One of the strengths of the paper by Kalisch et al. on a framework for the neurobiology of resilience is the way in which it focuses on general resilience mechanisms that may be appropriate across different challenges and protect from stress-related mental illness. The authors' suggestion that the influences of diverse resilience factors on good outcomes are mediated through a limited number of resilience mechanisms is intriguing. The model they propose potentially lends itself to testing beyond the neurobiological community, and we would welcome the opportunity to explore the strengths and limitations of broadening the theory.

Kalisch et al. suggest that numerous factors have been identified that influence resilience. In relation to later life, we suggest that relatively few factors have been identified (Windle Reference Windle2011; Windle & Bennett Reference Windle, Bennett and Ungar2011). This is in part because research on resilience in later life has received far less attention than at the other end of the age spectrum (Windle Reference Windle2012). We would argue, therefore, that examining whether or not this may offer an advantage to further understanding resilience mechanisms is a worthy extension of the authors' theory.

Kalisch et al. focus on an intraindividual mechanism. They comment that they regard socioenvironmental factors as distant influences, although they acknowledge that these influences are well documented. And in speaking of these as “influences,” Kalisch et al. distance them from the resilience factors that they discuss, suggesting that the former are not central. The key factor that Kalisch et al. focus on is appraisal, an intraindividual factor. We would strongly argue, however, that these interpersonal and socioenvironmental influences are instead important and proximal factors in resilience (Bennett Reference Bennett2010; Windle Reference Windle2012).

We propose an ecological model of resilience that explains resilience more fully in older adults. In our formulation, we refer to these factors as resources. We identify these resources at three, nonhierarchical, interacting levels, which we call “individual,” “community,” and “societal” (Windle Reference Windle2012; Windle & Bennett Reference Windle, Bennett and Ungar2011) (see Fig. 1). At the individual level, we identify the following resources: psychological, material, biological, health behaviours, age, and gender. Appraisal, as outlined by Kalisch et al., can be seen as a psychological resource. At the community level, we believe that social support, social cohesion and participation, and housing are important. Finally, at the societal level, resources such as social, health, and welfare services; social policy; the economy, and neighbourhood all have the potential to contribute to resilience.

Figure 1. The Resilience Framework (adapted from Windle & Bennett Reference Windle, Bennett and Ungar2011). The figure shows the relationships between the antecedents of resilience and the presence (or absence) of resources at the individual, community, and societal levels and their consequences for resilience.

Taking one aspect of Kalisch et al.'s argument, we can illustrate how the ecological model can effectively explain resilience in older adults. If one takes Kalisch et al.'s own example of social support, we argue that this is a community-level resource. Although, their theory reduces it to a mediator of appraisal, it remains an important component of resilience. When one looks at the literature both on stress and on resilience, social support remains a significant factor (Bennett Reference Bennett2010; Fuller-Iglesias et al. Reference Fuller-Iglesias, Sellars and Antonucci2008; Netuveli et al. Reference Netuveli, Wiggins, Montgomery, Hildon and Blane2008; Wells Reference Wells2010). This remains the case when aspects of social support – for example, its availability, quality, and quantity – are examined. We would argue that social support has a greater influence than its mere appraisal. One can extend that argument to other resources such as social participation, social and health service provision, and culture, including religion (Lee et al. Reference Lee, Brown, Mitchell and Schiraldi2008; Seddon et al. Reference Seddon, Robinson, Reeves, Tommis, Woods and Russell2006; Ungar Reference Ungar2011). We do not believe that the majority of people can be resilient without these resources, and indeed, the utilisation of these resources.

Drawing on our own work with older adults in the field of resilience, we recognise the broader antecedents, processes, and outcomes of resilience. Whereas intraindividual factors such as mastery (Windle & Woods Reference Windle and Woods2004), personal control, competence, and self-esteem are important in older adults for what has been described as psychological resilience (Windle et al. Reference Windle, Markland and Woods2008), so are community and societal factors, as well (Windle & Bennett Reference Windle, Bennett and Ungar2011). Bennett (Reference Bennett2010) found that amongst older widowed men, social support, both formal and informal, was a key factor in determining who became resilient and who did not. Similarly, in work with spousal carers of people with dementia, Donnellan et al. (Reference Donnellan, Bennett and Soulsby2014) found that social support from friends (as opposed to family) was a key factor in determining whether a carer became resilient or not. Societal factors such as social policy, culture, and religion also have been identified as important contributors. For example, Donnellan et al. (Reference Donnellan, Bennett and Soulsby2014) found the provision of support groups by the third sector and local government was an important factor in resilience, as was the availability of respite services, as well.

Kalisch et al.'s formulation suggests that fundamentally, resilience lies within the individual and that the individual is required to act in achieving resilience by appraising situations in a particular way. We would argue, however, that the acquisition of resilience does not necessarily require the action of the individual; that is, the individual may be at least initially passive. We would argue that in some circumstances, resilience could be facilitated by external agency (Bennett Reference Bennett2010). For example, in the case of some older widowed men, it required external agency and the involvement of neighbours, strangers, and health and welfare professionals to provide the first steps towards a resilient outcome (Bennett Reference Bennett2010). These widowers would not have been able to achieve resilience alone, and certainly not by appraisal alone (Bennett Reference Bennett2010). Hence, we would argue that the environment/context is an important aspect of positive development and subsequent resilience. It follows, then, that individual-level factors (e.g., traits and characteristics), whilst important, are strongly influenced by the content and function of these environments. Any subsequent appraisal processes are then determined by the interaction between the individual and that individual's environment.