Is Self–Other Overlap the Key to Understanding Empathy?

Abstract

Preston and Hofelich (2012) suggested that researchers disagree on the role of self–other overlap in empathy due to a failure to differentiate among neural overlap, subjective resonance, and personal distress; they also developed a framework for tying neural and subjective overlap to various aspects of functioning they include in the construct of empathy. Although we found their discussion of different processes that have been labeled empathy interesting and helpful, we found their discussion of self–other overlap to be somewhat less useful for conceptualizing differences among empathy-related processes. In addition, we provide an alternative perspective to their reasoning regarding the role of experience and neural overlap in an aspect of empathy-related responding (e.g., concern or compassion).

Preston and Hofelich argue that “there is still disagreement in the field, most of which stems from problems associated with the term ‘self–other overlap’… which occurs at both neural and subjective levels” (2012, p. 24). They note that “subjective overlap (i.e., a consciously experienced resonance that observers can notice, feel, and reflect upon) is the form typically discussed in psychology…” (2012, p. 24). Although we found the discussion of when empathy occurs or does not occur interesting and useful, we do not believe that the self–other overlap construct provides as much traction as is suggested by Preston and Hofelich and we discuss our reasons below. In addition, we present an alternative perspective on some of the thorny definitional issues that Preston and Hofelich tackle in their article.

First, we had difficulty understanding the authors’ definition of neural self–other overlap. The authors say that “At the neural level, the observation of another’s affective state can activate neural regions in the observer that are also activated when the observer directly experiences the state…” (Preston & Hofelich, 2012, p. 24), but also say that “Neural-level overlap occurs when the observer uses personal representations of experience to understand the target” (Preston & Hofelich, 2012, Table 1). These two definitions may involve different processes (e.g., mirror neurons that give rise to an understanding of emotion experience versus top–down cognitive processes such as perspective taking that are used to understand another’s emotional reactions), and therefore it is difficult to know precisely what is meant by neural self–other overlap. Moreover, based on the second mentioned statement regarding neural overlap, it is unclear whether affective experience is a necessary component of neural overlap or whether only cognitive understanding is required.

In addition, the authors define empathy as “Processes by which observers come to understand and/or feel the state of another…” (Preston & Hofelich, 2012, p. 25). This definition fails to make a distinction between shared understanding (i.e., perspective taking, which we take to be a cognitive rather than an affective process) and empathy, which we believe must also involve an affective component (Eisenberg, Shea, Carlo, & Knight, 1991). The lack of this distinction, despite the arguments for defining empathy broadly, makes it difficult to understand how neural and subjective self–other overlap differentially relate to empathy. If perspective taking is defined as a cognitive process, we would argue that only the affective component of empathy requires subjective overlap (which the authors described as involving “consciously experienced resonance”), whereas both perspective taking and the affective component of empathy probably often involve neural self–other overlap. This is an issue to be resolved by research, but it is quite possible that affective empathy and the cognitive process of perspective taking differ in the type of overlap that is typical or required. Perhaps we are misinterpreting the authors’ definition of perspective taking, but we were unsure. The authors said that “cognitive empathy is necessarily affective” (Preston & Hofelich, 2012, p. 29), but they defined cognitive empathy in Table 1 as “Understanding the other by engaging one’s representations through effortful top–down processes” (Preston & Hofelich, 2012, Table 1)—a definition that does not refer to affect and sounds purely cognitive. Consensus in the field on the definition of perspective taking is needed for consensus on its role in empathy and its neural bases.

We would also suggest that the authors’ view of the self–other overlap construct is not always consistent with that of others in the field. For example, Preston and Hofelich asserted that “those who exclude self–other overlap from empathy construe it as a distressing, subjectively felt state that undermines the argument for other-oriented aid” (2012, p. 26) and cite Batson’s work in this regard. However, Batson (C. D. Batson, personal communication, February 9, 2011) does not believe he has said this; the authors seem to confuse Batson’s construct of personal distress with their discussion of self–other overlap. We had difficulty understanding the relation of personal distress to the authors’ constructs of empathy and self–other overlap. In Table 1, emotional contagion (which we assume from other text involves some self–other overlap) is stated to overlap with the construct of personal distress; however, the authors also state that distress or self-focus is independent of self–other overlap per se (Preston & Hofelich, 2012). We have argued that empathic overarousal can lead to personal distress, but that it can also stem from cognitive processes incited by exposure to upsetting emotions or events (Eisenberg et al., 1994; Eisenberg et al., 1991). The field would benefit from more discussion of the relations among self–other overlap, personal distress, and emotional contagion; Preston and Hofelich’s article provides a starting point.

In our opinion, another factor that diminishes the usefulness of the neural self–other overlap construct is that it is viewed by the authors as involved in nearly all instances of empathy, broadly defined, except inaccurate empathy (or no empathy; Preston & Hofelich, 2012, Table 2). Thus, the construct of neural overlap is not very useful for differentiating various aspects of responding related to empathy.

In discussing the construct of empathy, the authors said: “Without a similar prior experience, it is presumed that the observer does not have the necessary representations to process and understand the target’s state, let alone share in their experience in a meaningful way” (Preston & Hofelich, 2012, p. 26); and “Of course, an observer can use top–down processes to derive an internal image for the scene that visually simulates the experience, but the observer cannot really understand, relate, or empathize without a representation for how it feels to be insulted or rejected from past experience” (Preston & Hofelich, 2012, p. 27). Coming from a perspective that differentiates sympathy and empathy, we do not agree with the assumption that a person requires “similar prior experience” to feel sympathy. The authors seem to include sympathy (“compassionate concern, feeling ‘sorry for’ the other”) and true empathy (“a compassionate, other-oriented state that requires a distinction in the observer between self and other”) in the broader construct of empathy (Preston & Hofelich, 2012, Table 1). In our opinion, people can experience sympathy/compassion for people in very different contexts from those they have experienced themselves. For example, one can feel sympathy for a person in the aftermath of war or an earthquake without ever having experienced such a disaster. People learn about others’ situations from symbolic and visual media without experiencing a similar situation; such information can be accessed to generate concern. In some contexts, a person may even feel concern for another person who is unaware of pending problems or current threats. For example, an observer may feel acute sympathy and emotions like sadness for a child who is about to lose a parent, even if the child has no knowledge of the pending death and is quite content at the moment.

Similarly, Preston and Hofelich argued that “neural-level overlap is necessary but not sufficient for true empathy” (2012, p. 27). They defined “true empathy” as involving compassion (2012, Table 1). However, we believe it is possible that people can experience sympathy or compassion without much or any real overlap in representations (self–other overlap is defined as involving “the observer’s personal representations for experiencing the observed state”; 2012, p. 24). For example, using our prior example, based on retrieving stored cognitions about death and its consequences, an observer could feel considerable compassion for a child who has lost a parent, even if the observer has not directly experienced the loss of a family member.

In summary, Preston and Hofelich do the field a service by trying to disentangle a number of thorny issues. The difficulty in doing so partly reflects the inherent difficulty in considering phenomena as diverse as conceptual overlap, emotional contagion, and sympathy under a single rubric and mapping of cognitive and emotional processes onto an array of complex processes.