R1.1. The BCD should not incorporate all behaviours associated with SES, and SES-behaviour associations may vary with context

In our target article, we presented a range of behaviours that appear to be somewhat consistently socioeconomically patterned, with their defining common feature being that they can all be conceived of as a response to being faced with trade-offs between present and future outcomes (our BCD). The BCD is not intended to incorporate all behaviours that vary with SES, nor is the contextually appropriate response perspective, presented in relation to the BCD behaviours, intended to explain all socioeconomic differences in behaviour. We simply offered a possible means of understanding this particular cluster of behaviours (the BCD behaviours listed in sect. 2 of the target article) as a contextually appropriate response to a common explanatory factor – the strength of the present-future trade-offs that people face.

We also did not mean to suggest that all variation in BCD behaviours can be explained by these present-future trade-offs. It is unlikely that any phenomenon, particularly one as complex as patterns of human behaviour, can be pinned to a single explanatory factor. We simply propose that a meaningful proportion of the variation in BCD behaviours should be attributable, via appropriate psychological pathways, to individuals' experiences of future-limiting factors. (This does not mean that we do not acknowledge that some other meaningful portion of variation in BCD behaviours might be explained by other factors, such as the experiences of relative disadvantage discussed by Novakowski & Mishra).

In response to our target article, Breugelmans, Plantinga, Zeelenberg, Poluektova, & Efremova (Breugelmans et al.) argue that the BCD “(1) overestimates the coherence of the various behaviors associated with poverty and (2) underrepresents the range of behaviors that should be included in such a constellation” (para. 1). This would certainly be true, had we intended to argue that all socioeconomically patterned behaviours should be included in the BCD. Perhaps the response of Breugelmans et al. indicates that our chosen title, “the behavioural constellation of deprivation,” is a little aggrandised. Nonetheless, we believe that this label is more memorable and relatable than a more-accurate but less-wieldy title, such as “a cluster of often socioeconomically patterned behaviours united by their connection to temporal trade-offs.”

Several commentaries raise the issue of socioeconomic variation in social and cooperative behaviour, either assuming it to be part of the BCD or arguing for its inclusion. For example, Grossmann & Varnum point out that lower SES is associated with greater accuracy in identifying others' emotions and in compassion towards others – perhaps because in-group coordination can help buffer against scarcity and external threats. Robinson & Piff argue that increased prosocial behaviour among lower-SES individuals may serve to provide increased control in the context of more-threatening social environments. Srivastava & Srinivasan highlight the trend towards more charitable and prosocial behaviour in lower-SES individuals, suggesting that our target article ought to have put more emphasis on the existence of such positive behaviours among people of lower SES. In response to these commentaries, we wish to emphasise that we are open to the extension of the BCD to incorporate social and cooperative behaviours that vary with SES. However, such extension would require demonstrating that these behaviours share the same underlying logic as the BCD behaviours we outlined in the target article. If they do not, they may well be socially patterned but not usefully viewed as part of the BCD as we define it (see also section R4.1).

Pearson & van der Linden discuss the specific example of pro-environmental concern – something they say is more common among those of lower SES than of higher SES. They rightly point out that environmental behaviours entail an intertemporal trade-off. However, these behaviours also involve other dimensions, such as common goods problems, and the harms and the costs of mitigation fall differentially on different social groups. As such, the proportion of variation in environmental behaviour that could be explained by exposure to future-limiting factors may be somewhat limited. Thus, as highlighted by these commentaries, to expand the BCD to incorporate a broader range of socioeconomically patterned behaviours, other explanatory factors, such as social context, would need to be taken into account in our explanation. We prefer in this instance to put forward a simple causal account of a narrow range of behaviours rather than a complete but multifactorial treatment of all SES-patterned behaviours.

Just as we did not mean to imply that the BCD should include all socioeconomically patterned behaviours, we did not intend to suggest that the elements of the BCD should be apparent among all lower-SES individuals or in all contexts. In our target article, we attempted to dispel such ideas with section 8.1, “The BCD only applies on average.” However, we may not have made our point sufficiently clear: There will be situations in which there are limited or no SES differences in experiences of the intertemporal trade-offs that we argue drive the BCD. That is, if the BCD is driven by trade-offs that tend to be more common at lower SES, rather than by SES per se, then the association between SES and those trade-offs is required for the existence of any relationship between SES and BCD behaviours.

If the BCD is not driven by SES per se, developing hypotheses about which specific facets of lower- or higher-SES life affect the BCD behaviours we are interested in can lead to more precise measures, allowing us to account for more of the variation in those behaviours. For example, our contextually appropriate response perspective would lead us to predict that exposure to extrinsic (uncontrollable) mortality risk (which tends to be greater at lower SES) should be a better predictor of health behaviour than SES itself (which is associated with, but does not necessarily directly influence, health behaviour). Indeed, in a survey of North American adults, we found that those of lower subjective SES reported making less effort to look after their health and perceived a greater portion of their mortality risk to be extrinsic. This relationship between SES and self-reported health behaviour was entirely mediated by respondents' perceived extrinsic mortality risk. That is, SES was no longer a predictor of health behaviour once perceived extrinsic mortality risk was controlled for (Pepper & Nettle Reference Pepper and Nettle2014b). Although this finding requires further replication and elucidation, it suggests that, in some cases, perceived extrinsic mortality risk may provide a more precise predictor or control measure for studies of health behaviour than measures of SES. Our target article and its commentaries list various other future-limiting factors that, like perceived extrinsic mortality risk, have the potential to account for more of the variation in BCD behaviours than measures of SES itself.

Some BCD behaviours may be more strongly influenced by other factors at times, making explanations regarding intertemporal trade-offs less applicable. For example, Daly Ramos, & Perry point out that, although research from wealthier countries supports the existence of the BCD as we define it, the patterns we have described do not seem to generalise to Latin American populations. They state that research in Argentina and Mexico shows the use of alcohol, drugs, and tobacco to be more common among those of higher SES – perhaps because poverty in such countries is more extreme, and those at the bottom of the SES ladder lack the financial means to use such substances. Such examples highlight the need to consider context: Moderating and limiting factors will apply.

R1.2. We didn't set out to list all of the proximate mechanisms that might underlie the BCD

Our commentators have done an admirable job of adding richness to the picture regarding potential proximate mechanisms of the BCD. They have discussed potential proximate psychological mechanisms, such as optimism and pessimism (Mittal & Griskevicius), envy (Novakowski & Mishra), and self-control (Carmel & Leiser), as well as physiological mechanisms, such as differential immunological investment (Garcia & Blackwell) and DNA methylation (Brown & Olding). In the target article, we didn't attempt to list all possible proximate mechanisms that might underlie the BCD. We hoped that our commentators would add detail in this regard, and we are grateful to them for having done so.

Many of our commentators appear to have understood that our attempt to explain the BCD was largely at the ultimate rather than at the proximate level. That is, we aimed to explain the BCD in terms of the payoffs to BCD behaviours in the context of deprivation rather than in terms of such factors as self-control, which we view as psychological mechanisms underpinning BCD behaviours. In section 4 of the target article, we emphasised that proximate explanations do not preclude ultimate ones – rather, they are complementary, with ultimate explanations telling us why a behaviour should occur in a given context, and proximate ones telling us how that behaviour is delivered via cognitive or physiological mechanisms (Scott-Phillips et al. Reference Scott-Phillips, Dickins and West2011). Despite this effort, some commentators proffered proximate explanations for the BCD as though they were mutually exclusive of our ultimate explanation. For example, Doebel, Michaelson, & Munakata (Doebel et al.) suggest that we made a false dichotomy by contrasting personal control with self-control abilities. However, under the perspective we put forward, self-control is considered a proximate psychological mechanism underpinning the link between limited personal control (over future-limiting factors) and BCD behaviours. We cover this issue in more detail in R2.1, where we discuss whether the “alternative” explanations offered by commentators are truly separate accounts.

R1.3. We didn't intend to pretend that mortality is the beginning and the end

In section 2.3 of the target article, we focussed on the specific example of extrinsic mortality risk as a future-limiting factor driving the BCD. We did this not only because uncontrollable mortality risk is an extreme example of a future-limiting factor, but also because it has been extensively examined in pertinent evolutionary models. In section 2.4 of the target article, we acknowledged that many other future-limiting factors will be important, giving some examples. Nonetheless, such commentators as Uggla and Riis-Vestergaard & Haushofer object to our focus on extrinsic mortality risk as an example. Uggla highlights studies that show small effects of area-level extrinsic mortality in comparison to individual-level measures of SES. Meanwhile Riis-Vestergaard & Haushofer suggest that observed rates of temporal discounting across countries are too high to be accounted for by extrinsic mortality risk alone. Of course, they are most likely right: Although extrinsic mortality risk offers a convenient illustrative example that is easy to grasp, in reality, people living with deprivation face a multitude of subtle and varied future-limiting factors that are likely to have a cumulative effect. Indeed, Riis-Vestergaard & Haushofer offer some additional examples of such factors, including liquidity constraints, income uncertainty, and the unpredictability of the social environment. Other commentaries expand upon this list by discussing such influences as capital constraints (Rickard), low cultural consonance (Quinlan), and the limited control experienced as a result of operating close to budgetary boundaries (Srivastava & Srinivasan).

R2.1. “Alternative” hypotheses that are not mutually exclusive to ours

A number of our commentators proffered explanations for the BCD, which they either explicitly state or imply are alternative to ours. For example, Srivastava & Srinivasan agree that differences in temporal discounting may underlie a portion of observed socioeconomic differences in behaviour, but they suggest that extrinsic mortality risk is not sufficient to explain these differences. They instead posit that “our alternative hypothesis is that the reason low SES subjects demonstrate steeper time discounting is that they have greater experience with planning and control failures caused by always operating close to budgetary boundaries, which in turn arise inevitably from having to consistently operate with limited budgets of money, status, trust, or other forms of social utility” (para. 5; emphasis added). We agree that extrinsic mortality risk cannot be sufficient to explain all socioeconomic variation in behaviours related to temporal discounting. Indeed, we emphasise in section R1.3 that we did not intend to suggest this.

We also agree that people of lower SES are more likely to experience what Srivastava & Srinivasan term “planning and control failures” (para. 5) that affect their temporal discounting. Indeed, we discussed this in section 2.2 of our target article, which outlines the ways in which having limited wealth, education, or social connections can curtail personal control and thereby the ability to influence future outcomes. We used limited control over mortality risk as an example in the target article but also included a section on other factors (2.4). To this section, we could easily add the explanation given by Srivastava & Srinivasan: “Low SES individuals will naturally find planning and controlling actions at long time scales inefficient, because plans with low resource reservoirs underpinning them are more susceptible to being overturned by small random socioeconomic fluctuations” (para. 5). This usefully illustrated another future-limiting factor experienced at lower SES – limited reserves increase vulnerability to socioeconomic perturbations. As such, this is not an alternative hypothesis, but a complementary one.

Some other commentators suggested that proximate mechanisms, such as self-control, are alternative explanations to the ultimate one we outlined in the target article. Doebel et al. suggest that “contrasting contextual factors with self-control may be a false dichotomy” (para. 4). This is not a dichotomy we intended to make, because we view self-control as a mechanism that calibrates behavioural responses to contextual factors. In the same vein, Carmel & Leiser “substitute a different psychological explanation and argue that self-control rather than ‘extrinsic mortality risk’ explains those behaviors” (para. 1). They argue that “short-sighted” decisions can be “explained by the limited attention span brought about by financial scarcity” (para. 2). However, in our target article (sect. 4), and in section R1.2 of this response, we emphasise that we view individual psychological mechanisms, such as self-control, as proximate pathways delivering the BCD in response to its ultimate cause – lack of control over future outcomes. As such, we do not view explanations invoking such concepts as self-control as being alternative to our explanation. These explanations work together at different levels, ours being ultimate, theirs being proximate.

As we stated in the target article, viewing these proximate and ultimate explanations as part of the same picture alters our perspective regarding such concepts as self-control: Rather than being a mysterious internal resource that sometimes fails us due to cognitive constraints, such as limited attention span, self-control can be viewed as a psychological mechanism guiding our behavioural response to the trade-offs we face. This is the essence of the contextually appropriate response perspective. It is important, because it suggests that rather than pejoratively viewing present-oriented decisions as resulting from a failure to muster sufficient cognitive resources to be future-oriented, we might view them as appropriate decisions for that person given their circumstances.

In addition to offering putatively alternative explanations that conflate levels of explanation, some commentators mistakenly suggest that evidence in support of phenomena not explained in our target article can be taken as evidence against the perspective we outlined. For example, Rad & Ginges present their finding that Muslim participants who were fasting for Ramadan became more risk and loss averse than those who were not. Because fasting entails the experience of self-induced scarcity (it is within individual control), they propose that it should not alter decision making if low control does elicit BCD behaviours. Their finding is interesting for a number of reasons, but it does not undermine our hypothesis that low control drives the BCD: Evidence that people become more risk averse under self-induced scarcity is not tantamount to evidence that people do not become more risk averse under involuntary scarcity. Moreover, there are many cases in which humans activate through voluntarily action mechanisms that evolved to respond to involuntary experiences. Self-imposed dieting makes people hungry, and self-imposed viewing of horror movies makes them scared. This hardly refutes the idea that fear is an evolved mechanism for responding to threats experienced involuntarily.

R2.2. We intentionally avoided referring to life history theory

We actively avoided characterising our explanation for the BCD as “life history theory” in our target article, although the term might be relevant. Despite this omission, many of our commentators' responses suggest that they assumed our contextually appropriate response explanation for the BCD to be life history theory repackaged (Brown & Olding; Grossmann & Varnum; Jones; Mittal & Griskevicius; Quinlan; Uggla). We therefore take this opportunity to clarify why the BCD should not be conceptualised as a life history strategy and why we do not wish to claim that it can be explained by life history theory.

Our chief motivation was that in the human sciences, life history theory has come to have several distinct referents. Life-history theory was originally developed to explain species-typical patterns of growth and reproduction in terms of fitness maximisation in a given ecological context (Promislow & Harvey Reference Promislow and Harvey1990; Stearns Reference Stearns1977; Reference Stearns1992). In this sense, life history theory is really a set of methods (formal approaches for explicitly modelling how selection would be expected to act on patterns of growth and reproduction) rather than any particular set of empirical claims. Indeed, one of the main lessons of this kind of modelling is that selection can favour many different things, depending on detailed assumptions about the ecology, the biology of the organism, demography, and other such factors. Thus, specific life history models make specific predictions, but it is hard to make statements about the general predictions of life history theory, in the sense described above (Baldini Reference Baldini2015). For this reason, although we discussed some relevant life history models, we would not describe the existence of the BCD as “predicted by life history theory” without further specification.

A second sense of life history theory that one finds is the idea that a broad group of human behaviours covary along a single axis of “slow” to “fast” (e.g., Dunkel & Decker Reference Dunkel and Decker2010; Dunkel et al. Reference Dunkel, Summerville, Mathes and Kesserling2014; Figueredo et al. Reference Figueredo, Vásquez, Brumbach and Schneider2007; Giosan Reference Giosan2006). Many of these behaviours have no obvious connection to growth, reproduction, or somatic maintenance, and this sense of “life history theory” does not follow in a simple way from the first sense (described above). Our explanation for the BCD suggests that some of these behaviours might be correlated due to common ecological drivers more than shared biological mechanisms. We also note that the questionnaire scales measuring this putative continuum often conflate potential causes of behaviour with the behavioural responses themselves. For example the “mini-K” scale of the Arizona Life History Battery (http://www.u.arizona.edu/~ajf/pdf/Arizona%20K-Battery.pdf; Figueredo Reference Figueredo2007) measures agreement with statements of experience, such as, “While growing up, I had a close and warm relationship with my biological mother,” in sum with self-reported behaviours, such as, “I avoid taking risks.” Thus, the construct validity and conceptual utility of such scales can be questioned (Copping et al. Reference Copping, Campbell and Muncer2014).

Still a third sense of life history theory is the idea that particular childhood experiences lastingly calibrate adults' behavioural strategies (Ellis et al. Reference Ellis, Figueredo, Brumbach and Schlomer2009; Griskevicius et al. Reference Griskevicius, Tybur, Delton and Robertson2011b; Hill et al. Reference Hill, Prokosch, DelPriore, Griskevicius and Kramer2016). In our explanation for the BCD, we do not exclusively privilege childhood experiences or irreversible developmental plasticity. Our account is compatible with day-to-day ongoing experience of low SES in adulthood playing a large role in the behaviours of the BCD. Thus, we do not wish to identify our account with life history theory in this sense, although our account does not necessarily conflict with it, either. In summary, all three senses of life history theory have potential relevance to our proposed explanation for the BCD, but, given the various uses of the term, to have used it as a descriptor of our account would invite confusion about our precise claims.

R2.3. We didn't claim that the BCD is adaptive, rational, logical, or optimal

Section 8.4 of our target article is titled “The BCD is not necessarily adaptive and perceptions are not necessarily accurate.” Nonetheless, many of our commentators used the word “adaptive” in relation to the BCD (Białek & Reddy; Brown & Olding; Gassen, Bradshaw, Leyva, & Hill [Gassen et al.]; Grossmann & Varnum; Jones; Quinlan; Rad & Ginges; Rickard; Robinson & Piff). Meanwhile, others used the word “rational” (Carmel & Leiser; Doebel et al.; Srivastava & Srinivasan), assuming that this is what we meant when we wrote that BCD behaviours can be seen as contextually appropriate. However, we actively avoided using such terms as “adaptive,” “rational,” and “logical” in our target article, because these words have very specific (although sometimes multiple) meanings that we did not wish to evoke.

To say that a behaviour is “adaptive” in the evolutionary sense is to suggest that it maximises Darwinian fitness. As we outlined in the target article (sect. 8.4), it is plausible that the tendency to prioritise more immediate outcomes over delayed ones, given certain environmental contexts, may have been adaptive in ancestral environments. However, any psychological mechanisms we have for making such decisions evolved to deal with cues and cue-world mappings that are potentially different from those faced in contemporary societies. Various features of our current environments may skew perceptions and behaviours away from what is strictly adaptive. Indeed, the commentary by Lewis & Lewis gives a more detailed account as to how maladaptive behaviours might arise as a nonfunctional by-product of adaptive mechanisms. The commentary by Rickard is also relevant here, as it outlines how the ability to accumulate capital to the advantage of one's descendants is a relatively new phenomenon, meaning that responses to this possibility may not be adaptive in the Darwinian sense.

In our target article we also avoided describing BCD behaviours as “rational,” because doing so might imply that the behaviours ought to maximise utility. This assumption of rationality is one that Carmel & Leiser accuse us of making; they state, “The ‘poor but neoclassical’ approach treats poor people as utility-maximizing agents and focuses on the structural constraints that affect decision making as a consequence of reduced opportunities and incomplete information” (para. 1). Carmel & Leiser are close to the truth in saying that we present an evolutionary take on this line of thought. However, we do not assume that people are necessarily operating as utility-maximising agents any more than they are fitness-maximising agents – hence, our description of the BCD as contextually appropriate rather than rational or adaptive.

We also chose to not use the term “logical” to describe the BCD, because, for some, this term can imply a consciously reasoned decision-making process. As we discussed in the target article (sect. 4.1), the BCD may be delivered by reflective or automatic psychological processes. That is, the mechanisms underlying the BCD need not always involve conscious reasoning.

Having discounted the use of the words “adaptive,” “rational,” and “logical,” we were forced to contemplate alternative terms. We decided against the use of the word “optimal,” because it provokes thoughts of the sort of fitness- and utility-maximising models that led us to avoid using the words “adaptive” and “rational,” and because, to some of our readers, it might imply a value judgement about whether a given behaviour is socially desirable. Left with a somewhat constrained vocabulary, we settled on “contextually appropriate.” We hoped this term would convey the idea that BCD behaviours are understandable given the contexts in which they are expressed, without implying that they necessarily maximise current fitness or utility, involve conscious reasoning processes, or should be valued differently than other behaviours. The exact entailments of a claim of “contextual appropriateness” are, we admit, rather underspecified, given that it does not necessarily mean utility or fitness maximising in the current environment, nor does it mean consciously thought through. Instead, the notion of contextual appropriateness must ultimately be grounded in the normal functioning of context-dependent psychological mechanisms in response to particular classes of environmental experience (Lewis & Lewis). Further theorizing of what exactly contextual appropriateness means is required if, as Brezina and Lewis & Lewis suggest, the term is to become a useful explanatory one for social issues.

R4.1. The BCD as a risk-management strategy

In the target article, we focussed on temporal discounting as a core concept connecting the BCD behaviours, which we suggest are driven by the experience of uncontrollable future-limiting factors. We touched briefly on the concept of risk acceptance (defined as a willingness to accept options associated with variable outcomes over less-variable options of equal expected value), noting that the contextually appropriate response perspective does not, without further assumptions, make predictions about this form of risk acceptance. However, as we mentioned in the target article, and as the commentaries of Jones; Mell, Baumard, & André (Mell et al.); and Amir & Jordan point out, evidence suggests that preferences for immediate rewards are driven by the inherent uncertainty of future outcomes (Andreoni & Sprenger Reference Andreoni and Sprenger2012; Weber & Chapman Reference Weber and Chapman2005). That is, temporal discounting is a response to perceived collection risk (Mell et al.).

Amir & Jordan make the excellent suggestion that temporal discounting can be conceptualised as a risk-management strategy by which people respond to uncertainty by becoming more present-oriented. They point out that by conceptualising the BCD as a risk-management strategy, we can more parsimoniously account for behaviours involving temporal trade-offs as well as those involving risk and social preferences. As motivation for risk aversion at lower SES, Amir & Jordan highlight the marked effects of fluctuations in resource availability for those living in deprivation: “small and moderate fluctuations in resources (income, calories, etc.) are unavoidable, but only those at the margins feel the full effects of such fluctuations and, consequently, must be more attentive to variability in their environment and the downside risk of their decisions” (para. 2). Conceptualising the BCD as a risk-management strategy, as Amir & Jordan suggested, results in a more inclusive BCD, thus offering a more comprehensive framework for understanding socioeconomic variation in behaviour (and thereby addressing some of the concerns expressed by commentators regarding the restricted range of behaviours considered in the BCD – see sect. R1.1). For example, it might allow us to accommodate Robinson & Piff's suggestions that prosocial behaviours among lower-SES individuals are a strategy to increase personal control to buffer against external threats.

Conceptualising the BCD as a risk-management strategy also allows us to account for the role of unpredictability, as suggested by Gassen et al. and by Mittal & Griskevicius. As Gassen et al. point out, both actual control and perceived control are reduced when outcomes are unpredictable, and unpredictability is greater in lower-SES environments. Further, as Srivastava & Srinivasan emphasise, lower-SES people more frequently operate close to budgetary boundaries, meaning that unpredictable changes in circumstances have more severe consequences. Nonetheless, as stated by Mittal & Griskevicius, “Harshness and unpredictability might have distinct effects, because the adaptive methods to deal with a consistently harsh environment are different from the methods to deal with a rapidly changing and inconsistent environment” (para. 4). Therefore, more empirical work is needed to establish whether harshness and unpredictability produce distinct effects or whether these are different factors that produce similar-looking consequences. Fortunately, Frankenhuis et al. (Reference Frankenhuis, Gergely and Watson2013) have provided a developmental theoretical model suggesting that harshness and unpredictability produce distinct contingency profiles, which offers testable predictions about the maturation of infants under these conditions. Such models provide useful starting points for devising empirical tests and improving our understanding of the effects of different environments.

The commentary by Mell et al. also contributes to our understanding of the BCD in terms of risk management. They suggest that although we explained the BCD in terms of the collection risk associated with future rewards, we should also consider the costs of waiting for future rewards, even when the future rewards are guaranteed. They provided some excellent examples of situations in which waiting would not be contextually appropriate, because the costs of waiting would outweigh the rewards of doing so. For example, they present the following thought experiment: “Imagine a farmer who participates in an economic study in which he is offered a choice between receiving $1,000 now or $2,000 in a month. Because this particular farmer does not own any expensive agricultural equipment, he is only able to sow half of his fields simultaneously. However, $1,000 now would allow him to buy new equipment and exploit his whole farm. This would yield him an expected $2,500 increase in revenue by the end of the month. Hence, our farmer should prefer the smaller-sooner reward, even though the collection risk in our example could be close to zero and the larger reward is only delayed by a month. Instead, the fact that his current level of capital is associated with a particularly high opportunity cost in productivity determines his choice” (para. 4). In this manner, the commentary by Mell et al. gives several clear accounts of ways, other than collection risk, in which economic deprivation may increase the value of more immediate rewards.

On a similar theme, Sear & Schaffnit call for an increased focus on the costs and benefits of behaviours in the present. They suggest that “the costs and benefits of behaviours in the present differ by environmental context and that it is not just the size of the present versus future reward that matters but that the costs and benefits may also be realised in different currencies in different environments” (para. 3; emphasis original). Because of this, they emphasise, we cannot always simplistically characterise the decisions that people face as smaller-sooner versus later-larger rewards. Sometimes the payoff in the present is the larger of the two, or it is in a favoured currency, in which case patience still does not pay. They illustrated this idea with the example of investment in education. In the target article, we suggested that it is less beneficial for people to invest in education if they are unlikely to reap much reward from it later in life. Sear & Schaffnit add that lower-SES individuals may benefit from leaving school early by gaining immediate earnings and opportunities to acquire other skills, which may be more valuable in their context. In this sense, the additions offered by Sear & Schaffnit tie in with those of Mell et al., who emphasise the effects of opportunity costs.

Collectively, the commentaries reviewed in this section make a compelling case for viewing the BCD as the behavioural result of strategies for managing risks, costs, and benefits under conditions of limited resources, as well as limited control. This perspective is not incompatible with the one we presented in the target article – rather, it is a superordinate explanation, which allows us to parsimoniously account for a broader constellation of behaviours, including those involving temporal trade-offs. As such, it is a welcome extension of the perspective we originally presented, and one that we wish to champion.

R4.2. The physiological constellation of deprivation

In our target article, we suggested that socioeconomic inequalities in life expectancy can be understood as resulting from a combination of differential extrinsic mortality risk and the intrinsic mortality risk it causes via a double disinvestment in behavioural and physiological investments in health. In response, Garcia & Blackwell suggest a subtly different perspective on our concept of physiological disinvestment: that “development in a deprived environment may lead to not just disinvestment in repair and immune mechanisms but also investment into alternate kinds of immune defense and repair” (para. 1; emphasis original). They go on to explain how the increased pro-inflammatory responses associated with experiences of early-life stress “might represent a predictive adaptive response that evolved in ancestral environments in which uncertainty was coupled with greater risks of injury and illness” (para. 2). Garcia & Blackwell explain that this innate, more general, inflammatory response is quicker and cheaper to establish than more specific adaptive (in the immunological sense) responses, which take time to develop. They suggest that the costs and time required to generate a more specific immune response mean that a stronger innate immune response may be preferred when time is short.

In the sense that it is focussed on differential investment rather than disinvestment, Garcia & Blackwell's perspective has something in common with the strengths-based approach suggested by Frankenhuis & Ellis (see R4.3 for further discussion). We applaud this suggestion, as it moves thinking away from deficit models and more firmly in the direction of contextually appropriate allocation of resources.

Relatedly, Brown & Olding discuss the epigenetic processes that might underlie the behavioural and physiological constellations of deprivation. They propose that the costs and benefits of the various possible contextually appropriate responses to environment are moderated by age and that DNA methylation profiles might act as an epigenetic clock, regulating such responses in line with age.

R4.3. Taking a more strengths-based approach

In our target article, we advocated moving away from purely deficit-based thinking regarding the behavioural effects of deprivation. We emphasised that we view the present-oriented behaviours of the BCD as a contextually appropriate response to structural and ecological factors rather than as a pathology or a failure of willpower. Yet the responses of our commentators suggest to us that, perhaps, we did not go far enough. Frankenhuis & Ellis go a step further to promote a strength-based approach, suggesting that “contextually appropriate responses may also include the development of enhanced skills and abilities that are ecologically relevant in harsh, unpredictable environments” (para. 2; emphasis original). They explore the potential effects of people's adapting to stressful environments, highlighting that people from unpredictable environments may become better able to shift attention between tasks or to track novel information. They also underscore the implications of taking a fuller perspective that acknowledges the strengths developed by people who have experienced various adversities. They concluded that understanding the enhanced skills and abilities developed in environments of deprivation could allow us to design classroom and work-training environments that work with rather than against the strengths of stress-adapted people.

Related to the call from Frankenhuis & Ellis to focus on the strengths developed by people in harsh and unpredictable environments, several commentators emphasise the socially desirable behaviours that are more common in lower-SES communities. For example, Robinson & Piff review the prosocial tendencies that are more common among lower-SES individuals, suggesting that “cooperative communities may provide a form of ‘collective control’ over threatening environments” (para. 9). Srivastava & Srinivasan cite the charitable behaviour observed among lower-SES individuals and called for a greater focus on such positive behaviours. Similarly, Pearson & van der Linden focus on environmental concern as an example of a socially desirable attitude, which they say is stronger among economically disadvantaged groups (although cf. Gifford & Nilsson Reference Gifford and Nilsson2014). We chose not to focus on these behaviours in our BCD, because they are not so straightforwardly joined by the conceptual threads of low control and temporal discounting as the other behaviours. However, we certainly do not wish to paint a negative picture regarding those behaviours that are more frequently seen at lower SES, and we agree that it is important to emphasise the socially desirable traits and abilities that are more common in lower-SES people.

R4.4. Early-life learning experiences are beyond control

In our target article, we proposed that limited control over future outcomes should be a driving factor of the BCD. We also outlined some ways in which the effects of early-life adversity can become embedded and amplified through feedback loops (sect. 3.2). However, we did not emphasise the fact that early-life experiences, especially for altricial species, are beyond individual control. This idea is highlighted in the commentary by Kurkul & Corrieveau, who discuss how early-learning experiences are beyond individual control and consequently may contribute to the BCD. They describe how lower-SES children have fewer opportunities to engage in new activities that help them acquire information, creating inequalities in conceptual knowledge and information-seeking behaviours, which are then reinforced by feedback loops. As they put it, “A virtuous cycle of learning occurs for children who have access to the type of numerous, rich, and varied experiences that support acquisition of knowledge about the world” (para. 3). They propose that for those who lack access to such rich early-learning opportunities, the potential for later educational success is constrained; as we said in our target article, constraints breed constraints.

R4.5. Additional routes for intergenerational transmission

In section 4.4 of the target article, we briefly touched upon the possible biological mechanisms by which stresses and disadvantages may be passed down through generations. Rickard helpfully expands upon this picture in his discussion of the pathways by which various forms of capital may be intergenerationally transmitted. He also explains that the evolutionarily novel complexity of our societies and their varied routes for transmission of capital may mean that behavioural responses to having capital (or lack thereof) are maladaptive – another reason for us to label the BCD as a contextually appropriate response rather than an adaptive one (see section R2.3).

Relatedly, Białek & Reddy extend our discussion of the intergenerational transmission of disadvantage through biological mechanisms (sect. 4.4 of the target article) to include what infants learn about the world through their parents. Their commentary lists a number of ways in which patterns of temporal discounting may be transmitted intergenerationally via social learning mechanisms (an intergenerational case of the social learning processes we discussed in sect. 4.3 of the target article).