Kalisch et al. offer a novel framework for resilience research, wherein appraisal is regarded as the “global” resilience mechanism par excellence, serving as the “proximal cause” for mental health. I applaud their transdiagnostic approach and the focus on resilience mechanisms over resilience factors.
By their own admission, however, the authors make some radical claims, and sometimes they go too far. The notion that a “positive appraisal style” mediates all other resilience factors is a bold one, but it cannot be true. For example, genetic factors may influence the stamina of an individual, which in turn may affect cortisol levels in the body. This is not mediated by appraisal. Similarly, Kalisch et al. claim that there is no other route between an event and an emotion than appraisal, but it is possible to arouse emotions through pharmacological agents (Izard Reference Izard1993). This does not detract, though, from the fact that most emotions are produced by some kind of stimulus evaluation.
Here the authors make another radical claim, which is the primary focus of this commentary: “that there is a single mediating mechanism” (Figure 2B). I will argue that resilience researchers need to consider a range of mechanisms for stress and emotion in a more process-specific manner, because each of the mechanisms influences emotions and stress in a different manner. Kalisch et al. seem aware that the evaluation of potential stressors can involve many different cognitive processes (sect. 4.3.5). Yet they do not appear to realize the consequences. The noted confusion about whether neural measures show appraisal content or process reflects the fact that studies have neglected specific mechanisms. Hence, LeDoux (Reference LeDoux1995) suggested that we should “abandon discussion of cognitive-emotional interactions in terms of vague monolithic cognitive processes, and instead consider exactly which cognitive processes are involved” (p. 224).
A recent framework along these lines aimed at explaining musical emotions, but with much wider applicability, might serve to illustrate this approach. The framework posits nine mechanisms, which range from simple reflexes to complex judgments and which are implemented by several, partially overlapping brain networks that developed in an evolutionary progression: BRECVEMAC – brain stem reflex, rhythmic entrainment, evaluative conditioning, contagion, visual imagery, episodic memory, musical expectancy, aesthetic judgment, and cognitive appraisal (Juslin Reference Juslin2013; see also Juslin & Västfjäll Reference Juslin and Västfjäll2008).
One implication of this framework is that because several of the mechanisms are “implicit” in nature and may occur in parallel, researchers cannot rely only on phenomenological report to explain how events cause emotions. Moreover, conflicting outputs from mechanisms that use different “data” to evoke emotions may explain the occurrence of “mixed emotions”.
Much of the discussion in the target article revolves around various appraisal dimensions, yet many of the mechanisms do not operate in terms of dimensions. Whether appraisal dimension is the right unit of analysis depends on the mental representation involved in each mechanism (see Table 2 in Juslin Reference Juslin2013). If appraisal dimensions are applied to emotion episodes aroused by mechanisms other than “cognitive appraisal,” the dimensions lack grounding in reality, and they take on a purely “metaphorical” status. This can explain the occasionally poor correspondence between specific emotions and predicted appraisal patterns (Frijda & Zeelenberg Reference Frijda, Zeelenberg, Scherer, Schorr and Johnstone2001).
A multimechanism framework has important consequences for the question of how appraisal style can be operationalized and measured. Empirically determining which emotion-induction mechanism is responsible for a stress response, and which information produced the outcome, is difficult. Current evidence of cognitive appraisal comes mostly from post hoc verbal reports regarding appraisal dimensions, but these do not warrant causal conclusions, because subjects may infer the “appropriate” appraisal pattern from the affective state (e.g., fear, stress) they are experiencing. Because verbal report may not be able to distinguish between various mechanisms, we simply do not know the precise proportion of instances in everyday life when emotions are caused by, say, conditioning, as opposed to a cognitive appraisal. To draw causal conclusions, we have to manipulate mechanisms under controlled conditions.
A multimechanism framework also has some implications for the development of new and better prevention strategies: It does not make sense to subsume different negative appraisal patterns under an umbrella term of “negative appraisal bias.” The same person could have a different “appraisal style” for each emotion-mechanism. To measure all of these would be a challenge indeed! It also means that a specific intervention (e.g., active reappraisal) will be ineffective if it targets the wrong psychological mechanism.
In addition, how does one decide that an appraisal style is “overly negative”? Is the resilience outcome (e.g., subjective well-being) the criterion? If so, the test is circular. Or is the style to be evaluated in relation to “objective” circumstances of the individual? If so, “realism” is key, rather than mere “aversion reduction.” Positive appraisals of potentially aversive events could indeed arouse less stress now, but might lead to negative consequences and hence more stress later. One could argue that, more important than the original appraisal of an aversive event, is how, if at all, the ensuing emotion leads to adaptive behavior to resolve the problem. In some situations, resilience as an outcome might be more effectively achieved by changing external life factors of the individual than by trying to change the appraisal styles at multiple levels of cognitive functioning. It remains to be seen which of these – appraisal style or external factors – will explain most variance in resilience outcomes.
My claim – that appraisal cannot be meaningfully addressed at a global level, disregarding the specific mechanisms involved – does not undermine the overall validity of the novel approach, but shows that researchers must make choice: Either they deal with the mechanisms and their characteristics at a “global” appraisal level, in terms of “appraisal dimensions” that might have little reality and lack the required precision to develop effective interventions; or they address appraisal at the level of specific mechanisms – even though this adds considerable complexity to the whole endeavor. The latter alternative is clearly more in line with the kind of ambitious approach advocated by Kalisch et al.
Kalisch et al. offer a novel framework for resilience research, wherein appraisal is regarded as the “global” resilience mechanism par excellence, serving as the “proximal cause” for mental health. I applaud their transdiagnostic approach and the focus on resilience mechanisms over resilience factors.
By their own admission, however, the authors make some radical claims, and sometimes they go too far. The notion that a “positive appraisal style” mediates all other resilience factors is a bold one, but it cannot be true. For example, genetic factors may influence the stamina of an individual, which in turn may affect cortisol levels in the body. This is not mediated by appraisal. Similarly, Kalisch et al. claim that there is no other route between an event and an emotion than appraisal, but it is possible to arouse emotions through pharmacological agents (Izard Reference Izard1993). This does not detract, though, from the fact that most emotions are produced by some kind of stimulus evaluation.
Here the authors make another radical claim, which is the primary focus of this commentary: “that there is a single mediating mechanism” (Figure 2B). I will argue that resilience researchers need to consider a range of mechanisms for stress and emotion in a more process-specific manner, because each of the mechanisms influences emotions and stress in a different manner. Kalisch et al. seem aware that the evaluation of potential stressors can involve many different cognitive processes (sect. 4.3.5). Yet they do not appear to realize the consequences. The noted confusion about whether neural measures show appraisal content or process reflects the fact that studies have neglected specific mechanisms. Hence, LeDoux (Reference LeDoux1995) suggested that we should “abandon discussion of cognitive-emotional interactions in terms of vague monolithic cognitive processes, and instead consider exactly which cognitive processes are involved” (p. 224).
A recent framework along these lines aimed at explaining musical emotions, but with much wider applicability, might serve to illustrate this approach. The framework posits nine mechanisms, which range from simple reflexes to complex judgments and which are implemented by several, partially overlapping brain networks that developed in an evolutionary progression: BRECVEMAC – brain stem reflex, rhythmic entrainment, evaluative conditioning, contagion, visual imagery, episodic memory, musical expectancy, aesthetic judgment, and cognitive appraisal (Juslin Reference Juslin2013; see also Juslin & Västfjäll Reference Juslin and Västfjäll2008).
One implication of this framework is that because several of the mechanisms are “implicit” in nature and may occur in parallel, researchers cannot rely only on phenomenological report to explain how events cause emotions. Moreover, conflicting outputs from mechanisms that use different “data” to evoke emotions may explain the occurrence of “mixed emotions”.
Much of the discussion in the target article revolves around various appraisal dimensions, yet many of the mechanisms do not operate in terms of dimensions. Whether appraisal dimension is the right unit of analysis depends on the mental representation involved in each mechanism (see Table 2 in Juslin Reference Juslin2013). If appraisal dimensions are applied to emotion episodes aroused by mechanisms other than “cognitive appraisal,” the dimensions lack grounding in reality, and they take on a purely “metaphorical” status. This can explain the occasionally poor correspondence between specific emotions and predicted appraisal patterns (Frijda & Zeelenberg Reference Frijda, Zeelenberg, Scherer, Schorr and Johnstone2001).
A multimechanism framework has important consequences for the question of how appraisal style can be operationalized and measured. Empirically determining which emotion-induction mechanism is responsible for a stress response, and which information produced the outcome, is difficult. Current evidence of cognitive appraisal comes mostly from post hoc verbal reports regarding appraisal dimensions, but these do not warrant causal conclusions, because subjects may infer the “appropriate” appraisal pattern from the affective state (e.g., fear, stress) they are experiencing. Because verbal report may not be able to distinguish between various mechanisms, we simply do not know the precise proportion of instances in everyday life when emotions are caused by, say, conditioning, as opposed to a cognitive appraisal. To draw causal conclusions, we have to manipulate mechanisms under controlled conditions.
A multimechanism framework also has some implications for the development of new and better prevention strategies: It does not make sense to subsume different negative appraisal patterns under an umbrella term of “negative appraisal bias.” The same person could have a different “appraisal style” for each emotion-mechanism. To measure all of these would be a challenge indeed! It also means that a specific intervention (e.g., active reappraisal) will be ineffective if it targets the wrong psychological mechanism.
In addition, how does one decide that an appraisal style is “overly negative”? Is the resilience outcome (e.g., subjective well-being) the criterion? If so, the test is circular. Or is the style to be evaluated in relation to “objective” circumstances of the individual? If so, “realism” is key, rather than mere “aversion reduction.” Positive appraisals of potentially aversive events could indeed arouse less stress now, but might lead to negative consequences and hence more stress later. One could argue that, more important than the original appraisal of an aversive event, is how, if at all, the ensuing emotion leads to adaptive behavior to resolve the problem. In some situations, resilience as an outcome might be more effectively achieved by changing external life factors of the individual than by trying to change the appraisal styles at multiple levels of cognitive functioning. It remains to be seen which of these – appraisal style or external factors – will explain most variance in resilience outcomes.
My claim – that appraisal cannot be meaningfully addressed at a global level, disregarding the specific mechanisms involved – does not undermine the overall validity of the novel approach, but shows that researchers must make choice: Either they deal with the mechanisms and their characteristics at a “global” appraisal level, in terms of “appraisal dimensions” that might have little reality and lack the required precision to develop effective interventions; or they address appraisal at the level of specific mechanisms – even though this adds considerable complexity to the whole endeavor. The latter alternative is clearly more in line with the kind of ambitious approach advocated by Kalisch et al.