Extended parenting and the evolution of cognition

Abstract

Traditional attempts to understand the evolution of human cognition compare humans with other primates. This research showed that relative brain size covaries with cognitive skills, while adaptations that buffer the developmental and energetic costs of large brains (e.g. allomaternal care), and ecological or social benefits of cognitive abilities, are critical for their evolution. To understand the drivers of cognitive adaptations, it is profitable to consider distant lineages with convergently evolved cognitions. Here, we examine the facilitators of cognitive evolution in corvid birds, where some species display cultural learning, with an emphasis on family life. We propose that extended parenting (protracted parent–offspring association) is pivotal in the evolution of cognition: it combines critical life-history, social and ecological conditions allowing for the development and maintenance of cognitive skillsets that confer fitness benefits to individuals. This novel hypothesis complements the extended childhood idea by considering the parents' role in juvenile development. Using phylogenetic comparative analyses, we show that corvids have larger body sizes, longer development times, extended parenting and larger relative brain sizes than other passerines. Case studies from two corvid species with different ecologies and social systems highlight the critical role of life-history features on juveniles’ cognitive development: extended parenting provides a safe haven, access to tolerant role models, reliable learning opportunities and food, resulting in higher survival. The benefits of extended juvenile learning periods, over evolutionary time, lead to selection for expanded cognitive skillsets. Similarly, in our ancestors, cooperative breeding and increased group sizes facilitated learning and teaching. Our analyses highlight the critical role of life-history, ecological and social factors that underlie both extended parenting and expanded cognitive skillsets.

This article is part of the theme issue ‘Life history and learning: how childhood, caregiving and old age shape cognition and culture in humans and other animals’.

1. Introduction

The diverse and flexible cognitive abilities of our species are considered pivotal for our evolutionary history [1]. Our extended childhood is a key life history trait that impacts on human cognition, giving humans a period of cognitive flexibility (changeability) to explore more options during learning [1–6]. Life history describes the age-specific patterns that are central to an organism's life, including the reproductive allocation strategies, development and ageing patterns [7]. However, extended developmental periods evolved not only in humans and other primates, but also in bats, cetaceans, elephants and several bird families [7]. Thus, we ask here two questions: (i) are the effects of extended childhood on cognition specific to humans, or can they be generalized to other species with relatively expanded cognitive skillsets? and (ii) what is the particular role of parenting on the development of cognition in individuals?

While the term ‘cognition’ has no agreed definition [8], we follow here Allen [9, p. 43], defining it as the brain's ‘synthesis of information from diverse sensory and memory sources to produce appropriate responses'. We refer to skillsets as the combination of cognitive abilities possessed by a species or an individual. Evolutionary cognitive studies investigate cognitive changes over time in species. The ‘embodied cognition’ approach, which views cognition as a result of interactions between an individual's body and its environment [10,11], is an increasingly influential view in cognitive science [11–13]. However, the interactions between life history, ecology and learning have rarely been considered in cognitive evolution research. Thus, we use here an integrative approach to propose a new model of cognitive evolution based on extended childhoods and extended parenting typical of family-living species.

Many scholars consider brain size as a useful proxy for cognitive abilities because species with larger brains tend to show more diverse and flexible behaviours [13], but the validity of this proxy is debated [11,14–16]. Because brain morphology can differ between lineages [14,17], it is recommended to relate brain sizes to cognitive abilities only within lineages, and to consider behaviour when inferring cognitive abilities in individuals or species [14]. Work focusing on brains showed that they are paradoxical adaptations for two reasons. First, brain tissue is energetically very expensive to grow and maintain [18], and thus, large brains can only evolve if their costs are buffered through prolonged developmental periods and/or allomaternal energy inputs (expensive brain framework [19]). Second, it takes time for an individual's brain to develop the cognitive abilities that make this adaptation worthwhile. Thus, cognitive adaptations are constructed developmentally, and their development relies on access to learning opportunities [20–22]. The combination of these factors stresses the importance of ontogeny, and a constant, reliable access to resources during development to offset the cost of large brains [23]. Thus, selection on brain size and learning ability is associated with extended developmental periods [24,25]. Relevant for models of cognitive evolution is that individuals with extended learning periods are predicted to develop larger cognitive skillsets, allowing them to more successfully reproduce and avoid extrinsic mortality [24].

Here, we focus on the coevolutionary links of life history traits with extended skill-learning periods. While learning is costly, including increased parental investment and higher vulnerability of naive juveniles [26], several studies showed fitness benefits of faster learning ability [27,28]. Previous work assessed particular drivers of cognitive abilities (usually using brain size as a proxy), including ecological, social and cultural drivers, or assessed very specific aspects of cognition, such as innovations [3,16,22,23,29–31], but rarely considered the evolutionary interplay between developmental trajectories, sociality, cognition and brain evolution [4,19,30,32,33]. Previous work showed that enlarged brain size evolved in species with a larger body size, larger group sizes, longer reproductive lifespan, more reliance on social learning and more variable environments [34]. These data are largely in line with the cognitive buffer hypothesis, which proposes that larger brains provide more abilities that help to survive unfavourable conditions [30]. Nevertheless, what remains to be clarified is the fitness benefit of extended parenting for the evolution of cognitive abilities. We develop below, the extended parenting hypothesis, which extends and complements the ‘extended childhood’ hypothesis [2,3,35–38] based on a phylogenetic comparison of corvids with other passerines, and insights from two long-term field studies on corvids.

2. The extended parenting hypothesis

We will argue that the ontogenetic development of expanded and more flexible cognitive skillsets requires: (i) a fitness benefit, (ii) ample and reliable social learning opportunities (facilitated by extended childhood and extended parenting), (iii) the buffering of the costs of developing and maintaining a large brain, and (iv) ecological conditions that facilitate extended parenting (figure 1). The comparative data below show that extended parenting is critical for the evolution of large brains and large cognitive skillsets in most cases. Previous models and comparative work identified a set of ecological, social and life history parameters that are associated with larger brains: (i) ecological and social challenges that favour certain cognitive abilities [30], (ii) a life history that provides the opportunity to develop more cognitive and sensory–motor abilities and covers their developmental and maintenance costs [23,32], and (iii) a fitness benefit from having these abilities [39]. These requirements are found not only in species with increased levels of alloparental care [40], but also in species with extended parenting. In birds, extended parenting provides an evolutionary steppingstone for transitions to cooperative breeding, where others than parents provide parental care [41], alleviating the development and maintenance costs of large brains. Comparative work showed that family living is associated with a larger body size, an increased lifespan and productive, mild environments [41]. Thus, building upon the cultural intelligence idea [22], we propose here that parenting itself is pivotal for the social environment that favours the evolution of cognitive adaptations through learning, and that it links to environmental conditions which support the costs of large brains and provide a benefit from increased skillsets [39] (figure 1).

Figure 1.

The extended parenting hypothesis. Grey box: aspects related to living in family groups. Green stippled boxes: ecological aspects; purple checkered boxes: life history aspects; blue boxes: extended parenting and social aspects; orange box: cognitive aspects; brown dashed box: costs of learning and brains; black squared boxes: links to other hypotheses.

For extended parenting to evolve, it must confer benefits that offset the costs of delayed maturation [7,42], for example, increased survival [43]. Increased flexibility and innovativeness, such as rates of novel feeding behaviours in natural populations, can also facilitate survival in changing environments [39,44]. An extended developmental and childhood phase provides more time to learn difficult skills, supporting the evolution of socioecological niches with specialized foraging techniques [45–47]. In humans, cognitive flexibility is greatest in preschool-aged children but decreases subsequently, which suggests that the evolution of an extended childhood favoured the variable use of a more exploratory learning strategy early on and fixed learned strategies later on in life, creating a fruitful balance between innovation and imitation in cultural learning [2]. An extended childhood has been an integral part of human life history patterns for at least 600 000 years, as shown by fossil evidence for body mass, brain size and dental development of human ancestors and related extinct species [48]. Our ancestors' technological skills included the most difficult to learn technologies documented until that time (i.e. involving the most elements and hierarchical levels of action sequencing, and demanding the longest learning times), as evidenced by archaeological tool analyses and transmission experiments [49–53]. Learning a skill, including stone tool-making that demands extensive time investment, is costly to the individual and the group. In humans, apprenticeship strategies evolved to reduce these costs [49,50,54]. Because offspring must be fed until they have acquired a self-sustaining level of foraging skills [24], a lengthening of provisioning time increases the costs to the feeders. Taken together, these data illustrate the critical role of costs in our model for both parents and learners.

3. Corvid life histories: comparison with other passerines

Research on human cognitive evolution benefits from consideration of non-primate species [23,30,55]. Birds, which have a 300 million-year-long evolutionary history independent of mammals, provide an excellent outgroup to untangle the factors involved in cognitive evolution. We focus on corvids because they show convergent cognitive abilities that rival apes in many domains, such as in tool manufacture, planning and insight [56]. These abilities have been linked to an increase in neuron density in their neocortex, particularly the telencephalon [17]. Thus, corvids provide us with a comparison to early human ancestors as they present an intriguing example of convergent evolution in innovative problem-solving, increased manipulative dexterity, technical skills and sociocultural transmission of skills that lead to cultural variation [55,57].

Corvids are a globally distributed passerine family that includes 127 species. Corvids and 30 related bird families form the super-clade of Corvides, which includes other bird species with large brain sizes (e.g. drongos, currawongs, Australian magpies). Corvids originated in the Australo-Papuan region around 14 Ma in rainforest habitats [58], and interestingly, they differ in a large number of traits from other passerines (table 1). Corvids are among the largest passerines, and have much larger relative brain sizes. The incubation time is longer than the passerine mean, and the time spent in the nest is almost double the mean for other passerines. Importantly for learning, offspring have extended periods beyond fledging where they remain associated with their parents compared to other passerines. Finally, a high number of corvid species breed cooperatively.

Table 1.

Basic life history characteristics of corvids (127 possible species) compared to other passerines (4796 possible species). (Data were compiled from published datasets [16,18,41,59,60], the online version of the Handbook of the Birds of the World [61]; unpublished data on brain size compiled by K. Isler 2014, personal communication; unpublished data on corvid social characteristics compiled by A. Krone and A. Clark 2019, personal communication; and a recent phylogeny [62]. Parenting time: time offspring remain associated with their parents beyond independence. Residual brain size: log(brain mass (g))/log(body mass (g)). Phylogenetic controlled comparisons (corvids versus other passerines) using phylogenetic generalized least squares (PGLSs) regressions in the R-package Geiger for continuous parameters (providing t-values), respectively phylogenetic logistic regressions in the R-package phylolm (providing z-values). Bold denotes traits with statistically significant differences (p < 0.05).)

			trait mean ± s.e.
trait	n corvid species	n other passerine species	corvids	other passerines	λ	estimate	s.e.	t-value/z-value	p-value
adult body weight (g)	70	1999	267.8 ± 23.83	37.9 ± 3.14	1	−89.58	163.18	−0.55	0.58
median clutch size	64	1577	4.2 ± 0.11	3.2 ± 0.03	0.85	−1.20	0.59	−2.04	0.041
incubation time (days)	42	1019	18.3 ± .0.21	14.5 ± 0.10	0.92	−1.17	1.42	−0.83	0.41
nestling time (days)	41	997	28. ± 91.29	16.0 ± 0.16	0.93	−8.28	2.46	−3.37	0.0008
parenting time (days)	32	557	300.8 ± 70.4	98.1 ± 6.16	0.74	−181.21	86.49	−2.10	0.036
lifespan (years)	30	701	17.7 ± 0.97	9.8 ± 0.16	0.70	−7.66	2.21	−3.46	0.0006
family living (%)	62	1820	81 ± 0.05	48 ± 0.01	0.70	−1.49	0.59	−2.51	0.012
cooperative breeding (%)	63	1819	48 ± 0.06	16 ± 0.01	0.69	−1.00	0.59	−1.71	0.087
residual brain size	28	426	0.31 ± 0.01	−0.09 ± 0.01	0.95	−0.26	0.1	−2.61	0.009

Although we lack detailed data on the brain structure of many species, our phylogenetic analyses show that bird species with extended family life have larger telencephalons, and a larger proportion of neurons located in their telencephalon [17] (table 2). Thus, corvids stand out from other passerines not only in terms of their cognitive abilities, but also in having a set of life history features linked to extended parenting. But how does extended parenting affect learning, and importantly, survival? Two case studies that directly addressed learning in corvids are presented below. Crucially, the data from these species are not confounded by potential effects of cooperative breeding, which has been previously identified to facilitate the evolution of large brains [30,40].

Table 2.

Phylogenetic regression models using PGLSs in the R-package Geiger, assessing the effect of family system (non-family living versus family living, [41]; this information is available for all n = 26 bird species included) and the post-fledging association time of offspring with their parents (data available for n = 17 species [61]; data for kea: A. Bond 2019, personal communication) on relative size (a,b) and relative number of neurons (c,d) in the telencephalon of birds. (Body mass is included to control for its effects on brain parameters. Telencephalon data from [17] on n = 26 altricial bird species; models including post-fledging association time. Bold denotes traits with statistically significant differences (p < 0.05).)

trait	estimate	s.e.	t-value	p-value
(a) telencephalon size in relation to total brain mass (%)
intercept	60.44	3.95	15.29	<0.001
non-family versus family living	−12.23	3.96	−3.09	0.005
body weight	0.01	0.01	2.04	0.053
(b)
intercept	45.23	3.94	11.48	<0.001
post-fledging association time	0.06	0.02	2.75	0.016
body weight	0.01	0.01	1.63	0.12
(c) number of neurons in telencephalon in relation to all neurons (%)
intercept	75.01	2.28	32.93	<0.001
non-family versus family living	−6.53	2.36	−2.76	0.011
body weight	0.01	0.00	2.06	0.051
(d)
intercept	66.44	2.28	29.08	<0.001
post-fledging association time	0.03	0.01	2.43	0.029
body weight	0.01	0.01	1.73	0.1

4. Siberian jays: learning opportunities matter

Siberian jays (Perisoreus infaustus; electronic supplementary material, video S1) are sedentary corvids that occur throughout the northern Palaearctic [63]. Their social system has two unique facets that provide insights into the benefits of family living. First, the species does not breed cooperatively (i.e. only parents incubate and feed young), despite the fact that offspring can remain with their parents for years after fledging, which is associated with cooperative breeding in most species [41]. Second, Siberian jay groups consist of a breeding pair, retained offspring and/or unrelated non-breeders. The latter are forced by the socially dominant retained offspring to disperse from the natal territory one to two months after fledging and settle in another group [64]. Retained offspring can remain up to an age of 4 years with their parents, which is well beyond the mean lifespan of 2.2 years [63].

Parents differ in their behaviour towards kin and non-kin (electronic supplementary material, figures S1 and S2) as they are nepotistic and only provide access to resources and predator protection to kin [63], resulting in a higher survival of kin compared to non-kin [43].

Field experiments on predator recognition and problem-solving in juveniles reveal how fitness-related proxies depend on kinship. Predation by goshawks is the main reason for mortality in Siberian jays, especially affecting juveniles [43,65]. Juveniles do not respond to perched predators when encountering them on their own, but when other group members start mobbing a perched predator, particularly retained offspring immediately copy the behaviour of their parents [66]. Natural mobbing events are very brief, as mobbed predators quickly move off, so that unrelated non-breeders have much fewer opportunities than retained offspring to observe mobbing. However, in experimental settings where predator models are presented, mobbing could last as long as 4 min, providing also unrelated non-breeders with learning opportunities. Consequently, all juveniles survived their first winter of life [67]. Moreover, retained juveniles learn to access a feeding device (electronic supplementary material, figure S3) faster than unrelated juveniles [68], as their learning is facilitated by more tolerant role models (i.e. their parents; electronic supplementary material, video S1). Thus, a life history involving extended parenting is a critical support for learning: parents provide their offspring with a safe haven, access to food and reliable learning opportunities, which together boost the long-term survival of retained offspring [67].

5. New Caledonian crows: tool manufacture matters

New Caledonian crows (Corvus moneduloides; electronic supplementary material, video S1) are endemic to the tropical South Pacific island of New Caledonia, where they live in family groups with extended dependency periods: offspring can be fed by their parents for up to 2 years [69], which might explain why this species does not breed cooperatively [69]. During this extended developmental period, juveniles have access to tolerant role models (both related and unrelated [70]), from whom they can learn tool-use and tool-making skills [70,71] (electronic supplementary material, figure S4 and video S1). Social learning experiments showed that adults and juveniles learn about the appropriate context for certain actions from each other, demonstrating the potential for lifelong learning ability [72]. In the wild, adults scaffold learning of juveniles by allowing them to be in physical contact during foraging (electronic supplementary material, figure S5), and by sometimes leaving a tool in a tree hole that juveniles can then use with success (electronic supplementary material, video S1). However, this extended tool-making learning period comes with a high cost: juveniles are unable to make functional tools until they are at least six months old. Adult-level proficiency is reached at 10–12 months of age [70], requiring that parents provision their offspring through their first year of life. Thus, wild juveniles grow up in a safe haven, surrounded by tolerant role models that constantly make and use tools (electronic supplementary material, figure S6), and juveniles have ample occasions to borrow and use other birds' tools (electronic supplementary material, video S1).

New Caledonian crows have the largest relative brain size among corvids [73], suggesting selective pressures on some aspects of cognitive performance [17]. Compared to other corvids, they have significantly larger brain areas subserving learning, action sequencing and fine motor control functions [74], all of which support the tool-making and tool-use behaviours enacted by New Caledonian crows. In experimental settings, New Caledonian crows have been shown to excel at problem-solving, physical cognition and causal reasoning in tests such as the trap-tube, string pulling and Aesop's fable tasks (in which suitable objects must be dropped into a beaker of water to raise the water level to obtain a floating reward) [75].

Several evolutionary adaptations were crucial for the emergence of tool-making skills in New Caledonian crows and representative human ancestors [55], including increases in brain size and brain networking potential, stable groups with a high social tolerance providing opportunities for social learning, and extended parenting and development periods. Increased social learning opportunities allow individuals to acquire more skills and to become more effective individual learners [20,22,76].

6. Cognitive consequences of extended parenting

In the corvid case studies we presented, Siberian jays tolerate retained offspring more than unrelated non-breeders, which increases the learning opportunities and survival prospects of retained offspring [43]. New Caledonian crows are highly tolerant to juveniles independent of their kinship [70], and adults scaffold the learning process of juveniles acquiring tool-making skills (‘education by master–apprenticeship’ [77]). In addition, juveniles of both species are highly proactive in ways that foster opportunities to observe others and to practice skills. In Siberian jays, retained juveniles pay more attention to the behaviour of the adults in their group than unrelated juveniles do [66]. Juvenile New Caledonian crows actively follow group members, direct begging behaviours towards them, steal foraged food from them and also steal ready-made tools that they can use to get food (electronic supplementary material, figures S4, S5, S6 and video S1). These behaviours are tolerated by the adults and ensure survival of juvenile corvids that are not yet nutritionally independent. Thus, learning opportunities arise from the interplay between extended childhood and extended parenting. The safe haven provided by extended parenting is critical for learning opportunities, and creates extended developmental periods that feed back into the extended childhood.

Centred on extended parenting, our model (illustrated in figure 1) extends and complements the ‘extended childhood’ model [2,3,35–38], while highlighting the active nature of both adults and juveniles. Our model integrates previous models of cognitive and brain evolution [4,16,19,29–32,78] but adds crucial details on the interactions with ecological traits, life history traits and fitness benefits. More productive environments facilitate low extrinsic mortality, leading to a longer lifespan and an associated reduction in reproductive allocation [7,42]. These conditions facilitate extended family time [41,79], and a safe haven through access to food and protection from predators [41], which in turn facilitate extended developmental periods with ample learning opportunities provided by tolerant role models [20] (i.e. the ‘playful protected learning environment’ of humans [2]). These conditions are likely to favour the evolution of larger cognitive skillsets [32,80]. The costs of extended learning periods are paid for by extended provisioning, which is enabled by productive environments. Moreover, extended family time also facilitates the evolution of cooperative breeding, which additionally can enable the evolution of larger brains (table 2) that sustain larger skillsets [40]. Extended learning time, more role models to learn from and a release from the costs of large brains allow the development of larger or more flexible cognitive skillsets, which in turn give species chances to move into new ecological niches [39,81], or allows them to cope with environmental degradation leading to variable environments [41]. This later factor is also associated with the evolution of cooperative breeding [41]. Together, these adaptations can incur fitness benefits through the safe haven, an expansion of the cognitive skillset and cooperative breeding.

The life history filter model [32] proposes that survival challenges only result in cognitive adaptations if the life history of species include a low extrinsic mortality, which in turn allows them to reallocate resources into the development and maintenance of a larger brain. Drawing upon this idea, we propose that integrating the family filter into this model can explain the absence of large brains in long-lived species owing to short family times, or the occurrence of large brains in short-lived species owing to extended parenting (figure 2). Thus, not only extrinsic mortality and size constraints [32], but also extended parenting, can explain the occurrence of grade-shifts (i.e. deviations from body sized-based expected brain sizes) in animals.

Figure 2.

The life-history and family filter explaining cognitive adaptations (modified after [32]), illustrated with eight altricial hypothetical bird lineages. Cognitive challenges can only lead to cognitive adaptations in long-lived species (allowing them to compensate for the energetic costs of increased brains and larger skillsets), which are exposed to ample learning opportunities in a safe environment (provided by extended parenting). (Online version in colour.)

7. Discussion and conclusion

The comparative data on corvids support the fundamental role of extended parenting for the evolution of cognitive skillsets through learning, and family living links to critical life-history and ecology features that have been previously identified to facilitate cognitive evolution. Extended parenting is only possible if: (i) the life history pace is slow enough so that the optimal onset of independent reproduction is not at the first opportunity [42], (ii) parents can afford extended offspring association [79], and (iii) the ecological setting provides productive environments or adaptations that buffer possible ecological costs. Particularly, the access to tolerant role models is a key variable that affects between-individual variability in learning within and between species. Extended parenting allows these high-quality relationships to develop, and has a direct impact on fitness, as shown by the restricted learning opportunities of unrelated juveniles in Siberian jays, and their lower survival [66,68]. Similarly, adult New Caledonian crows are highly tolerant to their own but also unrelated juveniles, scaffolding the sensory–motor and embodied learning process of juveniles.

This special issue examines whether and how the life history of different species is connected with their cognitive abilities. Cognitive flexibility throughout the lifetime is central to our cognitive abilities [2], and might also underlie those observed in corvids and other species, allowing individuals to adjust their behaviour in response to varying inputs. Experience continues to shape the phenotype through adulthood also beyond sensitive developmental periods, just to a lesser extent [2,35]. In experimental studies, adult New Caledonian crows can adapt their tool-use skills on a daily basis according to changing local conditions [82]. Adult Siberian jays must learn new skills, such as recognizing a rare predator species, or accessing new food sources [68]. Cognitive flexibility in humans and corvids is based on plastic alteration of brain networks, which are highly interconnected and self-organize by interacting dynamically to adapt to changing conditions [17,83–86]. Variation in the cultural environment affects brain growth in humans [87], and impacts on birds' cognitive development through the enhancement of neural plasticity mechanisms [33]. Our analysis shows that prolonged parenting is associated with a higher neuron density in the telencephalon of their neocortex across species (table 2), which underlies some of the cognitive skills that characterize corvids or parrots. Thus, extended parenting fosters an expansion of the cognitive skillset via an increase in neuron density of critical brain regions. At the individual level, we speculate that inter-individual variation in brain growth patterns could cause sociocultural diversification at the group level thanks to the sociocultural setting of tolerant group members (safe haven), combined with extended developmental periods.

In the context of this article, we need to know how developmental changes in each individual might affect evolutionary change in a whole species. For example, epigenetic changes that increase neuronal plasticity (metaplasticity) in response to new environmental and social triggers [88,89] in offspring are heritable via both biological and cultural routes [88]. These epigenetic changes might allow for increased learning ability in juveniles [90]. We also need to understand much better how learning varies through individual life stages, between species, comparatively across multiple species, and evolutionarily through time [13,91]. Most fundamentally, we are in need of more developmental studies that examine how the social environment impacts on brain development, and consequently on the skills expressed by juveniles [92].

To summarize, we presented an integrative model of cognitive evolution that incorporates life history, ecology, sociality and learning. We have discussed the evolutionary trade-offs of a life history involving extended development (including the brain) that on the one hand offers individuals longer learning periods to acquire necessary survival skills, but on the other hand is energy-expensive, often requiring elaborate food processing skills that take time to learn. Adopting a comparative approach based on life history and learning in corvids, we find notable similarities with the unusual life history of humans, with its extended childhood that has characterized our lineage for the last 600 000 years. Corvids have key characteristics that make them a relevant comparison family to understand human evolution [55]. Enlarged brains and reliance on sociocultural learning of skills, enabled by extended development periods in species with prolonged parenting and access to tolerant role models in a safe haven, are likely to result in expanded cognitive skillsets. These conditions were also present in our ancestors, for whom cooperative breeding led to a safe haven where juveniles could learn skills from extended family, including grandparents [36–38,46,93], and increased group sizes opened more learning opportunities for individuals [91,94]. The case studies on Siberian jays and New Caledonian crows show that extended family life is crucial to provide the social learning opportunities where juveniles acquire vital skills. We propose that extended parenting could well have led to the extended, lifelong learning found in humans, given its life history, ecological and social links that previous work identified to facilitate the evolution of flexible and expanded cognitive skillsets across species.

Data accessibility

Data were compiled from the published sources listed in the table captions.

Authors' contributions

N.U. and M.G. wrote the paper, with inputs from J.F. and R.D.G. All authors approved the final version.

Competing interests

We declare we have no competing interests.

Funding

N.U.'s fieldwork and the writing of this paper were made possible by support from the Max Planck Society and the support of a grant from Templeton World Charity Foundation (https://www.templetonworldcharity.org/) no. 0271. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of Templeton World Charity Foundation Inc. M.G.'s Siberian jay fieldwork was funded by the Swiss National Research Foundation (grant nos PPOOP3_123520, and PPOOP3_150752) and the University of Zurich.