hasya (the mood of jocularity), and adbhuta (the mood of wonder). Later Abhinavagupta propounded the ninth rasa, shanta (the mood of total freedom), where one neither experiences happiness nor unhappiness. The corresponding bhavas are rati, hasya, shoka, krodha, utsaha, phaya, jugupsa, vismaya, and saama.

According to Keltner and Haidt (1999), emotions play individual, dyadic, group, and cultural-level functions. At individual level it prepares the person to express, while at dyadic level it is crucial for understanding other’s emotions. At group level it serves the purpose of knowing group-specific rage and status and at the cultural level it ultimately helps learn cultural ideologies. The emotions arising out of transgression of cultural ideologies boil down to shame/self-condemnation, guilt/remorse, and regret (Fisher & Exline, 2010). This compels us to expand the scope of empirical study to other complex emotions such as shame, guilt, and remorse. The description of these complex emotions in the Western literature differs from the way it has been conceived in the Indian literature to some extent. Both the views are succinctly presented here.

Lewis (2000) describes shame as “a highly negative and painful state that also results in the disruption of ongoing behaviour, confusion in thought, and inability to speak” (p. 629). Shame has been explained as self-conscious emotion that represents “the effect of indignity, of defeat, of transgression, and of alienation…[it] is felt as an inner torment, a sickness of the soul” (Tomkins, Sedgwick, & Frank, 1995, p. 133). Considering its social and psychological significance, it has been referred to a ‘master emotion’. It has been argued that anger, fear, grief, and love will not be expressed if the person is ashamed of them (Cooley, 1922). It demands that the person be self-aware and emerge when the person or significant others evaluate the person to fall short of the moral standards. Thus, goodness, desirability, and appropriateness get attached to this moral standard (Tangney, Stuewig, & Mashek, 2007).

Shame needs to be reliably distinguished from similar constructs such as guilt, embarrassment, and low self-esteem (Rizvi 2010). Some researchers have argued that it may be labelled as self-conscious emotion and misclassified with guilt, embarrassment, and pride, or may be labelled as moral-emotion and classified along with guilt, regret, and remorse (Kroll & Egan, 2004; Tracy & Robins, 2004).

Researchers have distinguished shame–guilt on the basis of the public–private nature of the situation that elicits emotion (Benedict, 1946; Combs, Campbell, Jackson, & Smith, 2010; Smith, Webster, Parrott, & Eyre, 2002; Wolf, Cohen, Panter, & Insko, 2010). This viewpoint considers shame as a public and guilt as a private emotion. While shame arises out of publicly being exposed for something that is disapproved by others, guilt is a private emotion that is self-generated out of one’s conscience. On the other hand, many researchers suggest that shame and guilt inducing events are mostly public (Smith et al., 2002; Tangney & Dearing, 2002; Tangney, Miller, Flicker, & Barlow, 1996; Tangney, Youman, & Stuewig, 2009) as transgression or failure is known to people. The other basis to differentiate between shame and guilt is the focus; it could be on the self or the behaviour. According to Lewis (1971), negative focus on the self results into shame, whereas negative focus on a specific behaviour results into guilt. While shame has been considered more painful and emotionally disruptive, guilt is more adaptive. Despite being painful, guilt is relatively easy to mitigate as it is derived out of negative focus on a specific behaviour and not the self. Empirical studies support the difference between shame and guilt proposed by Lewis (1971). However, some studies report otherwise (Luyten, Fontaine, & Corveleyn, 2002; Sabini & Silver, 1997). Literature suggests shame and guilt to be related yet “distinct negative ‘self-conscious’ emotions” (Tangney, Stuewig, & Hafez, 2011, p. 2) which has been largely studied as dispositional constructs. According to Ketelaar and Au (2003), shame and guilt yield contrasting motives. While shame motivates a person to deny, hide, or escape situation that induces it, guilt motivates a person to confess, apologies, or try to undo the action that induced it. “The experience of shame is directly about the self, which is the focus of evaluation. In guilt, the self is not the central object of negative evaluation, but rather the thing done or undone is the focus” (Lewis, 1971, p. 30). According to Cohen, Wolf, Panter, and Insko (2011) guilt and shame have the self-behaviour and the public–private distinctions. Publicly exposed transgressions lead to shame, whereas transgressions not exposed publicly yield guilt (Combs et al., 2010). Both these distinctions have been empirically supported by researchers. The distinction between shame and guilt has been considered as the difference between attitude and intent (Fishbein & Ajzen, 1975). Shame and guilt overlap in the sense that both elicit repair and withdrawal behaviour (Schmader & Lickel, 2006). Shame is likely to foster avoidance and concealment, whereas guilt might foster apology and reparation. This has been empirically found using Test of Self-Conscious Affect (TOSCA: Tangney, Dearing, Wagner, & Gramzow, 2000; Tangney, Wagner, & Gramzow, 1989).

Remorse has been considered as a primary component of guilt experience. Researchers have also referred to self-condemnation as “an offence-specific version of shame” and remorse as “an offence-specific version of guilt” (Fisher & Exline, 2010). Shame and guilt differ in their orientation. While shame is oriented towards the whole self, guilt is other-oriented emotion focusing on specific transgression (Leith & Baumeister, 1998; Tangney & Dearing, 2002). Remorse, on the other hand, has been considered a situation-specific guilt associated with more repentance and acceptance of responsibility (Fisher & Exline, 2006). Studies suggest that remorse prompt reparative behaviour which in turn helps resolve guilt and shame (Fisher & Exline, 2010). If reparative action taken as a result of experiencing remorse can help in resolving guilt and shame, then it clearly states that one single situation or event can give rise to all the three emotions. On a similar note, although researchers have linked guilt to moral transgressions and shame to failures, both moral and non-moral, (Ferguson, Stegge, & Damhuis, 1991; Sabini & Silver, 1997; Smith et al., 2002) they are not reliably distinguishable on the basis of event types. This has been found for both, children and adults (Keltner & Buswell, 1996; Tangney, 1992; Tangney & Dearing, 2002; Tracy & Robins, 2006). The same event type induces guilt in one and shame in the other. Thus, the assessment of guilt and shame becomes very challenging and the consideration of event type while studying these three emotions becomes all the more necessary.

‘Lajja’ is the equivalent of shame in the Indian context. The Western literature refers to it as embracement, shame, or shyness (Hejmadi, Davidson, & Rozin, 2000). Lajja does not find a mention in the list of bhavas derived out of Natya Shastra but has been referred at several places in other texts. Unlike the Western connotation, Lajja has been accepted as a positive emotion. Menon and Shweder (1994) claimed that “Lajja will not be felt in a culture that sees hierarchy and exclusive prerogative of others as unjust than as powerful object of admiration and respect” (p. 183). According to Sibia and Misra (2011), “to experience lajja is to experience sense of graceful submission and virtuous, courteous well-mannered self…of the three emotions—shame, happiness, and anger—Hindu Indians view shame more positively than their European–American counterparts” (p. 295). We did not find similar indigenous study of aparaadhbodh (guilt) and pashchaataap (remorse).

Adopting multi-method approach Bhawuk (2017) analysed three epics, namely the bhagavadgItA, drugA saptazatI, and kAmAyanI to explain the concept of lajja and concluded that “lajjA has both internal and external aspects and synthesizes guilt and shame” (p. 2), both. This view is different from the view adopted by Western researchers who consider guilt and shame as distinct constructs. Other Indian researchers have also analysed Indian scriptures, and it seems that they agree that these scriptures play “important role…in the development of indigenous Indian psychological constructs” (Bhawuk, 2017, p. 3). BhagavadgItA and durgA saptazatI “converge on many ideas about lajjA [and present it]… as a reflective self-evaluation process that guides people in following the path shown by the zAstras or scriptures…lajjA mediates between desire and action, and is the wisdom that guides us to choose desires that are appropriate and reject those that are not” (Bhawuk, 2017, p. 21). The sixth sarga or canto of kAmAyanI refers to lajja. The fact that lajja is glorified in the scriptures and the prevailing cultural norms endorse its appropriateness necessitates the study of shame and guilt. While Lajja finds place in the work of some of the Indian researchers, we did not find empirical study of aparaadhbodh (guilt) and pashchaataap (remorse). Nor did we find any indigenous tool measuring shame (lajja), guilt (aparaadhbodh), or remorse (pashchaataap).

Initially Western researchers did not differentiate between shame and guilt (Tangney, 1995a; 1995b; Tangney, Burggraf, & Wagner, 1995). The early measures such as Buss–Durkee Guilt Scale (Buss & Durkee, 1957) and Mosher Forced-Choice Guilt Inventory (Mosher, 1966) also considered them as a single construct. Later, researchers started enumerating difference between the two. Researchers differ in their approach towards differentiating and assessing guilt and shame (Tangney & Dearing, 2002; Wolf et al., 2010). Researchers have used both quantitative and qualitative techniques to establish the distinction between shame and guilt. Those adopting qualitative method have used case study and content analysis of narratives, whereas rating of personal experiences, analyses of attributes, and counterfactual thinking have been used by researchers using quantitative method. But no such well agreed upon tool for measuring remorse was also found in the Western culture.

It merits mention that researchers suggest that shame and guilt are difficult to be distinguished on the basis of events. Further, the verbal-linguistic representation of shame and guilt has also been criticized as it does not allow making distinction between the two. In addition, the Western and the Indian connotations of shame (Lajja) and guilt are very different. This makes the mapping of similarity/dissimilarity between these three emotions difficult. Further, very little research is available on remorse and it has not been examined vis-à-vis shame and guilt. Considering the verbal-linguistic issue and the fact that social desirability is also likely to affect the self-rating of these three complex emotions, we attempted assessing observed similarity/dissimilarity among these three complex emotions. This was an exploratory study attempting to explore—(i) whether same scenario can result to two or more of these three complex emotions, and (ii) the observed similarity/dissimilarity between guilt, shame, and remorse. As no indigenous tools for measuring guilt, shame and remorse were found, scenarios were generated by gathering life experiences/narratives of people, rather than employing tools from the Western culture. The emotions of guilt, shame and remorse are fairly common and experienced by people on a regular basis. Studying and comprehending their intricate interplay, specifically in the Indian culture, would provide a more elaborate understanding of these emotions.

Method

The study was conducted in three phases—(i) generation of scenarios (referred as event above), (ii) development of illustrations and determination of item (scenario) equivalence, and (iii) classification and rating of the scenarios. As we intended to map the observed similarity/dissimilarity between the three complex emotions which are culture-sensitive, the participants at all levels (generation of scenarios, conversion to illustrations, and statistically measuring similarity/dissimilarity) were Indians.

Phase I: Generation of Scenarios

Ten young adults (eight males and two females) with an age range of 20-25 years (Mean = 22.5 years, SD = 1.8) were individually approached for this purpose. Three of them were undergraduates, while the remaining seven were postgraduates. All of them were native Indians. Initially, they were given proper explanation of the three emotions, namely guilt (aparaadhbodh), shame (lajja) and remorse (pashchaataap). They were asked to write as many life experiences of guilt, shame and remorse they could recall on separate sheets of paper in the next 3 to 4 days. Thereafter, these sheets were collected. A total of 34 scenarios were obtained of which 12 were for shame and 11 each for guilt and remorse. These narratives (scenarios) were further made precise by eliminating repetitions and grammatical errors. The word count of all 34 scenarios ranged between 40 and 55 words.

Phase II: Development of Illustrations and Item (Scenario) Equivalence

The scenarios were converted into illustrations. This was an iterative process, and the illustrations were finally selected after 2–3 rounds of modification. Each scenario was depicted using three illustrations thus completing a storyboard. This was done to take care of verbal-linguistic issues. However, few illustrations had couple of English words. The words were used to minimize speculation while understanding the event being illustrated. The English words were used in the present study due to the nature of sample. For the purpose of assessing item equivalence between the narrations and the corresponding storyboard, 14 native Indian participants (seven males and females, respectively) were recruited for this phase of the study. The mean age of the participants was 20.36 years (SD 1.01), and they were all undergraduates (mean education 15.5 years, SD 0.52). The participants were first shown the narratives for identification and rating. The exposure duration for each scenario (narration) was 12 s. The exposure duration was determined based on the previous work of Just, Carpenter, and Woolley (1982) suggesting that the exposure duration for one word is 240 ms. As the scenarios had approximately 50 words 12 s exposure duration was arrived at (50 × 240 = 12,000 ms; 12 s). After a gap of 24 h, they were presented with the storyboards and asked to identify and rate the emotion elicited by them. The scenarios were randomly presented following the table of random numbers. Each illustration was exposed for 5 s; thus, the total exposure duration for each storyboard was 15 s. After completion of each storyboard, the screen said “Enter Your Response” and the participants had to label the elicited emotion (shame, guilt, or remorse) and rate it on a 5-point Likert scale.

Phase III: Classification and Rating of the Scenarios

In the third phase, a separate set of participants were exposed to the scenarios in the form of storyboard and their responses were recorded.

Participants

A total of 138 male participants volunteered for the third phase of the study. The participants were inducted through an advertisement. The demographic characteristics of the participants were alike. They were native Indians from middle and upper-middle socioeconomic background. The mean age of the participants was 20.9 years (SD 2.07). All of them were undergraduates with the mean year of education 15.5 years (SD 0.70). They were explained the difference between the three emotions, shame, guilt and remorse. Borrowing from Proeve and Tudor (2016) for the purpose of this study, the three complex emotions were operationally defined as follows:

  • Guilt (aparaadhbodh): Wrongdoing as infringement of a rule or disobeying some imposing command.

  • Shame (lajja): Wrong action as a matter of failing to live up to some close standard that one expected of oneself.

  • Remorse (pashchaataap): Viewing the wrong done as a wrong done to someone, especially where the wrong somehow harms or hurts another person or creature.

After ensuring that they understood the three emotions as described in the literature, they were instructed to identify the emotion that best suited the scenarios and also indicate its intensity on a response sheet. The study protocol was duly approved by the Institutional Ethics Committee involving Human Participants (IEC) of the Indian Institute of Technology Kanpur.

Procedure

After explaining the difference between the three emotions, the following illustrative example was presented to ensure that the participants understood the three emotions.

“If I physically assault someone,

I might feel guilty (aparaadhbodh) if I feel that physically assaulting someone is not a correct thing to do, that I am disobeying or breaking certain rule (internalized rule that I have learned from a very young age);

I might feel ashamed (lajja) if I feel that it was not what I expected from myself, that I have failed to live up to my standards that how did I assault someone; or

I would feel remorse (pashchaataap) if I think in terms of the person I have hurt, that I have hurt another person and the person might be in pain and I am more concerned about the person here rather than have broken any rule.”

Thereafter the scenarios were presented in a random order. Randomization was done using the table of random numbers. The total exposure duration for each storyboard was 15 s, i.e. 5 s for each illustration. In their attempt to create a database “for researchers who are making decisions regarding exposure duration in the many studies that are currently using affective pictures as cues to elicit emotion in the laboratory” (p. 737), Codispoti, Mazzetti, and Bradley (2009) assessed “the effects of exposure duration on affective responding by parametrically varying exposure duration from 25 to 6000 ms while measuring evaluative, facial, and autonomic responses, as well as the late positive component of the ERP” (p. 731). For images not followed by visual masking stimulus, they found that “emotional engagement was strong and similar in magnitude, regardless of actual exposure duration. Even with very brief exposures, affective pictures elicited evaluative, cortical, and facial changes indicative of emotional engagement” (p. 737). Thus, the exposure duration of 15 s was arrived at for each storyboard. The exposure durations of narratives and illustrations were thus comparable. The participants were instructed to visualize themselves in the given scenario and state whether they would have experienced guilt (aparaadhbodh), shame (lajja), or remorse (pashchaataap). Each scenario was followed by the message “Enter Your Response” following which the participants had to tick on the radio-button thus identifying the emotion and also rate its intensity on a 5-point Likert scale where 1 represented minimum and 5 maximum intensity of that emotion. This was followed by another scenario. The inter-scenario interval was 3 s.

Results

Data Screening

While screening of data three univariate outliers were identified. They were removed from the data. Thus, the data obtained from 135 participants were analysed. Before exploring the answer to the two research questions at hand, it was important to statistically identify the scenarios that robustly assessed the three complex emotions.

Principal Component Analysis (PCA)

PCA was performed to identify which of the 34 scenarios robustly assessed shame, guilt, and remorse, respectively. Further, to analyse whether shame, guilt, and/or remorse can be elicited by same scenario or not, we performed PCA as a single scenario can elicit one complex emotion in some and another complex emotion in others only if they have similarities.

The Kaiser–Meyer–Olkin measure of sampling adequacy was .261, and Bartlett’s test of sphericity was significant (χ2 (5151) = 11,176.562, p < .001). Of the 102 correlation coefficients (34 items × 3 complex emotions), most of them were significant. All items correlated with at least eleven other items and communalities ranged between .03 to .3 indicating that the item shared some common variance with other items. Initial eigenvalues indicated that the first three factors explained 5.83%, 4.3 8%, and 3.99% of the variance, respectively. Further examination of the eigenvalues on the scree plot suggested a possible three factors solution. Those items with a primary load of .3 or above and a cross-load below .2 were only retained. This resulted into retention of only twenty of the thirty-four items. Tables 1 and 2 summarize the mean (SD) and the rotated component matrix, respectively.

Table 1 Descriptive statistics
Full size table
Table 2 Rotated component matrix
Full size table

As evident in Table 2, all twelve items measuring guilt had factor load ranging between .318 and .509. On the other hand, all the eight items measuring shame (two items) and remorse (six items) had factor load ranging between − .303 and − .450. The outcome indicated guilt and shame/remorse getting induced by same scenario depending on the positive–negative load. The positive score reflects guilt, while negative scores reflect shame and remorse, both, suggesting that same scenario can elicit one complex emotion in some, while in few others it can induce the other complex emotion.

Generalized Procrustes Analysis (GPA)

Having identified the items measuring guilt, shame, and remorse, we converted the scenarios into a storyboard. Each storyboard comprised three frames illustrating a given scenario (Fig. 1).

Fig. 1
figure 1

Storyboard of a scenario found to induce guilt or remorse

Full size image

To statistically calculate how precisely the illustrations represented the narrations, 14 judges were asked to rate the possibility of emotion evoked (shame, guilt, or remorse) for all the scenarios as well as their illustrations. However, GPA was performed on only those 13 scenarios which had robust factor load. The response was analysed to achieve a consensus configuration. GPA was chosen as it permits comparison of the proximity between the terms adopted by the 14 different judges to describe the invoked emotion. The added advantage is also the fact that it takes care of the scaling effects. Thus, the preference to use lower or higher end of the rating scale by different judges is taken care by transforming the data to obtain a consensus configuration. The GPA scaling transformation is summarized in the PANOVA table (Table 3).

Table 3 Procrustes Analysis of Variance (PANOVA) outcome
Full size table

Figure 2 illustrates the residuals by scenarios after the transformations. The figure indicates the smallest residual for scenarios 9 and 20, a little higher residual for scenario 30, a bit higher residual for scenarios 32 and 34. Scenarios 4, 8, 25, and 26 had almost similar residuals. It is interesting to see that all these scenarios measure guilt. These residuals indicate higher consensus among the judges for these scenarios measuring guilt. On the other hand, scenarios 26 (remorse) and 23 (shame) have very high residual, indicating lack of consensus among the judges.

Fig. 2
figure 2

Generalized procrustes analysis outcomes: Residuals by scenarios after the transformations (G guilt, S Shame, R Remorse)

Full size image

Figure 3a illustrates the residuals by configuration after the transformations and it clearly indicates that five judges have very high residuals suggesting that their rating did not match the consensus. Figure 3b illustrates scaling factors of the GPA transformations. Scaling factor lower than 1 indicates usage of wider scale than other judges, whereas factors higher than 1 indicate the reverse. It is apparent from the figure that six of the judges used wider scale than others, whereas the remaining eight judges used the rating scale as widely as the others.

Fig. 3
figure 3

Generalized procrustes analysis outcomes: a illustration of the residuals by configuration, b scaling factors of the GPA transformations

Full size image

Thereafter, a consensus test was performed to ensure if the consensus configuration is a true consensus by determining whether the observed Rc value is significantly higher than 95% of the results attained when permuting the data. The eigenvalues demonstrate the variability corresponding to each axis. 79.786% of the variability is represented on the first axis, and the remaining 20.214% on the second axis. This variability, when split between the judges, shows almost identical result across the judges. The correlations between dimensions in the initial consensus configuration and the factors were 0.892 (F1) and 0.102 (F2) for scenario (narrative) and 0.964 (F1) and − 0.001 (F2) for scenario (illustrations). Figure 4a illustrates these dimensions on axes F1 and F2, and Fig. 6 illustrates the biplot on the two axes.

Fig. 4
figure 4

Generalized procrustes analysis: a illustration of the correlations between dimensions in the initial consensus configuration, and b biplot on the two axes

Full size image

Multidimensional Scaling (MDS)

Although both MDS and PCA are not performed on the same data as the previous analyses similarities/dissimilarities while the latter considers similarities between the items (here scenarios), we performed MDS to understand the rated similarity or dissimilarity among the items scores depicting the three complex emotions. The summary statistics is presented in Table 4. The results indicated decrease in Kruskal’s stress as the iterations proceed. The iterations continue till the point when there is no more improvement in stress. The observed Kruskal’s stress is 0.367 with 330 iterations for two-dimensional solution, 0.269 with 262 iterations for three-dimensional solution, and 0.215 with 419 iterations for four-dimensional solution. The decrease in stress value indicates improvement. Traditionally, the value of .10 or less is considered a good fit for two dimensions and for three dimensions the stress may be .07 or so. We did not get the stress value close to .10 but it was close enough.

Table 4 Summary statistics of MDS for 135 observations
Full size table

The scree plot shows stress plotted against the number of dimensions. Although it is assumed that stress decreases with increasing number of dimensions, there is a trade-off. Figure 5 shows that there is an elbow at three dimensions indicating that a two-dimensional solution may be adequate. However, three or four dimensions might improve the representation of the dissimilarities between the emotions. The stress for three dimensions is 0.26 which falls to 0.21 at four-dimensional solution.

Fig. 5
figure 5

Multidimensional scaling (MDS) scree plot

Full size image

To interpret the findings more clearly, we examined the clusters and dimensions in the MDS picture. Figure 6 illustrates an interesting map of perceived similarities among guilt, shame, and remorse invoking scenarios. Spread over two dimensions, all guilt scenarios (except 9 and 33) were clustered close to each other. Ten of them (scenarios 4, 8, 20, 23, 25, 26, 29, 30, 32, and 34) clustered in the first zone (positive on dimension one and negative on dimension two). The remaining two (scenarios 9 and 33) were in the second zone (positive on both the dimensions). Three of the remorseful scenarios (4, 10, and 33) were in the third zone (negative on both the dimensions), while the remaining three remorseful scenarios (8, 26, and 30) were in the fourth zone (negative on dimension one and positive on dimension two). Further, both the shame inducing scenarios (9 and 23) were close to remorse inducing scenarios. Thus, it can be interpreted that perception of shame and remorse overlaps and this is distinct from guilt. An analysis of the content of these scenarios suggests expectations and significant others as two possible dimensions. While unmet expectations and underachieved expectations are the two extremes of the expectation dimension, reparation and lack of reciprocity are the two extremes of the dimension ‘significant others’.

Fig. 6
figure 6

Multidimensional scaling (MDS) outcome: a configurations on the 2-dimensional representation space, b modified 2-dimensional representation space with only scenarios with robust factor loading

Full size image

In a traditional experiment, random assignment of variable is desirable for manipulation check. However, our aim was to explore the observed similarity/dissimilarity between guilt, shame, and remorse, and see whether same scenario can result to two or more of these three complex emotions or not. Therefore, all 34 scenarios were presented to all the participants in a random order. The estimated population of the senior undergraduates in the institute (where the study was conducted) falling within this age range would be approximately 800. Given the sample size of 138 the margin of error increases by 2.59% (from 5 to 7.59% assuming 95% confidence level and 50% response distribution). Further, the standard deviation of both age and years of education was upheld to minimum.

Discussion

The series of analyses endorse that same scenario can induce shame or guilt in some and guilt or remorse in others. It also helped in mapping the observed similarity/dissimilarity between guilt, shame, and remorse. Of the 13 scenarios, 12 assessed guilt, two assessed shame, and six assessed remorse. GPA results endorse the equivalence of narration of the scenarios and their corresponding illustrations, thus enabling the researchers to take care of issues arising out of verbal-linguistic format of the scenarios and use only illustrations in the final phase of the study. By adopting this method, the social desirability could also be mitigated. Of significance is the finding that guilt items have positive load, while shame and remorse have negative load on the same scenarios. Our study clearly suggests the intertwined nature of guilt, shame, and remorse in as much as they can be induced by the same scenario. They get induced depending upon the nature of processing; positively oriented processing results to guilt, while negatively oriented processing results in either shame or remorse. It is evident that rather than the situation the appraisal directs us towards complex social emotions. Siemer, Mauss, and Gross (2007) have also reported that a particular situation might elicit different emotional reactions depending on its appraisal.

Our findings are also in line with the findings of Tangney (1992) who focused on the ambiguity between the shame and guilt inducing situations highlighting that most of the situations can engender both. Other researchers have also endorsed that the same event might elicit both guilt and remorse (Lewis, 1971; Taylor, 1996). Even the indigenous construct of lajja synthesizes both guilt and shame (Bhawuk, 2017). Further, remorse has been interpreted as “an offence-specific version of guilt” (Fisher & Exline, 2010). These “moral emotions provide the motivational force—the power and energy—to do good and to avoid doing bad” (Tangney, 2003, p. 386).

Proeve and Tudor (2016) have referred to three conceptions of wrongness, the ones done to others, the ones felt as disobedience to an authority, and the ones that fall short of one’s own standards. They suggest that the retractive emotional state of remorse, guilt, or shame is based on the appreciative wrongness and a “person may quite coherently believe all three at once in relation to the same episode” (p. 32). The findings of this study suggest the significance of reparation in experiencing guilt. Irrespective of the nature of expectations, whether underachieved or unmet, the possibility of amend(s) results into guilt. On the contrary, the absence of scope of restitution results to shame and remorse. Unmet expectations along with lack of reciprocity might lead to the feeling of shame or remorse. On the other hand, the combination of underachieved expectations and lack of reciprocity results only to remorse.

Guilt and shame have been classified as self-conscious emotions which precipitates as a result of evaluation by significant others. Their primary function is adherence to social norms. In collectivist cultures, such as India, different phases of life are associated with some doable activities as well as some activities that should be avoided. As explicated by Bhawuk (2017), “lajjA is a gentle transient emotion but doSa is a fault or sin calling for prAyazcitta or penance” (p. 6). It is this reparation that might connect it to remorse. LajjA is also accepted as “the internal governor that guides one not only in not doing what is inappropriate, but also in doing what is appropriate” (Bhawuk, 2017, p. 21). This better explains why same situation invokes positive appraisal in some and negative in few others.

Although this study has revealed that the same event can elicit any of the three emotions, there is need to adequately differentiate the three. While many researchers have extensively focused on distinguishing between guilt and shame, very few works have attempted differentiating guilt from remorse. To add to that many a times guilt and remorse have been used interchangeably. Hence, there is a need for an empirically drawn line of distinction between these concepts and this study has catered to this need. This study significantly contributes in two ways. Besides showing the observed similarity/dissimilarity between guilt, shame, and remorse, it establishes that the same scenario can result into two or more of these three complex emotions. In contrast to the established Western research finding differentiating shame and guilt, it suggests that the three emotions differ in terms of positive–negative direction of processing. Even the Indian studies which banks on the connotation of Lajja as a positive emotion does not provide any measure to distinguish between Lajja and corollary of such scenarios. The expectation-reparation-reciprocity dimensions of this study help delineate this. Its merits mention that the participants were chosen from an institute of national importance with Pan-India representation of students. A random selection of participants from such an institution ensures both diversity and inclusiveness. The variation in the unyielding or slack social code of conduct, if any, in a diverse culture such as India, and especially those relating moral transgressions is likely to be accommodated due to participant’s diverse cultural background. As mentioned earlier, aparaadhbodh (guilt) and pashchaataap (remorse) have not been empirically studied in the Indian context and this study fills this gap to some extent. They have been used only as an empirical approach to map semblance among self-rated sense of Lajja (shame), aparaadhbodh (guilt) and pashchaataap (remorse). We also acknowledge that this study is confined to young population and thus the findings might not be generalized across the board.