YEARBOOK OF PHYSICAL ANTHROPOLOGY 40:63–89 (1997)

Evolutionary Hypotheses for Human Childhood

BARRY BOGIN Department of Behavioral Sciences, University of Michigan-Dearborn, Dearborn, Michigan 48128 KEY WORDS childhood; heterochrony; neoteny; insertion model

ABSTRACT The origins of human childhood have fascinated scholars from many disciplines. Some researchers argue that childhood, and many other human characteristics, evolved by heterochrony, an evolutionary pro- cess that alters the timing of growth stages from ancestors to their descen- dants. Other scholars argue against heterochrony, but so far have not offered a well-developed alternative hypothesis. This essay presents such an alterna- tive. Childhood is defined as a unique developmental stage of humans. Childhood is the period following infancy, when the youngster is weaned from nursing but still depends on older people for feeding and protection. The biological constraints of childhood, which include an immature dentition, a small digestive system, and a calorie-demanding brain that is both relatively large and growing rapidly, necessitate the care and feeding that older individuals must provide. Evidence is presented that childhood evolved as a new stage hominid life history, first appearing, perhaps, during the time of Homo habilis. The value of childhood is often ascribed to learning many aspects of human culture. It is certainly true that childhood provides ‘‘extra’’ time for brain development and learning. However, the initial selective value of childhood may be more closely related to parental strategies to increase reproductive success. Childhood allows a woman to give birth to new offspring and provide care for existing dependent young. Understanding the nature of childhood helps to explain why humans have lengthy development and low fertility, but greater reproductive success than any other species. Yrbk Phys Anthropol 40:63–89, 1997. ௠ 1997 Wiley-Liss, Inc.

Behold the Child among his newborn blisses, A six-years’ Darling of a Pygmy size! See, where ’mid work of his own hand he lies, Fretted by sallies of his mother’s kisses, With light upon him from his father’s eye —W. Wordsworth: Intimations of Immortality (1802–1804)

Childhood fascinates scholars and practi- these is accepted universally. Perhaps the tioners from many disciplines. Virtually all lack of agreement is due to the nature of human cultures recognize a time of life that human evolutionary biology. Allison Jolly, may be called ‘‘childhood.’’ Many historical author of The Evolution of Primate Behavior, sources from Egyptian times to the 19th states that ‘‘human evolution is a paradox. century, including Wordsworth in the poem We have become larger, with long life and above, mention that ‘‘childhood’’ occupies the immaturity, and few, much loved offspring, first 6 to 7 years of life (Boyd, 1980). Some and yet we are more, not less adaptable.’’ In explanations for the origins and functions of an attempt to resolve the paradox of human childhood have been proposed, but none of evolution and our peculiar life history Jolly

௠ 1997 WILEY-LISS, INC. 64 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997

Fig. 1. Idealized mean velocity and distance curves of growth in height for healthy girls (dashed lines) and boys (solid lines) show- ing the postnatal stages of the pat- tern of human growth. Note the spurts in growth rate at mid-child- hood and adolescence for both girls and boys. The stages of postnatal are abbreviated as follows: I, in- fancy; C, childhood; J, juvenility; A, adolescence; M, mature adult. Data used to construct the curves come from Prader (1984) and Bock and Thissen (1980). concludes in the next sentence that ‘‘mental do our offspring not take a more direct and agility buffers environmental change and rapid path to maturity? has replaced reproductive agility’’ (1985, p. This essay attempts to answers these 44). The reference to reproductive agility questions and paradoxes in terms of the means that we are a reproductively frugal evolution of human growth and develop- species compared with those that lavish ment. Throughout the essay the focus is on dozens, hundreds, or thousands of offspring children and childhood. It is argued here on each brood or litter. However, a paradox that childhood is a unique stage of the remains, for the emphasis on ‘‘brain power’’ human life cycle, a stage not to be found in rather than reproduction is an apparent the life cycle of any other living mammal. It exception to Darwin’s rules of natural selec- is important to define clearly what is meant tion. Evolutionary success is traditionally by childhood, for often the terms ‘‘child,’’ measured in terms of the number of off- ‘‘juvenile,’’ and ‘‘adolescent’’ are used inter- spring that survive and reproduce. Biologi- changeably in the literature. cal and behavioral traits do not evolve un- Childhood is defined here as the period less they confer upon their owners some following infancy, when the youngster is degree of reproductive advantage in terms of weaned from nursing but still depends on survivors a generation or more later. If we older people for feeding and protection. In- are truly interested in our immortality, or at fants by definition are exclusively or par- least that of our DNA, then why do we not tially breast-fed. One survey of the world- produce more offspring, instead of a few wide variation in the age for termination of mentally agile offspring? breast-feeding, that is weaning age, finds it Another question is, why do our offspring to occur at a median age of 36 months after take so long to reach reproductive age? After birth (Dettwyler, 1995). Though no longer all, our genetic immortality requires that we breast-fed, children are still dependent on have grandchildren, great-grandchildren, older people for feeding. Several biological and so on. Yet we take two decades of and behavioral characteristics of the young- postnatal development to reach reproduc- ster necessitate this dependency of child- tively successful adulthood. Moreover, our hood. In terms of feeding there are three path from birth to maturity is sinuous, major biological factors (Fig. 2): 1) children meandering through alternating periods of continue the rapid brain growth experienced rapid and relatively slow development. This by infants and need a diet dense in energy may be seen in the examples of the distance and protein to support this brain growth, 2) and velocity curves of growth for healthy children possess deciduous teeth, with thin boys and girls illustrated in Figure 1. Why enamel and shallow roots, and cannot pro- Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 65

Fig. 2. Growth curves for different body tissues. The ‘‘brain’’ curve is for total weight of the brain (Cabana et al., 1993). The ‘‘dentition’’ curve is the median matu- rity score for girls based on the seven left mandibular teeth (I1, I2, C, PM1, PM2, M1, M2) using the reference data of Demir- jian (1978). The ‘‘body’’ curve represents growth in stature or total body weight. The ‘‘reproductive’’ curve represents the weight of the gonads and primary repro- ductive organs (Scammon, 1930). cess the adult-type diet, and 3) children may be thought of as a slowing down of the have relatively small body size, and hence a rate of development. Neoteny produces an small digestive system, which limits total adult descendant that retains many imma- food intake, furthering the requirement for ture characteristics of its ancestor. Hyper- nutrient-dense foods. These physical charac- morphosis may be defined, at this point, as teristics are coupled with motor and cogni- an extension of the growth and development tive immaturity, as well as social inexperi- period of the descendant beyond that of the ence. All of these factors render the child ancestor. Hypermorphosis produces a de- dependent on older individuals for care. scendant with features that are hyperma- The childhood stage precedes the juvenile ture compared with the ancestor. This essay stage of the life cycle. A juvenile stage is will show that both the Peter Pan scenario of common to most species of social mammals neoteny and the Methuselah-like develop- (Pereira and Fairbanks, 1993), and is a time ment of hypermorphosis are inadequate to of feeding independence from older individu- account for the evolution of childhood. More- als prior to the onset of reproductive matu- over, neither neoteny nor hypermorphosis rity (Pereiera and Altmann, 1985). Based on can unravel the paradox of human evolu- a variety of biological, behavioral, and cogni- tion. To resolve the puzzle of why we have so tive traits and abilities that are described in few offspring, why they take so long to detail below, human children enter the juve- develop and reproduce, and why our rate of nile stage at about 7 years of age. By these growth takes a serpentine path to adulthood criteria, the human childhood stage of life requires a new, and more mature, view of spans the time from about 3 to 7 years of age childhood and its place in human evolution. (Bogin, 1990, 1995; Bogin and Smith, 1996). HUMAN ONTOGENY AND More detailed evidence for this biological HETEROCHRONY definition of childhood is presented later in this essay. One point to stress at this junc- Ontogeny refers to the process of growth, ture, however, is that childhood is likely to development, and maturation of the indi- be a new life cycle stage, that was evolved de vidual organism from conception to death. It novo into hominid life history. This view of is virtually axiomatic that every species has childhood stands in diametrical opposition its own unique pattern of ontogeny (Bonner, to older, but still widely held views that 1965; Gould, 1977). During hominid evolu- childhood evolved by altering the develop- tion the form and function of our ancestor’s mental timing of the preexisting life stages structural and regulatory DNA was re- of our primate ancestors. The two most worked to produce the genetic basis for the popular hypotheses in this regard invoke ontogeny of the human species. The litera- either neoteny or hypermorphosis as the ture is replete with proposals for how the primary agent of human evolution. Neoteny reworking occurred. One tradition in the 66 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997 study of human evolution looks for a single down and juvenile stages of ancestors be- major cause or process. It has been argued come adult features of descendants. Many that humans evolved when we became big- central features of our anatomy link us with brained apes, terrestrial apes, killer apes, the fetal and juvenile stages of [non-human] hunting apes, aquatic apes, toolmaking apes, primates . . .’’ (Gould, 1981, p. 333). symbolic apes, monogamous apes, food- sharing apes, and, even, apes with ventral- NEOTENY AND HUMAN EVOLUTION ventral copulatory behavior. None of these, That humans of all ages are essentially or any other single-factor hypothesis, proves child-like in morphology, behavior, and cog- to be helpful to understand human evolu- nitive potential is the essence of the concept tion, for a non-human primate exception can of neoteny. Notions of human neoteny may always be found. Another tradition looks be traced back as far as biblical writings instead at the pattern of ontogeny. In the (Montagu, 1989), and the concept has domi- book Size and Cycle, J.T. Bonner (1965) nated popular and scientific attitudes to- develops the idea that the stages of the life ward the evolution of human growth for cycle of an individual organism, a colony, or hundreds of years. William Wordsworth, for a society are ‘‘the basic unit of natural example, praised the concept of neoteny in selection.’’ Bonner’s focus on life cycle stages 1802 with words of innocence and hope: follows from the research of several 19th- and 20th-century embryologists who pro- My heart leaps up when I behold posed that speciation is often achieved by A rainbow in the sky: altering rates of growth of existing life stages So was it when my life began; and by adding or deleting stages. So is it now I am a man; A history of research on life cycle evolution So be it when I shall grow old, was published by S.J. Gould in the book Or let me die! Ontogeny and Phylogeny (1977). Gould hand- The child is father of the man; And I could wish my days to be ily summarizes the mechanisms for biologi- Bound each to each by natural piety. cal change over time by stating, ‘‘Evolution occurs when ontogeny is altered in one of The term ‘‘neoteny’’ was coined by Julius two ways: when new characters are intro- Kollman in 1885 to describe the sexual duced at any stage of development with maturation of the axolotl, a urodele amphib- varying effects upon subsequent stages, or ian (salamander), while still in its aquatic, when characters already present undergo gill breathing stage of development. Ashley changes in developmental timing. Together, Montagu (1989) states that Kollman in- these two processes exhaust the formal con- tended neoteny to mean ‘‘retaining youth’’ tent of phyletic change.’’ Gould contends but that Kollman confused the Greek word that it is the second process that accounts teinein (to stretch) with the Latin word for human evolution. This process is called tenere (to retain). Figure 3 presents contem- heterochrony. Quoting Gould again, ‘‘this porary dictionary definitions, as well as an- book is primarily a long argument for the other current medical usage of the Latin evolutionary importance of heterochrony— root of the term. Despite the scatological changes in the relative time of appearance humor that may be found in Kollman’s ety- and rate of development for characters al- mological error, Montagu and Gould believe ready present in ancestors’’ (author’s italics). that Kollman had the right idea, that is, Gould explains that there are several types neoteny is the process for human evolution. of heterochronic processes, but only one The idea that neoteny is the primary accounts for human evolution. This is neo- process for humanization was first formal- teny, defined in the glossary of Gould’s book ized scientifically by Louis Bolk in 1926 (see as ‘‘paedomorphosis (retention of formally Montagu, 1989, for a concise review of his- juvenile characters by adult descendants) torical sources). Gould acknowledges that produced by retardation of somatic develop- much of Bolk’s neoteny is really an argu- ment.’’ In a subsequent publication Gould ment for scientific racism and sexism. Never- provides a somewhat more readable defini- theless, Gould tries to retain the baby of tion: ‘‘In neoteny rates of development slow neoteny while discarding the bath water of Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 67 with our ancestor, but retain the shape of an immature ancestor. Gould makes very clear his interpretation of the consequences of neoteny. ‘‘If humans evolved, as I believe, by neoteny,...then we are, in a more than metaphorical sense, permanent children’’ (1981, p. 333). Could it be true that we ‘‘permanent children’’? Some eminent scholars believe it. Benjamin Frank- lin wrote, ‘‘Our whole life is but a greater and longer childhood.’’ Sigmund Freud was more circumspect: ‘‘In our innermost soul we are children and remain so for the rest of our lives.’’ Montagu is the most emphatic: ‘‘We are intended to remain in many ways child- like, we were never intended to grow up into the kinds of adults we have become....Our uniqueness lies in always remaining in a state of development’’ (all quotes from Mon- tagu, 1989). Given the scientific respectability of works on neoteny by Gould and Montagu, the term Fig. 3. Definitions and etymology of the term ‘‘neo- and the concept have been popularized in teny.’’ western society and adopted into many other arenas. The following are some recent ex- amples. Philosophers posit neoteny as the racist and sexist science. To do so, Gould process leading to the human capacity for suggests that the major difference between language (Goldsmith, 1993; Brown, 1995). human and non-human primate growth is Psychiatrists allege it is the reason humans that humans mature sexually while still in are so playful (Brown, 1995). Some types of an infantile or juvenile stage of physical neurological dementia are described as ‘‘a development. failure of neoteny’’ (Bemprad, 1991). Human Gould expressed his ideas on hetero- female sexual attractiveness is claimed to be chrony via neoteny in terms of a ‘‘clock a function of neoteny (Jones, 1995). Taking model.’’ The clock has two hands, one that this cue, the advertising industry employs calibrates a size arc and the other that human infants, children, and neotenous measures a shape arc. The clock sits above a adult women to sell all sorts of products. type of bar-shaped calendar that records Moreover, the product designers make the time in the form of biological age, that is, age inanimate objects (everything from tea pots at sexual maturation. The clock may be used to automobiles) more appealing and purchas- to compare an ancestor with its descen- able by making them seem neotenous, that dants. The ancestor’s clock is fixed with the is, shaped like a human infant or child two hands at the zenith of the arcs and age (Boym, 1994, and see Box 1). Finally, a of sexual maturation at the mid-point of the prominent social anthropologist claims that calendar. The descendant’s clock measures human evolution via neoteny allows for the changes in size, shape, and age at matura- development of human culture, especially tion by differences in these three vectors. religion (La Barre, 1991)! The major point that Gould emphasizes HYPERMORPHOSIS AND HUMAN with the clock model is that size, shape, and EVOLUTION maturation are disassociable from each other and can evolve independently. Figure 4a is Despite the intellectual weight that biol- Gould’s clock for human evolution via neo- ogy (Gould), social philosophy (Franklin), teny. Humans have larger size (body, brain, psychoanalysis (Freud), anthropology (Mon- etc.) and mature at a later age, compared tagu), and advertising bring to this issue, 68 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997 neoteny clock, the hypermorphosis clock pos- its that humans have larger size, later matu- ration, and a more ‘‘mature’’ shape than our putative ancestor. Although McKinney and McNamara do admit that there are a num- ber of non-hypermorphic features of humans ‘‘hypermorphosis seems to best explain most of those traits that make us human: large body size, large brain, long learning stage and life span’’ (p. xi). According to heterochrony, humans are not permanent children; rather we are devel- opmentally delayed and/or growth-prolonged apes. Adolph Schultz (1960) proposed ex- actly this more than three decades ago. Figure 5 is his classic illustration of the progressive delay in the onset and offset of primate life stages. Hypermorphosis, there- fore, is hardly a new idea, but is it the right idea? Sue Taylor Parker (1996) thinks that hy- permorphosis is the right idea, and she uses it to account for the evolution of human cognitive capacities. Parker applies Piaget’s stage theory (Piaget and Inhelder, 1969) to Fig. 4. a: Clock model for human evolution by neo- analyze the cognitive development of human teny. b: Clock model for human evolution by hypermor- phosis. Adapted from Godfrey and Sutherland (1996), and non-human primates. She finds that all with permission. primates share the first one or two stages (sensorimotor and preoperational). Only the human species progresses to the higher Pia- getian stages (concrete operational and for- and despite the popularity of neoteny as a cause ce´le`bre for so many human traits and mal operations). Parker explains that the behaviors, the proposition that adult hu- uniquely human stages were added by hyper- mans are permanent children is not ac- morphosis, that is, by extending the period cepted by all scholars. Michael McKinney of cognitive development past that of the and Kenneth McNamara (1991), in their monkeys and apes. This negates neoteny, book Heterochrony: The Evolution of Ontog- which in Parker’s view would mean that eny, provide a detailed case against neoteny human cognitive capacities evolved by halt- in human evolution. McKinney and McNa- ing mental development at early or interme- mara argue instead for another type of diate stage of non-human primate develop- heterochronic process to account for human ment. To quote Parker, ‘‘Cognitively, humans growth and evolution, namely hypermorpho- are overdeveloped rather than underdevel- sis. They state their position as follows: oped apes’’ (p. 377). ‘‘Neoteny is the process of growing slower. Elizabeth Vrba (1996) agrees that hyper- Yet humans do not grow particularly slow morphosis is the key process in human (relative to either chimp or our ances- evolution, but she adds a twist to McKinney tors. . .). What we do is delay the offset of and McNamara’s argument. Vrba’s model of virtually all developmental events (growth human evolution is set in the more general phases) so that each phase is longer. This is context of mammalian evolution in Africa. hypermorphosis’’ (p. xi). Figure 4b illus- Vrba tries to show that climate change, trates evolution by hypermorphosis using specifically global cooling after 2.9 million Gould’s clock model. Compared with the years ago, resulted in an enlargement of Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 69 of growth prolongation, or time hypermor- phosis, as it acts on characters with different ancestral growth profiles in the same body plan.’’ Vrba continues by stating that this ‘‘can result in a major reorganization—or ‘shuffling’—of body proportions such that some characters become larger and others smaller, some hyperadult and others more juvenilized’’ (1996, p. 1). Many of the ‘‘ances- tral growth profiles’’ of the hominids are, according to Vrba, still to be found in the great apes. She states, ‘‘I do not imply that the chimpanzee itself is ancestral, but only that its growth profile resembles that of the common ancestor’’ (1996, p. 17). Vrba pre- dicts that by maintaining chimpanzee growth rates for legs, arms, torso, skull, brain, etc., and prolonging the total time for growth it is possible to derive a modern human morphology from an African pongid morphology.

CRITIQUES OF NEOTENY AND HYPERMORPHOSIS The concepts of neoteny and hypermorpho- sis posit that modern humans evolved by either a delay or an extension of ancestral patterns of growth. Human anatomy, physi- ology, and behavior, however, cannot be ex- plained by either of these simple processes. Fig. 5. Schultz’s diagram of the proportional in- crease in the length of life stages across the scala Criticisms of the application of neoteny to naturae of living primates. Note that Schultz did not human evolution are not new. Kummer recognize the childhood stage for humans. Indeed, all (1953), Bertalanffy (1960), and Starck and species have the same life stages, which just increase in length from prosimian to human. The estimates for total Kummer (1962) argued against neoteny, or length of life are based on average expectations rather what they called fetalization, on the basis of than theoretical maximums. The data for ‘‘Early Man’’ human craniofacial and postcranial growth: are entirely speculative as no species is given and very little data were available when Schultz prepared this ‘‘. . . the concept of ‘fetalization’ is to be re- figure. Adapted from Schultz (1960) and Smith et al. fused with respect to the ontogeny and evolu- (1994). tion of the human . . . for this is not the result of an arrest of growth at an early phase but of a differentiation in changed body size along with a relative decrease in direction....Hence we may speak of ‘retar- limb length for several species of African dation’ but not of fetalization in human mammals. Such morphological change in development’’ (Bertalanffy, 1960, p. 250). response to a cooler climate accords with The phrase to emphasize in this quote is Bergmann’s and Allen’s rules. According to ‘‘differentiation in changed direction.’’ Vrba, the African hominids alive at that One specific example of the failure of time also conform to these rules, but the neoteny and hypermorphosis to account for hominids are notable for a concurrent in- the direction of human development is the crease in brain size relative to body size. The ontogeny of cranial growth related to lan- general mammalian and the specifically guage development. The human newborn hominid morphological change is explained cannot produce the speech sounds (pho- to result from ‘‘the same evolutionary event nemes) used by adult speakers of any lan- 70 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997 guage. Lieberman et al. (1972) and Laitman the rest of the body did not evolve by neo- and Heimbuch (1982) believe that the shape teny. In Shea’s view, a variety of hetero- of the basicranium is the reason for this. chronic processes are responsible for human They argue that newborns possess a basicra- evolution. The others may be hypermorpho- nium with a relatively large angle of flexion sis, acceleration (defined as an increase in (the angle formed by the junction of the the rate of growth or development), and occipital and vomer bones). This angle influ- hypomorphosis (defined as a delay in growth ences the shape of the soft tissues of the with no delay in the age at maturation). In vocal track, especially the pharynx, and Figure 6 are illustrated Shea’s estimates for determines the nature of the vocal sounds body size and shape as a consequence of the newborn can produce. During growth neoteny and two types of hypermorphosis. the angle of flexion becomes more acute, as None of these acting as a single process can the skull assumes child, juvenile, adoles- produce the human adult size and shape cent, and, finally, adult proportions. As this from the human infant size and shape. The growth process takes place a greater range same holds true for acceleration and hypo- of linguistically recognizable phonemes is morphosis. In agreement with Schultz, Shea produced. In this one aspect of growth of the states that ‘‘we [humans] have extended all skull and its functional correlates neither of our life history periods, not merely the neoteny nor hypermorphosis is a useful con- embryonic or juvenile ones’’ (pp. 84–5). Hu- cept. Non-human primates never possess mans have also altered rates of growth from the human type of basicranial anatomy at those found in other primates and possible any stage of their development, nor would ancestors. To accomplish all this required, in they develop this anatomy if they prolonged Shea’s view, several genetic changes or ad- any of their stages of growth. As a conse- justments during human evolution. Since quence, non-human primates cannot pro- the hormones that regulate growth and de- duce human-like phoneme sounds. velopment are, virtually, direct products of A graphic case against neoteny and hyper- DNA activity, Shea proposes that the best morphosis can be made by considering the place to look for evidence of the evolution of ontogeny of human body proportions. From ontogeny is in the action of the endocrine fetus, to child, and to adult, human body system. According to Shea and others (e.g. proportions are so much altered that the Bogin, 1988) differences in endocrine action mature morphology cannot be simply pre- between humans and other primates negate dicted from earlier stages of growth. Allom- neoteny or hypermorphosis as unitary pro- etry, differential rates of growth of parts of cesses and instead argue for a multiprocess the body relative to that of the body as a model for human evolution. whole (Huxley, 1932), is the rule in primate, In the male chimpanzee, for example, the including human, development. Both posi- concentration of testosterone in blood serum tive and negative allometry take place in the prior to puberty (from 1 to 6 years of age) ontogeny of human development (e.g., leg averages 13 ng/dl (Martin et al., 1977). For vs. trunk growth, or head vs. body growth). the human male, the prepubertal serum These allometric changes bring about func- testosterone concentration (from ages 1 to tional differences between the adult and 12 years) averages 9 ng/dl (Winter, 1978). child in physical appearance and perfor- The peak velocity in long bone growth of mance. eight male chimpanzees studied by Watts Brian Shea (1989) published the most and Gavan (1982) occurred at a mean age of cogent allometric analysis to date that re- 10.96 years, with a standard deviation of jects both neoteny or hypermorphosis as a 1.31 years. At this age, serum testosterone ‘‘grand unification theory’’ for all of human averages about 400 ng/dl (Martin et al., growth and evolution. Shea is not anti- 1977). Peak height velocity in human boys neoteny or anti-hypermorphosis per se, and from western Europe occurs at a mean age of in fact he allows that the human brain and 14.06 years, with a standard deviation of cranium may have evolved by neoteny. But 0.92 years (Marshall, 1978), when serum he argues that the human face, jaws, and testosterone levels average about 340 ng/dl Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 71 formed a longitudinal analysis of the growth of individual limb segments in British boys. It was found that in contrast to the negli- gible velocity change of chimpanzees, the peak value in velocity during the human adolescent growth spurt ranged between 1.34 cm per year for the forearm and 2.44 cm per year for the tibia. From these findings one may propose that there are differences in the effect of testosterone on the skeletal growth of chimpanzees and humans. The growth response of the human skeleton to rising testosterone levels is greater than that of the chimpanzee skeleton, since the change in the serum hormone levels of the two primates differs only by a factor of about 1.23, but the change in the velocity of growth differs by a factor of at least 2.0 and as much as 4.9. The heterochronic models of neoteny and hypermorphosis predict that a greater or lesser amount of time for growth produces the differences in size and shape between humans and chimpanzees. But these empiri- cal data gathered from growth and endo- crine research show that it is the sensitivity of specific skeletal parts to testosterone, Fig. 6. Silhouettes of size and shape change during determined by DNA and cellular activity, human growth. Numbers under silhouettes indicate age in years. Numbers above or on silhouettes indicate that results in the differences in limb size relative shape. Top left: Actual size and shape change and shape between adult humans and chim- during normal human development. Top right: Neo- panzees. The time available for growth is teny. Note that at adult size shape 3 is still maintained. Bottom left: Time hypermorphosis. The growth period largely irrelevant. is extended to 36 years, yielding a peramorphic giant A closer inspection of human cognitive (size and shape of the descendant beyond that of the ancestor). Bottom right: Rate hypermorphosis. Growth development also fails to support hetero- ends at age 26 but proceeds at a faster rate, producing chrony, especially Parker’s hypermorphosis another peramorphic giant. Note that in both cases the argument. Human cognitive ontogeny from adult shape at 7ϩ is outside the range of normal development. Redrawn from Shea (1989), with permis- childhood to adulthood entails the develop- sion. ment of new competencies, such as those related to language and social intelligence. The stage theory of Piaget (1954; Piaget and (Winter, 1978). Thus, the serum testoster- Inhelder, 1969) provides a descriptive and one concentration of the chimpanzee in- theoretical understanding of the develop- creases about 31-fold from the prepubertal ment of human intelligence. The neoteny to pubertal state. In the human male, serum and hypermorphosis arguments correctly as- testosterone concentration increases about sert that adult humans possess an intellec- 38-fold, or 1.23 times the increase for the tual plasticity and curiosity usually found chimpanzee. only in the young of other species. Piaget’s According to Watts and Gavan (1982), stage theory of human development shows, chimpanzees have a relatively small in- however, that this is due to the maturation crease in the velocity of growth of individual of increasingly sophisticated and flexible long bones during puberty, ‘‘usually less cognitive processes, rather than to the reten- than a centimeter’’ (p. 58), and often less tion or extension of infantile or juvenile than 0.5 cm. Cameron et al. (1982) per- intellectual abilities. At the physiological 72 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997 level, Eccles (1979) shows that the adult sequent stages;’’ the second is by hetero- human potentials for playfulness, creativity, chrony. Much of human evolution, especially and intellectual advancement are not de- the evolution of childhood, is the result of rived via heterochrony; rather they are new the introduction of new life stages into the competencies derived from a constant remod- general pattern of primate growth and devel- eling, restructuring, and maturation of the opment. neurological architecture in the central ner- Consider first the general mammalian vous system. and primate patterns of growth and then The last word on heterochrony, so far at compare these with the human pattern. The least, is the work of Godfrey and Sutherland majority of mammals progress from infancy (1996). These researchers do not argue for or to adulthood seamlessly, without any inter- against one or the other of the many hetero- vening stages, and while their growth rates chronic mechanisms that may have influ- are in decline (Brody, 1945; Bertalanffy, enced human evolution. Instead, they de- 1960). This pattern of postnatal growth is velop a quantitative method for the fair illustrated in Figure 7 using data for the testing of any or all heterochrony hypoth- mouse. Highly social mammals, such as eses. Their methodology is a major innova- wolves, wild dogs, lions, elephants, and the tion in heterochrony research. They first primates, postpone puberty by inserting a show that Gould’s clock model is based on period of juvenile growth and behavior be- three mathematically definable vectors for tween infancy and adulthood (Bekoff and size, shape, and time. Godfrey and Suther- Beyers, 1985; Pereiera and Fairbanks, 1993). land then show that movement of the clock’s Juveniles may be defined as ‘‘prepubertal hands or age bar may be represented as linear vector distortions. The translation of individuals that are no longer dependent on the clock model to linear vectors makes their mothers (parents) for survival’’ (Pereira quantification relatively easy, and, more im- and Altmann, 1985, p. 236). This definition portantly, allows for the prediction of exact is derived from ethological research with differences between ancestor and descen- social mammals, especially non-human pri- dent species for the different types of het- mates, and applies to the human species as erochrony. The authors carry out several well. Juveniles are not children, as juveniles prediction analyses and find that many of are independent but children still require the assertions made in the past about the care from older individuals. In the highly role of heterochrony in human evolution are social mammals, puberty occurs while the not correct. In the end, the authors find no rate of growth is still decelerating and there support for any currently published het- is no readily detectable growth spurt in erochronic models for human evolution. skeletal dimensions (Fig. 8, but see Tanner Please read Godfrey and Sutherland’s ar- et al., 1990, who find that one sample of ticle for the details. captive rhesus monkeys have a rate of skel- etal growth at puberty that is greater than THE OTHER SIDE OF THE COIN that of human beings). In the tribulations and trials of hetero- Human growth and development from chrony the jury is currently out, and a birth to reproductive maturity may be char- verdict on what type of heterochrony has acterized by five stages: 1) infancy, 2) child- shaped which aspects of human growth and hood, 3) juvenility, 4) adolescence, and 5) development must await new research and adulthood (Bogin, 1988, 1995; Bogin and testing. Forgotten by all parties in the litiga- Smith, 1996). Thus, humans add childhood tion surrounding heterochrony is another and adolescence to the pattern found for process by which evolution works. I requote primates and other highly social mammals. Gould, as he stated most succinctly that Each of the human stages of growth can be ‘‘evolution occurs when ontogeny is altered defined by clear biological and behavioral in one of two ways: [the first is] when new characteristics, especially those related to characters are introduced at any stage of rate of growth, feeding, and reproductive development with varying effects upon sub- behavior. Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 73

Fig. 7. Velocity curves for weight growth in the mouse. In both sexes puberty (vaginal opening for females or sper- matocytes in testes of males) occurs just after weaning and maximal growth rate. Weaning (W) takes place between days 15 and 20. After Tanner (1962).

Fig. 8. Baboon crown-rump length velocity. The let- and ends by about 6.0 years in both sexes. Redrawn with ters indicate the stages of growth: I, infancy; J, juvenil- some data smoothing from Coelho (1985). The patterns ity; M, sexual maturity. In the wild the weaning (W) of growth for other primate species, including chimpan- process begins as early as 4 months of age and ends by zees, are similar to these for the baboon (see Bogin, 12 to 18 months (Altmann, 1980). Puberty begins at 1988). about 3.5 years in females (&) and 4.5 years in males (()

RATE OF GROWTH Following childhood, the rate of growth decel- Changes in the velocity of growth from erates during the juvenile stage, and eventu- birth to adulthood signal the transitions ally the rate drops to its lowest point since between these five developmental stages. birth. The onset of adolescence is marked by Idealized velocity curves are presented in a sudden and rapid increase in growth rate, Figure 1. During infancy growth rate plum- which peaks at a level unequaled since early mets, followed by a period of slower velocity infancy. The mature adult stage begins when decline in childhood. The end of childhood is growth of the skeleton stops. often marked by a small increase in velocity, The mid-growth spurt is an important called the mid-growth spurt (Tanner, 1947). feature of human childhood. This spurt is 74 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997 associated with an endocrine event called adrenal androgens was not related to sexual adrenarche, the progressive increase in the maturation. Rhesus monkeys aged 1 to 3 secretion of adrenal androgen hormones. years, and not sexually mature, had the Adrenal androgens produce the mid-growth same high concentrations of all three adre- spurt in height, a transient acceleration of nal androgens as older, sexually mature bone maturation, and the appearance of monkeys. The same was true for baboons. In axillary and pubic hair, and seem to regulate contrast, chimpanzees 7 years old or older the development of body fatness and fat had adrenal androgen concentrations that distribution (Katz et al., 1985; Parker, 1991). were, on average, 4.7 times greater than There is a little story that links the mid- those for chimpanzees less than 4 years old. growth spurt with neoteny. Louis Bolk Thus, among the animals examined so far, (1926), the ‘‘father’’ of the scientific hypoth- the chimpanzee and the human are the only esis for human evolution via neoteny, specu- species that show adrenarche. Only humans lated that for our early human ancestors are known to have a mid-growth spurt. sexual maturation took place at about 6 to 8 The primate data reported by Cutler et al. years of age. The mid-growth spurt was first (1978) and Smail et al. (1982) suggest a reported by Backman (1934) and following possible function for adrenarche. Chimpan- its discovery, several of Bolk’s followers, zees and humans have low serum levels of without any additional supporting evidence, adrenal androgens after infancy and prior to opined that the mid-growth spurt and ad- adrenarche. Moreover, chimpanzees and hu- renarche are vestiges of sexual maturation mans have a relatively slow growth and a from our evolutionary past. Much research, long delay in the onset of sexual maturation. however, from clinical medicine to anthropo- Perhaps, then, the evolution of reduced adre- logical fieldwork, shows that there is little or nal androgen production prior to adrenarche no connection between adrenarche and may be explained as a mechanism that sexual maturation, as each are indepen- maintains slow epiphyseal maturation and dently controlled events (Smail et al., 1982; skeletal growth in the face of the prolonga- Weirman and Crowley, 1986; Worthman, tion of the prepubertal stages of growth. In 1986; Bogin, 1988, in press; Parker, 1991). the human, the delay is so protracted that it Bolk’s idea may be wrong, but a connec- becomes possible to insert the childhood tion with the evolution of the human pattern stage of development between the infancy of growth is still a possibility. The mecha- and juvenile stages. nism controlling adrenarche is not under- Synthesizing all of these data, it is pos- stood as no known hormone appears to sible to view the combination of adrenarche cause it. There are connections, however, and the human mid-growth spurt as life between the production of adrenal hor- history events marking the transition from mones, growth, and maturation. Cutler et the childhood to the juvenile stage of growth. al. (1978) and Smail et al. (1982) measured In terms of physical growth, the effects of the plasma concentration of the adrenal the adrenal androgens, to increase rate of androgens dehydroepiandosterone (DHA), skeletal growth, stimulate body hair growth, dehydroepiandosterone sulfate (DHAS), and and regulate body fat distribution, are short- delta4-androstenodine (D4) before and after lived and quite small. Even so, these physi- sexual maturation in 14 species. These spe- cal changes may be noticed by the child and cies include samples of rodents (rat, guinea his/her intimates, such as parents, and rec- pig, hamster), domestic animals (rabbit, dog, ognized as markers of developmental matu- sheep, pig, goat, horse, cow), primates (ma- ration. More to the point is the fact that caques—including 76 Macaca mulatta and while adrenarche may have only transient 80 M. nemestrina—a few baboons, and 52 effects on physical development, there is a chimpanzees), and the chicken. The plasma more permanent and important effect on concentrations of DHA, DHAS, and D4 were cognitive function. Psychologists have long significantly higher in sexually mature pri- been interested in what is called the ‘‘5 to 7 mates species than in any of the other year old shift’’ in cognition (White, 1965; animals. However, the serum level of these Rogoff et al., 1975; Whiting and Whiting, Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 75 1975; Weisner and Gallimore, 1977; Whiting ized societies provide a poor indication of and Edwards, 1988; Weisner, 1996). Weisner weaning age because bottle-feeding and the (1996, p. 295) states that ‘‘the 5–7 shift manufacture of ‘‘baby foods’’ allow either involves changes in internal states and com- early termination of breast-feeding or no petencies of the maturing child—shifts in breast-feeding at all. Preindustrialized hu- cognitive capacities, self concept, visual/ man societies provide a better indication of perceptual abilities, or social abilities. The the age at weaning, and hence the transition transition marks the emergence of increas- from infancy to childhood. One study finds ing capabilities for strategic and controlled that the termination of breast-feeding oc- self-regulation, skills at inhibition, the abil- curs at a median age of 36 months in these ity to maintain attention and to focus on a societies (Dettwyler, 1995). Another review complex problem, and planfullness and re- of such research (Lee et al., 1991) finds that flection.’’ Using the terminology of Piaget in so-called ‘‘food-enhanced’’ societies, those (Piaget and Inhelder, 1969) the 5–7 shift where nutritional intake is good, weaning moves the child from the preoperational to takes place as early as 9 months of age. In concrete operational stage of cognition. In ‘‘food-limited’’ societies, where chronic under- short, following the shift, the child becomes nutrition occurs, weaning takes place at 36 a juvenile capable of much of his or her own months. There are two fascinating corollar- feeding and care. This 5–7 year old shift is ies of this comparison. The first is that in found in all cultures so far investigated both the ‘‘food-enhanced’’ and the ‘‘food- (Rogoff et al., 1975), and thus seems to be a limited’’ societies the mean weight of weaned human species phenomenon. In addition to infants is about the same, 9.0 kg and 9.2 kg self-care and -feeding, the juvenile becomes respectively, or about 2.7 times birth weight increasingly involved with domestic work (Lee et al. assume a mean birth weight of and ‘‘caretaking interactions with other chil- 3,400 g for full-term humans). The second is dren’’ (Weisner, 1996, p. 296). The associa- that some solid foods are introduced into the tion of adrenarche, the mid-growth spurt, diet when the infant achieves about 2.1 and the 5–7 shift all seem to mark the times birth weight. Lee and colleagues (Lee progression of the child to the juvenile stage et al., 1991; Bowman and Lee, 1995) com- of development. pare the human data with data from 88 species of large-bodied mammals (32 non- FEEDING AND BREEDING human primates, 29 ungulates, 27 pinni- peds). They find that for all these species Changes in human feeding and reproduc- solid food is introduced, again, at about 2.1 tive behavior complement the pattern of times birth weight, but weaning takes place human growth velocity. The 5–7 shift is but when the infant achieves between 3.2 and one example of a change in feeding behavior 4.9 times birth weight. For all primates the from dependency to independence. There mean value is 4.6 times birth weight (range: are other important changes related to feed- 2.3 for Micropithecus talapoin to 9.4 for ing between birth, childhood, and the juve- Gorolla gorilla). The other great apes wean nile stage. at the following multiples of birth weight: Pan troglodytes, 4.9; P. paniscus, 6.1; Pongo Infancy pygmaeus, 6.4. As for all mammals, human infancy is the Thus humans are similar to other mam- period when the mother provides all or some mals in that we introduce solid foods at nourishment to her offspring via lactation. about 2.1 times birth weight. However, hu- The infancy stage ends when the young mans are unlike other mammals, even other mammal is weaned. Weaning is defined here species of primates, in that preindustrial as the termination of lactation by the mother and traditional societies, including ‘‘food- (other researchers may define weaning as limited’’ groups such as !Kung hunter- the process of shifting from lactation to gatherers, wean at a relatively early stage of eating solid foods). In human societies the growth—before reaching 3.0 times birth age at weaning varies greatly. Industrial- weight. 76 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997 Childhood dards for growth, the rural Thai children and juveniles were delayed in growth for all Childhood is the period following wean- the dimensions studied. Disease histories, ing, when the youngster still depends on parasite infestations, and mortality during older people for feeding and protection. Chil- infancy and childhood were not significantly dren require specially prepared foods due to associated with growth in this sample. Bai- the immaturity of their dentition. The de- ley et al. concluded that the delays in growth ciduous dentition, often called ‘‘milk teeth,’’ were not due to disease or the lack of specific have thin enamel and shallow roots com- nutrients, such as protein, vitamin A, or pared with the permanent dentition. Smith iron; rather the delays were due to a defi- (1991a) and Smith et al. (1994) report that ciency in the total intake of calories. The given this dental morphology, young mam- most dramatic falloff in growth occurred at mals with only the deciduous dentition can- 18 months of age, which corresponds with not process the adult-type diet. Smith the average age at weaning in these villages. (1991a) also finds that for virtually all mam- Weaning foods were usually watered-down mals the deciduous dentition is in place versions of adult foods. Although there were during infancy, that is, when the young are no food shortages in the villages, Bailey et nursing. With eruption and occlusion of the al. reason that the small gastrointestinal first permanent teeth the infant mammal tracts of the weaned infants and young moves either to adulthood (most mammals children may not have been capable of digest- have only two postnatal growth stages) or to ing enough food to meet their caloric de- the juvenile stage (social mammals). Smith mands for maintenance and growth of the reports that mammals with some perma- body. These two studies (and others re- nent teeth are able to process the adult diet viewed in Bogin, 1988) show that without and are independent in terms of feeding. the use of appropriate weaning foods chil- When human infancy ends the deciduous dren will suffer calorie insufficiency leading dentition is still in place; thus the human to undernutrition, developmental delays, and child still requires a special diet and re- growth retardation. mains dependent on older individuals for Another reason that children need a spe- food. cial high-energy diet is the rapid growth of Children also require a special diet due to their brain (Fig. 2). Leonard and Robertson the small size of their digestive tracts rela- (1992) estimate that due to this rapid brain tive to that of the adult (Bogin, 1988). This growth, ‘‘a human child under the age of 5 need can be appreciated most acutely when years uses 40–85 percent of resting metabo- it is not met. Behar (1977) studied the lism to maintain his/her brain [adults use causes of growth retardation and undernutri- 16–25 percent]. Therefore, the consequences tion of children living in rural villages in of even a small caloric debt in a child are Guatemala. He found that the children had enormous given the ratio of energy distribu- access to sufficient food, but the traditional tion between brain and body’’ (p. 191). In a adult diet of corn and beans did not have the related study, Leonard and Robertson (1994) caloric density to meet their growth require- also show that the size of the human brain ments. Bailey et al. (1984) studied the growth relative to total body size necessitates an of more than 1,000 infants, children, and energy-dense diet. At all stages of life after juveniles living in 29 villages in northern birth, human beings have brains that are Thailand. The participants in the study significantly larger than expected given the lived in rural agricultural villages, with rice human body size (Figs. 2 and 9). Aiello and as the basic subsistence crop. The partici- Wheeler (1995) refine the relationship be- pants were between the ages of 6 months tween human brain size and body size by and 5 years at the start of the study, and noting that compared with other mammals, they were measured about every 6 months including monkeys and apes, humans have for 5 years. It was found that, compared exceptionally small gastrointestinal tracts with local or international reference stan- (guts) relative to brain size. Aiello and Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 77

Fig. 9. Adult body weight and brain weight plotted for 61 species of Cercopithecidae (Old World monkeys), apes, and hu- mans. The curve is logarithmic regression fit to the data for all species. The illustration is a sagital section of the human brain. Each part of the human brain enlarged during evolu- tion, especially the size of cere- bral cortex (data from Harvey et al., 1987).

Wheeler show that both brain tissue and gut of growth of the brain (in weight). First tissue are ‘‘expensive,’’ meaning that both molar eruption takes place, on average, be- types of tissue have relatively high meta- tween the ages of 5.5 and 6.5 years in most bolic rates. They present estimates of the human populations (Jaswal, 1983; Smith, percentage of total body basal metabolic rate 1992). Functional occlusion occurs some for several tissues utilized by the typical 65 weeks to months thereafter. Recent morpho- kg adult human male. For the brain the logical and mathematical investigation value is 16.1 percent and for the gut the shows that brain growth in weight is com- value is 14.8 percent. Given these values, plete at a mean age of 7 years (Cabana et al., Aiello and Wheeler propose that during hu- 1993). Thus, significant milestones of dental man evolution the gut reduced in size as a and brain maturation take place at about 7 trade-off allowing, in part, for expansion of years of age. At this stage of development brain size. Both Leonard and Robertson the child becomes much more capable of (1994) and Aiello and Wheeler conclude that processing dentally an adult-type diet given this brain to body size/gut size relation- (Smith, 1991a). Furthermore, nutrient re- ship, adult humans require a diet that is quirements for the maintenance and the nutrient-dense, especially in energy, and easy to digest. Children, of course, have a growth of both brain and body diminish to relatively larger disproportion between brain less than 50 percent of total energy needs. size and body/gut size than do adults (Fig. Finally, the child shifts to a new plateau of 2). Added together, each of these constraints cognitive function. of childhood, an immature dentition, a small The child then progresses to the juvenile digestive system, and a calorie-demanding stage. As described above, ethnographic and brain that is both relatively large and grow- psychological research shows that juvenile ing rapidly, necessitates a special diet that humans have the physical and cognitive must be procured, prepared, and provided abilities to provide much of their own food by older individuals. and to protect themselves from environmen- tal hazards such as predation (children are Juvenility especially vulnerable to predation due to Important developments that allow chil- small body size) and disease (Weisner, 1987; dren to progress to the juvenile stage of Blurton-Jones, 1993). In girls, the juvenile growth and development are the eruption of period ends, on average, at about the age of the first permanent molars and completion 10. This is 2 years before it usually ends in 78 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997 boys, the difference reflecting the earlier of mathematical functions that are needed onset of puberty in girls. to describe growth. The distance and veloc- ity curves for most mammalian species can Adolescence be estimated by a single function, such as a Human adolescence begins with puberty, simple polynomial or exponential function. marked by some visible sign of sexual matu- Even the monkeys and apes, with the addi- ration such as increased density of pubic tion of the juvenile stage, require no more hair (indeed the term is derived from the than two such relatively simple functions Latin pubescere: to grow hairy). The adoles- (Laird, 1967; Bogin, 1988). The insertion of cent stage also includes development of the the mid-childhood and adolescent spurts secondary sexual characteristics and the into human ontogeny means that at least onset of interest and activity in adult pat- three mathematical functions are needed to terns of sociosexual and economic behavior. adequately describe shape of the velocity These physical and behavioral changes of curve (Fig. 10). Not only more, but also more puberty occur in many species of mammals. complex functions are needed (Bock and What makes human adolescence different is Thissen, 1976; Karlberg, 1985). It is vitally that during this life stage both boys and important to stress here that all of this girls experience a rapid acceleration in the quantitative knowledge of the biology of growth of virtually all skeletal tissue—the human growth is well established and widely adolescent growth spurt. Human adoles- available. This information unequivocally cence and the growth spurt are treated in negates neoteny, hypermorphosis, or any detail in other publications (Bogin, 1988, other single heterochronic process as the 1994a, 1995; Bogin and Smith, 1996). primary process of human evolution. Lamen- tably, the works on neoteny cited above— Adulthood with the exception of Shea’s work—make Adolescence ends and early adulthood be- little or no reference to studies of the physi- gins with the completion of the growth spurt, cal growth and development of living people. the attainment of adult stature, the comple- tion of dental maturation (eruption of the HOW AND WHEN DID THE HUMAN third molar, if present), and the achieve- PATTERN EVOLVE? ment of full reproductive maturity (Figs. 1 The stages of the life cycle may be studied and 2). The latter includes both physiologi- directly only for living species. However, cal, socioeconomic, and psychobehavioral at- there are lines of evidence on the life cycle of tributes, which all coincide, on average, by extinct species. Such inferences for the about age 19 in women and 21 to 25 years of hominids are, of course, hypotheses based age in men (Bogin, 1988, 1993, 1994a). That on comparative anatomy, comparative physi- is, these are the median ages at first birth ology, comparative ethology, and archaeol- for women or age at fatherhood for men on a ogy. Examples of this methodology are found worldwide basis for both contemporary and in the work of Martin (1983) and Harvey et historic populations. al. (1987) on patterns of brain and body growth in apes, humans, and their ances- EMPIRICAL AND MATHEMATICAL tors. EVIDENCE AGAINST HETEROCHRONY Apes have a pattern of brain growth that Clearly, the human pattern of growth is rapid before birth and relatively slower from birth to maturity is qualitatively and after birth. In contrast, humans have rapid quantitatively different from the pattern for brain growth both before and after birth other primates. The quantitative differences (Fig. 11). This difference may be appreciated can be expressed in amounts, rates, and by comparing ratios of brain weight divided timing of growth events, and are so reported by total body weight (in grams). At birth this in many standard textbooks of human ratio averages 0.09 for the great apes and growth (e.g., Tanner, 1962; Bogin, 1988). 0.12 for human neonates. At adulthood the These quantitative differences may also be ratio averages 0.008 for the great apes and expressed in terms of the type and number 0.028 for humans. In other words, relative to Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 79

Fig. 10. Idealized velocity curve of human growth for boys (solid line): I, infancy; C, childhood; J, juvenility; A, adolescence; M, mature adult. The dashed line is a sixth-degree polynomial curve fit to the velocity curve data. The polynomial curve does not fit well to real growth data due to the pulses of the mid-childhood spurt (MCS) and the adolescent spurt (AS). The human velocity curve cannot be fit adequately by a single continuous mathematical function. At a minimum, three functions are required. body size human neonatal brain size is 1.33 mean rate of postnatal brain growth for times larger than the great apes, but by living apes, an 850-cc adult brain size may adulthood the difference is 3.5 times. The be achieved by all hominoids, including ex- human-ape difference is not due to any tinct hominids, by lengthening the fetal single heterochronic process, that is, not the stage of growth. At brain sizes above 850 cc result of delay, prolongation, or acceleration the size of the pelvic inlet of the fossil of a basic ape-like pattern of growth. Rather hominids, and living people, does not allow it is due to new patterns of growth for the for sufficient fetal growth. Thus, a period of human species. The rate of human brain rapid postnatal brain growth and slow body growth exceeds that of most other tissues of growth—the human pattern—is needed to the body during the first few years after reach adult brain size. birth (Fig. 2). Martin (1983) and Harvey et Martin’s analysis is elegant and tenable. al. (1987) also show that human neonates Nevertheless, the difference between ape have remarkably large brains (corrected for and human brain growth is not only a body size) compared with other primate matter of velocity; it is also a matter of life species. Together, relatively large neonatal history stages. Brain growth for both apes brain size and the high postnatal growth rate give adult humans the largest encepha- and humans ends at the start of the juvenile lization quotient (an allometric scaling of stage, which means that apes complete brain brain to body size) of all higher primates growth during infancy. Humans, however, (Fig. 9). insert the childhood stage between the in- Martin (1983) hypothesizes that a ‘‘human- fant and juvenile stages. Childhood may like’’ pattern of brain and body growth be- provide the time and the continuation of comes necessary once adult hominid brain parental investment, necessary to grow the size reaches about 850 cc. This biological larger human brain. Following this line of marker is based on an analysis of cephalo- reasoning, any fossil human, or any of our pelvic dimensions of fetuses and their moth- fossil hominid ancestors, with an adult brain ers across a wide range of social mammals, size above Martin’s ‘‘cerebral Rubicon’’ of including cetaceans, extant primates, and 850 cc may have included a childhood stage fossil hominids (Martin, 1983). Given the of growth as part of its life history. 80 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997

Fig. 11. Growth curve for human brain and body postnatal phase. In contrast, the rate of human body compared to that of the chimpanzee. The length of the growth slows after birth. If the human brain/body human fetal phase, in which brain and body grow at the growth rate were equal to the chimpanzee rate, then same rate for both species, is extended for humans. adult humans would weigh 454 kg and stand nearly 3.1 Chimpanzee brain growth slows after birth, but humans m tall (indicated by the Q). After Martin (1983). maintain the high rate of brain growth during the

Given this background, Figure 12 is my shares many anatomical features with non- Schultz-inspired summary of the evolution hominid pongid species including an adult of the human pattern of growth and develop- brain size of about 400 cc (Simons, 1989) and ment from birth to age 20 years (the evolu- a pattern of dental development indistin- tion of adolescence is not discussed in this guishable from extant apes (Smith, 1991b). article, but see Bogin, 1993, 1994a, and Therefore, the chimpanzee and A. afarensis Bogin and Smith, 1996). Figure 12 must be are depicted in Figure 12 as sharing the considered as ‘‘a work in progress,’’ as only typical tripartite stages of postnatal growth the data for the first and last species (Pan of social mammals—infant, juvenile, adult and Homo sapiens) are known with some (Pereira and Fairbanks, 1993). To achieve certainty. The patterns of growth of the the larger adult brain size of A. africanus fossil hominid species are reconstructions (442 cc) may have required an addition to based on published analyses of skeletal and the length of the fetal and/or infancy peri- dental development of fossil specimens that ods. The rapid expansion of adult brain size died before reaching adulthood (see Smith, during the time of Homo habilis (650 to 800 1991b, 1992; Bogin and Smith, 1996, for cc) might have been achieved with further reviews of this literature). One major mes- expansion of both the fetal and infancy sage to take from the figure is that the periods, as Martin’s ‘‘cerebral Rubicon’’ was prolongation of the total time for growth not surpassed. However, the insertion of a that plays such a prominent role in the brief childhood stage into hominid life his- hypotheses of neoteny and hypermorphosis tory may have occurred. This conjecture is is definitely a part of human evolution. based on a comparison of human and ape However, time prolongation is not sufficient reproductive strategies. to account for the insertion of the new stages There are limits to amount of delay be- of childhood and adolescence that are part of tween birth and sexual maturity, and be- human growth. tween successful births, that any species can Australopithecus afarensis is considered tolerate. The great apes are examples of this by most paleontologists to be a hominid, but limit. Chimpanzee females in the wild reach Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 81

Fig. 12. The evolution of hominid life history during the first 20 years of life. Abbrevi- ated nomenclature as follows: A. afar., Australopithecus afarensis; A. africa., Australopi- thecus africanus; H. habilis, Homo habilis; H. erec. 1, early Homo erectus; H. erec. 2, late Homo erectus; H. sapiens, Homo sapiens. menarche (the first menstruation) at 11 to The great apes, and fossil hominids such 12 years of age and have their first births at as Australopithecus or early Homo, seem to an average age of 14 years (Goodall, 1983). have reached this demographic dilemma by The average period between successful births extending infancy, forcing a demand on nurs- in the wild is 5.6 years and young chimpan- ing to its limit (Fig. 12). Extending infancy zees are dependent on their mothers for and birth intervals beyond the chimpanzee/ about 5 years (Teleki et al., 1976; Goodall, Australopithecus range may not have been 1983; Nishida et al., 1990). Actuarial data possible if hominids such as Homo habilis collected on wild-living animals indicate that were to remain extant. Early Homo may between 35 percent (Goodall, 1983) and 38 have overcome this demographic dilemma percent (Nishida et al., 1990) of all live-born by reducing the length of infancy and insert- chimpanzees survive to their mid-20s. Al- ing childhood between the end of infancy though this is a significantly greater percent- and the juvenile period. Free from the de- age of survival than for most other species of mands of nursing and the physiological brake animals, the chimpanzee is at a reproduc- that frequent nursing places on ovulation tive threshold. Goodall (1983) reports that (Ellison, 1990), mothers could reproduce soon for the period 1965 to 1980 there were 51 after their infants became children. This births and 49 deaths in one community of certainly occurs among modern humans. An wild chimpanzees at the Gombe Stream often cited example, the !Kung, are a tradi- National Park, Tanzania. During a 10-year tional hunting and gathering society of period at the Mahale Mountains National southern Africa. A !Kung woman’s age at her Park, Tanzania Nishida et al. (1990) ob- first birth averages 19 years and subsequent served ‘‘74 births, 74 deaths, 14 immigra- births follow about every 3.6 years, resulting tions and 13 emigrations’’ in one community. in an average fertility rate of 4.7 children Chimpanzee population growth is, by these per woman (Howell, 1979; Short, 1976). data, effectively equal to zero. Galdikas and Women in another hunter-gather society, Wood (1990) present data for the orangutan the Hadza (Blurton-Jones et al., 1992), have that show that these apes are in a more even shorter intervals between successful precarious situation. Compared with the 5.6 births, stop nursing about 1 year earlier, and years between successful births of chimpan- average 6.15 births per woman. For these zees, the orangutan female waits up to 7.7 reasons a brief childhood stage for Homo years, and orangutan populations are in habilis is indicated in Figure 12. decline. Lovejoy (1981) calls the plight of Further brain size increase occurred dur- great ape reproduction a ‘‘demographic di- ing H. erectus times. The earliest adult lemma’’ (p. 211). specimens have brain sizes of 850 to 900 cc. 82 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997 This places H. erectus at or above Martin’s explanations that emphasize learning, an ‘‘cerebral Rubicon’’ and seems to justify inser- idea that goes back at least to Spencer tion and/or expansion of the childhood pe- (1886), and even further back, to the dawn of riod to provide the biological time needed for written history (Boyd, 1980). These three the rapid, human-like, pattern of brain traditional explanations are that childhood growth. It should be noted from Figure 12 provides for: 1) an extended period for brain that the model of human evolution proposed growth; 2) time for the acquisition of techni- here predicts that from the Australopithecus cal skills, e.g. toolmaking and food process- to the H. erectus stage the infancy period ing; and 3) time for socialization, play, and shrinks as the childhood stage expands. If the development of complex social roles and the infancy stage did in fact shrink, H. cultural behavior. erectus, and all later hominids, would have These reasons are valid inasmuch as they enjoyed an even greater reproductive advan- confer an advantage to preadult individuals. tage over all other hominids. Later H. erec- However, this brain-learning list of explana- tus, with adult brain sizes up to 1,100 cc, are tions cannot account for the initial impetus depicted with further expansion of child- for the insertion of childhood into human life hood and the insertion of the adolescent history. A childhood stage of development is stage. In addition to bigger brains, the ar- not necessary for the type of learning listed chaeological record for later H. erectus shows here. The prolonged infancy and juvenile increased complexity of technology (tools, period of the social carnivores (Bekoff and fire, and shelter) and social organization Beyers, 1985) and apes (Bogin, 1994b) can (Klein, 1989). These techno-social advances serve that function. Rather, childhood may are likely to be correlates of changes in be better viewed as a feeding and reproduc- biology and behavior associated with further tive adaptation for the parents of the child, development of the childhood stage of life as a strategy to elicit parental care after (Bogin and Smith, 1996). The evolutionary infancy, as a strategy to minimize the risks transition to archaic and finally modern H. of starvation for the child, a means of shift- sapiens expands the childhood stage to its ing the care of offspring from the parents, current dimension. especially the mother, to juveniles and older, postreproductive, adults (i.e., grandmoth- WHO BENEFITS FROM CHILDHOOD? ers), and as a mechanism that allows for Brain sizes of extant and fossil hominoids more precise ‘‘tracking’’ of ecological condi- provide some idea of when human life stages tions via developmental plasticity during may have evolved, but do not explain why the growing years. they evolved. Bonner (1965) shows that the Thus in addition to the three ‘‘textbook’’ presence of a stage, and its duration, in the explanations given above, there are at least life cycle relate to such basic adaptations as five additional reasons for the evolution of locomotion, reproductive rates, and food ac- childhood (these update or replace reasons quisition. To make sense out of the pattern 4–7 in Bogin, 1988), as follows. of human growth one must look for the 1) Childhood is a feeding and reproductive ‘‘basic adaptations’’ that Bonner describes. adaptation. A childhood growth stage may The most basic of these adaptations are have originally evolved as a means by which those that relate to evolutionary success. the mother, the father, and other kin could This is traditionally measured in terms of provision dependent offspring with food. This the number of offspring that survive and frees the mother from the demands of nurs- reproduce. Biological and behavioral traits ing and the inhibition of ovulation related to do not evolve unless they confer upon their continuous nursing. This decreases the inter- owners some degree of reproductive advan- birth interval and increases reproductive tage, in terms of survivors a generation or fitness. more later. Bogin (1988) lists seven reasons Figure 13 is a comparison of several life for the evolution of human childhood from history events for female great apes and the perspective of reproductive success. The humans. The data are drawn from studies of first three are the traditional ‘‘textbook’’ wild-living animals for the great apes and Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 83 non-contracepting hunting and gathering or societies, juvenile girls associate with these horticultural populations for humans (cita- working groups and the girls provide much tion in the figure legend). Based on this of the direct care and feeding of children, but figure it is possible to reiterate and always under the guidance of adolescents strengthen some points made above about and adults. In some societies fathers provide ape and human reproductive patterns. Note significant child care, including the Agta, that the infancy dependency period for each who take their children on hunting trips, of the apes species is longer than that for and the Aka Pygmies, a hunting-gathering humans. Ape infancy ends after eruption of people of central Africa (Hewlitt, 1991). Sum- the first permanent molar, which is probably marizing the data from many human societ- a requirement so that the juvenile ape can ies, Lancaster and Lancaster (1983) call this acquire and process foods of the adult diet type of child care and feeding ‘‘the hominid (Smith, 1991a). Human infancy ends before adaptation,’’ for no other primate or mam- eruption of the first permanent molar, that mal does this. is, before the youngster can process adult 2) A stimulus to release these parental foods. The evolution of childhood as a stage behaviors toward children may be found in in human life history ‘‘fills the gap’’ between the very pattern of growth of the children the infant’s dependency on the mother for themselves, in that the allometry of the food via nursing and the feeding indepen- growth of the human child releases nurtur- dence of the juvenile. Note also from Figure ing and care-giving behaviors in older indi- 13 that compared with the apes, the addi- viduals. The central nervous system, in par- tion of a childhood stage and the prolonga- ticular the brain, follows a growth curve tion of the juvenile and adolescent stages of that is advanced over the curve for the body development for humans delay the age of as a whole (Fig. 2). The brain achieves adult both menarche and first birth. However, size when body growth is only 40 percent compared with the great apes, humans have complete, dental maturation is only 58 per- the potential to reduce the birth spacing cent complete, and reproductive maturation interval and, therefore, as discussed above, is only 10 percent complete. The allometry of each woman may produce more offspring the growth of the human child maintains an during her life than any female ape. This infantile appearance (large cranium, small results in an increase in reproductive fitness face and body, little sexual development), if the additional offspring survive to matu- which stimulates nurturing and care-giving rity. !Kung, Hadza, and all human parents behaviors in older individuals. A series of help to ensure survival of their offspring by ethological observations (Lorenz, 1971) and provisioning all their children with food, not psychological experiments (Todd et al., 1980; just their current infant, for a decade or Alley, 1983) demonstrate that these growth longer. The child must be given foods that patterns of body, face, and brain allow the are specially chosen and prepared and these human child to maintain a superficially may be provided by older juveniles, adoles- infantile (i.e., ‘‘cute’’) appearance longer than cents, or adults. In Hadza society, for ex- any other mammalian species (see Box 1). ample, grandmothers and great-aunts are The infantile appearance of children facili- observed to supply a significant amount of tates parental investment by maintaining food to children (Blurton-Jones, 1993). In the potential for nurturing behavior of older Agta society (Philippine hunter-gathers) individuals toward both infants and depen- women hunt large game animals but still dent children (Bogin, 1988, 1990; McCabe, retain primary responsibility for child care. 1988). They accomplish this by living in extended 3) Children are relatively inexpensive to family groups—two or three brothers and feed. The relatively slow rate of body growth sisters, their spouses, children and parents— and small body size of children reduces and sharing the child care (Estioko-Griffin, competition with adults for food resources, 1986). Women of all ages work together in because slow-growing, small children re- food preparation, manufacture of clothing, quire less total food than bigger individuals. and child care. In all of these hunter-gather A 5-year-old child of average size (the 50th 84 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997

Fig. 13. Hominoid developmental landmarks. Data Gorilla, Gorilla gorilla; Chimp, Pan troglodytes; Hu- based on observations of wild-living individuals, or for man, Homo sapiens. Developmental landmarks are: human, healthy individuals from various cultures. Note Infancy/B.I., period of dependency on mother for sur- that compared with apes, humans experience develop- vival, usually coincident with mean age at weaning mental delays in eruption of the first permanent molar, and/or a new birth (B.I., birth interval); Molar 1, mean age at menarche, and age at first birth. However, age at eruption of first permanent molar; Menarche, humans have a shorter infancy and shorter birth inter- mean age at first estrus/menstrual bleeding; 1st birth, val, which in apes and traditional human societies are mean age of females at first offspring delivery. Sources: virtually coincident. The net result is that humans have Bogin (1988, 1994a), Galdikas and Wood (1990), Nishida the potential for greater lifetime fertility than any ape. et al. (1990), Smith (1992), Watts and Pusey (1993). Species abbreviations are: Orang, Pongo pygmaeus; centile of the NCHS reference curves for adults, especially the mother, for subsis- growth) and activity, for example, requires tence activity, adult social behaviors, and 22.7 percent less dietary energy per day for further childbearing. maintenance and growth than a 10-year-old 5) A further important reason for the juvenile on the 50th growth centile (Ulijas- evolution of childhood is that childhood al- zek and Strickland, 1993; Guthrie and Pic- lows for developmental plasticity. Following ciano, 1995). Thus, provisioning children, the discussion in Stearns (1992) and Mascie- though time consuming, is not as onerous a Taylor and Bogin (1995), the term plasticity task of investment as it would be, for in- means a potential for change in the pheno- stance, if both brain and body growth were type of the individual caused by a change in both progressing at the same rapid rate. the environment. The fitness of a given Moreover, in times of food scarcity children phenotype varies across the range of varia- are protected from starvation by this unique tion of an environment. When phenotypes pattern of brain and body growth. are fixed early in development, such as in 4) The task of child care becomes even less mammals that mature sexually soon after onerous because ‘‘babysitting’’ is possible. weaning (e.g., rodents), environmental Since children do not require nursing any change and high mortality are positively competent member of a social group can correlated. Social mammals (carnivores, provide food and care for them. Early neuro- elephants, primates) prolong the develop- logical maturity vs. late sexual maturity mental period by adding a juvenile stage allows juveniles and young adolescents to between infancy and adulthood. Adult pheno- provide much of their own care and also types develop more slowly in these mam- provide care for children (Bogin, 1994a). mals. They experience a wider range of Grandmothers and other postreproductive environmental variation, and the result is a women also provide much child care (Bogin better conformation between the individual and Smith, 1996). Again, this frees younger and the environment. Fitness is increased in Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 85 that more offspring survive to reproductive ance of the genus Homo in the last 2 million age than in mammalian species without a years, patterns of growth were shaped by juvenile stage (Lancaster and Lancaster, natural selection to promote and enhance 1983; Pereira and Fairbanks, 1993). Hu- parental investment. One way this was ac- mans insert the childhood stage between complished was by stimulating what may be infancy and the juvenile period. This results called the ‘‘psychology of parenting.’’ in an additional 4 years of relatively slow Lorenz (1971) stated that the physical physical growth and allows for behavioral characteristics of mammalian infants, in- experience that further enhances develop- cluding small body size, a relatively large mental plasticity. The combined result is head with little mandibular or nasal progna- increased fitness (reproductive success). Lan- thism, relatively large round eyes in propor- caster and Lancaster (1983) report that hu- tion to skull size, short thick extremities and mans in traditional societies, such as hunt- clumsy movements, inhibit aggressive be- ers and gatherers and horticulturalists, rear havior by adults and encourage their care- about 50 percent of their live-born offspring taking and nurturing behaviors. Lorenz be- to adulthood. In contrast, the Lancaster and lieved that these infantile features trigger Lancaster find that monkeys and apes rear ‘‘innate releasing mechanisms’’ in adult between 12 and 36 percent of live-born off- mammals, including humans, for the protec- spring to adulthood. The initial human ad- tion and care of dependent young. Gould vantage may seem small, but it means that (1979) questions the innateness of the hu- between 14 to 38 more people survive out of man response to infantile features. Such every 100 born—more than enough over the behavior may be ‘‘learned from our immedi- vast course of evolutionary time to make the ate experience with babies and grafted upon evolution of human childhood an overwhelm- an evolutionary predisposition for attaching ingly beneficial adaptation. ties of affection to certain learned signals’’ (p. 34). The important point is that whether CONCLUSION innate or learned the resultant behavior is These five themes of childhood—feeding, the same. nurturing, low cost, babysitting, and plastic- There seems to be a pan-human ability to ity—account for much of the evolution of and perceive the five stages of human postnatal pattern of growth of our species. Understand- development and respond appropriately to ing these themes helps to resolve the para- each. An elegant series of experiments per- dox of human growth and evolution—lengthy formed by Todd and his colleagues (1980) development and low fertility. In reality show that human perceptions of body shape humans raise a greater percentage offspring and growth status are consistent between to adulthood than any other species. These individuals. When adult subjects (about 40 successfully reared young adults then begin college students, all childless) were shown a their own reproduction and thus ensure series of profiles of human skull proportions, some ‘‘intimation of immortality’’ for their they could easily arrange them correctly parents. In the center of it all is human into a hierarchy spanning infancy to adult- childhood. For the child is indeed, to para- hood. The subjects could also ascribe matu- phrase Wordsworth, parent to the reproduc- rity ratings to skull profiles that were geo- tively successful and well-adapted adult. metrically transformed to imitate the actual changes that occur during growth (Fig. B1.1). BOX 1: THE EVOLUTIONARY This perception was selective because a vari- PSYCHOLOGY OF CHILDHOOD ety of other types of geometrical transforma- Reproductive success is the major force tions elicited no reports of growth or matura- behind the evolution of all species. Part of tion. When the growth-like mathematical the reproductive success of the human spe- transformations were applied to profile draw- cies is due to the intense investment and ings of the heads of birds and dogs, human care that parents, and other individuals, subjects reported identical perceptions of lavish on infants and children. In the course growth and maturation, even though in real- of human evolution, at least since the appear- ity the development of these animals does 86 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997

Fig. B1.1. Two of the mathematical transformations of human head shape used in the experiments of Todd et al. (1980). The middle profile in each row was drawn from the photograph of a 10-year-old boy. The transformations were applied to this profile of a real child. The cardioidal strain transformation is perceived by most adults as growth. The affine shear transformation is not perceived as growth.

Fig. B1.2. a: The series of five shape-variant drawings used in the experiments of Alley (1983). These drawings show the typical body proportions of a male at (from left) birth, 2, 6, 12, and 25 years of age. b: An example of the size-variant pairs of drawings used in the experiments of Alley (1983).

not follow the human pattern of skull-shape viduals. In the first experiment, subjects change. Even more surprising is that sub- were shown two sets of drawings. One set jects reported the perception of growth when was based upon two-dimensional diagrams the growth-like mathematical transforma- depicting changes in human body proportion tions were applied to front and side view during growth. Alley’s version of these dia- profiles of Volkswagen ‘‘beetles,’’ objects that grams are called ‘‘shape-variant’’ drawings do not grow. (Fig. B1.2a). Alley’s second set of figures In another series of experiments, Alley were called ‘‘size variant’’ drawings (Fig. (1983) studied the association between hu- B1.2b). He used the middle-most, ‘‘6-year- man body shape and size, and the tendency old’’ profile in the shape-variant series to by adults to protect and ‘‘cuddle’’ other indi- construct sets of figures that varied in height Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 87 TABLE B1.1. Mean (and standard deviation) reported ments developed in this chapter for the willingness to defend or cuddle persons of different body proportions (Alley, 1983) evolution of human childhood. In particular, small body size and a superficially infantile Age portrayed (years) Defend Cuddle appearance promote appropriate parental Newborn 7.7 (1.7) 3.4 (2.3) behavior by older individuals toward chil- 2 7.1 (1.7) 3.4 (2.0) 6 5.8 (1.8) 3.2 (2.2) dren. 12 5.0 (1.9) 3.0 (1.9) 25 4.3 (1.9) 2.7 (2.2) LITERATURE CITED Aiello LC and Wheeler P (1995) The expensive-tissue hypothesis. Curr. Anthropol. 36:199–221. and width, but not in shape. Note that these Alley TR (1983) Growth-produced changes in body shape and size as determinants of perceived age and adult figures have no facial features or genitals. caregiving. Child Dev. 54:241–248. Perceptual differences between figures are Altman J (1980) Baboon Mothers and Infants. Cam- bridge: Harvard University Press. due to body shape or size alone. Backman G (1934) Das Wachstum der Korperlange des In the first experiment, the subjects were Menchen. Kunglicke Svenska Verenskapsakademiens shown pairs of the shape-variant drawings Handlingar 14:145. Bailey SM, Gershoff SN, McGandy RB, Nondasuta A, (i.e., profiles of a newborn and a 6-year-old, a Tantiwongse P, Suttapreyasri D, Miller J, and McCree 2-year-old and a 12-year-old, etc.) or pairs of P (1984) A longitudinal study of growth and matura- the size-variant drawings and asked to state tion in rural Thailand. Hum. Biol. 56:539–557. Behar M (1977) Protein-caloric deficits in developing which one of the pair they ‘‘would feel most countries. Ann. N. Y. Acad. Sci. 300:176–187. compelled to defend should you see them Bekoff M and Beyers JA (1985) The development of being beaten.’’ In another experiment they behavior from evolutionary and ecological perspec- tives in mammals and birds. Evol. Biol. 19:215–286. were asked about their feelings to ‘‘hug or Bemprad JR (1991) Dementia preacox as a failure of cuddle’’ the person depicted. The results of neoteny. Theor. Med. 12:45–53. Bertalanffy L von (1960) Principles and theory of growth. both experiments, summarized in Table B1.1, In WN Nowinski (ed.): Fundamental Aspects of Nor- find a fairly strong reported willingness to mal and Malignant Growth. Amsterdam: Elsevier, pp. defend ‘‘newborns’’ and ‘‘2-year olds,’’ and a 137–259. Blurton-Jones NG (1993) The lives of hunter-gather moderate willingness to defend ‘‘older’’ per- children: Effects of parental behavior and parental sons. The reported willingness to cuddle reproductive strategy. In ME Pereira and LA Fair- decreased with the ‘‘age’’ of the drawings. banks (eds.): Juvenile Primates. Oxford: Oxford Uni- versity Press, pp. 309–326. Placed in the context of the ethological study Blurton-Jones NG, Smith LC, O’Connell JF, and Han- of parental caregiving in mammals and birds, dler JS (1992) Demography of the Hadza, an increas- ing and high density population of savanna foragers. Alley believes that his results demonstrate a Am. J. Phys. Anthropol. 89:159–181. general tendency to protect or cuddle others Bock RD, and Thissen DM (1976) Fitting multi- based on the perception of maturational component models for growth in stature. Proc. Ninth Int. Biometric Conf. 1:431–442. status. Bock RD and Thissen D (1980) Statistical problems of McCabe (1988) reviews the work of Alley fitting individual growth curves. In FE Johnston, AF and other similar studies. Taken together, Roche, and C Susanne (eds.): Human Physical Growth and Maturation, Methodologies and Factors. New these studies indicate that adults are more York: Plenum, pp. 265–290. likely to protect or nuture individuals with Bogin B (1988) Patterns of Human Growth. Cambridge: Cambridge University Press. ‘‘neotenous’’ facial features. McCabe defines Bogin B (1990) The evolution of human childhood. such features as having a relatively large BioScience 40:16–25. ratio of cranium size to lower face size. Bogin B (1993) Why must I be a teenager at all? New Scientist 137:34–38. McCabe also cites studies of the facial fea- Bogin B (1994a) Adolescence in evolutionary perspec- tures of nursery school–aged children under tive. Acta Paediatr. (Suppl.) 406:29–35. court protection for abuse compared with Bogin B (1994b) The evolution of learning. In T Husen and TN Postlethwaite (eds.): The International Ency- non-abused age-matched controls. The clopedia of Education, 2nd ed. Oxford: Pergamon, pp. abused children had smaller ratios of the 2681–2685. Bogin B (1995) Growth and development: Recent evolu- cranium/lower face—that is, they were less tionary and biocultural research. In N Boaz and LD ‘‘neotenous’’ or ‘‘cute’’—than the non-abused Wolfe (eds.): Biological Anthropology: The State of the controls. Science. Bend, OR: International Institute for Human Evolutionary Research, pp. 49–70. These psychological experiments and case Bogin B (1996) Childhood in evolutionary and biocul- control studies provide support for the argu- tural perspective. In CJK Henry and SJ Ulijaszek 88 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997

(eds.): Long Term Consequence of Early Environment. Harvey PH, Martin RD, and Clutton-Brock TH (1987) Cambridge: Cambridge University Press, pp. 7–22. Life histories in comparative perspective. In BB Smuts, Bogin B (in press) The mid-childhood growth spurt. In DL Cheney, RM Seyfarth, RW Wrangham, and TT SJ Ulijaszek, FE Johnston, and MA Preece (eds.): The Struhsaker (eds.): Primate Societies. Chicago: Univer- Cambridge Encyclopedia of Human Growth. Cam- sity of Chicago Press, pp. 181–196. bridge: Cambridge University Press. Hewlitt BS (1991) Intimate Fathers: The Nature and Bogin B and Smith BH (1996) Evolution of the human Context of Aka Pygmy Paternal Care. Ann Arbor, MI: life cycle. Am. J. Hum. Biol. 8:703–716. University of Michigan Press. Bolk L (1926) Das Problem der Menschwerdung. Jena: Howell N (1979) Demography of the Dobe !Kung. New Gustav Fischer. York: Academic. Bonner JT (1965) Size and Cycle. Princeton, NJ: Prince- Huxley JS (1932) Problems of Relative Growth. London: ton University Press. Methuen. Reprinted 1972, New York: Dover. Bowman JE and Lee PC (1995) Growth and threshold Jaswal S (1983) Age and sequence of permanent tooth weaning weights among captive rhesus macaques. emergence among Khasis. Am. J. Phys. Anthropol. Am. J. Phys. Anthropol. 96:159–175. 62:177–186. Boyd E (1980) BS Savara and JF Schilke (eds.): Origins Jolly A (1985) The Evolution of Primate Behavior, 2nd of the Study of Human Growth. Eugene: University of ed. New York: Macmillan. Oregon Health Sciences Center, Dental School. Jones D (1995) Sexual selection, physical attractive- Boym C (1994) Cover story: By design. Metropolis 13:38. ness, and facial neoteny. Curr. Anthropol. 36:723–748. Brody S (1945) Bioenergetics and Growth. New York: Karlberg J (1985) The human growth curve decomposed Reinhold. into three additive and partly superimposed compo- Brown S (1995) Through the lens of play. ReVision nents: The FBP model. In Abstracts. Fourth Interna- 17:4–8. tional Congress of Auxology. London: Taylor and Fran- Cabana T, Jolicoeur P, and Michaud J (1993) Prenatal cis, p. 46. and postnatal growth and allometry of stature, head Katz SH, Hediger ML, Zemel BS, and Parks JS (1985) circumference, and brain weight in Que´bec children. Adrenal androgens, body fat and advanced skeletal Am. J. Hum. Biol. 5:93–99. age in puberty: New evidence for the relations of Cameron N, Tanner JM, and Whitehouse RH (1982) A adrenarche and gonadarche in males. Hum. Biol. longitudinal analysis of the growth of limb segments 57:401–413. in adolescence. Ann. Hum. Biol. 9:211–220. Klein RG (1989) The Human Career: Human Biological Coelho AM (1985) Baboon dimorphism: Growth in and Cultural Origins. Chicago: University of Chicago weight, length, and adiposity from birth to 8 years of Press. age. In ES Watts (ed.): Nonhuman Primate Models for Kummer B (1952) Untersuchungen uberdie ontogenet- Human Growth. New York: Alan R. Liss, pp. 125–159. ische Entwicklung des menschlichen Schadelba- Cutler GB Jr, Glenn M, Bush M, Hodgen GD, Graham siswinkels. Z. Morphol. Anthropol. 43:331–360. CE, and Loriaux DL (1978) Adrenarche: A survey of La Barre W (1991) Shadow of Childhood: Neoteny and rodents, domestic animals, and primates. Endocrinol- the Biology of Religion. Norman: University of Okla- ogy 103:2112–2118. homa Press. Demirjian A (1978) Dentition. In F Falkner and JM Laird AK (1967) Evolution of the human growth curve. Tanner (eds.): Human Growth 2: Postnatal Growth. Growth 31:345–355. New York: Plenum, pp. 413–444. Laitman JT and Heimbuch RC (1982) The basicranium Dettwyler KA (1995) A time to wean: The hominid of Plio-Pleistocene hominids as an indicator of their blueprint for the natural age of weaning in modern upper respiratory systems. Am. J. Phys. Anthropol. human populations. In P Stuart-Macadam and KA 59:323–343. Dettwyller (eds.): Breastfeeding: Biocultural Perspec- Lancaster JB and Lancaster CS (1983) Parental invest- tives. New York: Aldine de Gruyter, pp. 39–73. ment: The hominid adaptation. In DJ Ortner (ed.): Eccles JC (1979) The Understanding of the Brain. New How Humans Adapt. Washington, DC: Smithsonian York: McGraw-Hill. Ellison PT (1990) Human ovarian function and reproduc- Institution Press, pp. 33–65. tive ecology: New hypotheses. Am. Anthropol. 92:933– Lee PC, Majluf P, and Gordon IJ (1991) Growth, wean- 952. ing, and maternal investment from a comparative Estioko-Griffin A (1986) Daughters of the forest. Natural perspective. J. Zool. Soc. Lond. 225:99–114. History 95:36–43. Leonard WR and Robertson ML (1992) Nutritional Galdikas BM and Wood JW (1990) Birth spacing pat- requirements and human evolution: A bioenergetics terns in humans and apes. Am. J. Phys. Anthropol. model. Am. J. Hum. Biol. 4:179–195. 83:185–191. Leonard WR and Robertson ML (1994) Evolutionary Godfrey L and Sutherland MR (1996) Paradox of pera- perspectives on human nutrition: The influence of morphic paedomorphosis: Heterochrony and human brain and body size on diet and metabolism. Am. J. evolution. Am. J. Phys. Anthropol. 99:17–42. Hum. Biol. 6:77–88. Goldsmith J (1993) Language and learning as properties Lieberman P, Crelin ES, and Klatt DH (1972) Phonetic of our species. Modern Philosophy 90:34–38. ability and related anatomy of the newborn and adult Goodall J (1983) Population dynamics during a 15 year human, Neanderthal man, and the chimpanzee. Am. period in one community of free-living chimpanzees in Anthropol. 74:287–307. the Gombe National Park, Tanzania. Z. Tierpsychol. Lorenz K (1971) Part and parcel in animal and human 61:1–60. societies: A methodological discussion. In R Martin Gould SJ (1977) Ontogeny and Phylogeny. Cambridge, (ed. and trans.): Studies in Animal and Human Behav- MA: Belknap. ior, vol. 2. Cambridge, MA: Harvard University Press, Gould SJ (1979) Mickey Mouse meets Konrad Lorenz. pp. 115–95. Natural History 88:30–36. Lovejoy CO (1981) The origin of man. Science 211:341– Gould SJ (1981) The Mismeasure of Man. New York: 350. Norton. Marshall WA (1978) Puberty. In F Falkner and JM Guthrie H and Picciano MF (1995) Human Nutrition. Tanner (eds.) Human Growth, vol. 2. Postnatal St. Louis: Mosby. Growth. New York: Plenum, pp. 141–181. Bogin] EVOLUTIONARY HYPOTHESES FOR HUMAN CHILDHOOD 89

Martin DE, Swenson RB, and Colins DC (1977) Correla- Smith BH (1992) Life history and the evolution of tion of serum testosterone levels with age in male human maturation. Evol. Anthropol. 1:134–142. chimpanzees. Steroids 29:471–481. Smith BH, Crummett TL, and Brandt KL (1994) Ages of Martin RD (1983) Human Brain Evolution in an Ecologi- eruption of primate teeth: A compendium for aging cal Context. Fifty-second James Arthur Lecture. New individuals and comparing life histories. Yrbk Phys. York: American Museum of Natural History. Anthropol. 37:177–231. Mascie-Taylor CGN and Bogin B (eds.) (1995) Human Spencer H (1886) The Principles of Biology, vols. 1 and 2. Variability and Plasticity. Cambridge: Cambridge Uni- New York: D. Appleton. versity Press. Starck D and Kummer B (1962) Zur Ontogenese des McCabe V (1988) Facial proportions, perceived age, and Schimpansenschadels. Anthropol. Anz. 25:204–215. caregiving. In TR Alley (ed.): Social and Applied Stearns SC (1992) The Evolution of Life Histories. Aspects of Perceiving Faces. Hillsdale: NJ: Lawerence Oxford: Oxford University Press. Earlbaum Associates, pp. 89–95. Tanner JM (1947) The morphological level of personal- McKinney ML and McNamara KJ (1991) Heterochrony: ity. Proc. R. Soc. Med. 40:301–303. The Evolution of Ontogeny. New York: Plenum. Montagu A (1989) Growing Young, 2nd ed. Massachu- Tanner JM (1962) Growth at Adolescence, 2nd ed. setts: Bergin and Garvey. Oxford: Blackwell Scientific. Nishida T, Takasaki H, and Takahata Y (1990) Demogra- Tanner JM, Wilson ME, and Rudman CG (1990) Puber- phy and reproductive profiles. In T Nishida (ed.): The tal growth spurt in the female Rhesus monkey: Rela- Chimpanzees of the Mahale Mountains: Sexual and tion to menarche and skeletal maturation. Am. J. Life History Strategies. Tokyo: University of Tokyo Hum. Biol. 2:101–106. Press, pp. 63–97. Teleki GE, Hunt E, and Pfifferling JH (1976) Demo- Parker LN (1991) Adrenarche. Endocrinol. Metab. Clin. graphic observations (1963–1973) on the chimpanzees North Am. 20:71–83. of the Gombe National Park, Tanzania. J. Hum. Evol. Parker ST (1996) Using cladistic analysis of compara- 5:559–598. tive data to reconstruct the evolution of cognitive Todd JT, Mark LS, Shaw RE, and Pittenger JB (1980) development in hominids. In E Martins (ed.): Phylog- The perception of human growth. Sci. Am. 242(2):132– enies and the Comparative Method in Animal Behav- 144. ior. Oxford: Oxford University Press, pp. 433–448. Ulijaszek SJ and Strickland SS (1993) Nutritional An- Pereira ME, and Altmann J (1985) Development of thropology: Prospects and Perspectives. London: Smith social behavior in free-living nonhuman primates. In Gordon. ES Watts (ed.): Nonhuman Primate Models for Hu- Vrba ES (1996) Climate, heterochrony, and human man Growth and Development. New York: Alan R. evolution. J. Anthropol. Res. 52:1–28. Liss, pp. 217–309. Watts DP and Pusey AE (1993) Behavior of juvenile and Pereira ME and Fairbanks LA (eds.) (1993) Juvenile adolescent great apes. In ME Pereira, LA Fairbanks Primates. Oxford: Oxford University Press. (eds.): Juvenile Primates. Oxford: Oxford University Piaget J (1954) The Construction of Reality in the Child. Press, pp. 148–170. New York: Basic Books. Watts ES and Gavan JA (1982) Postnatal growth of Piaget J and Inhelder B (1969) The Psychology of the nonhuman primates: The problem of the adolescent Child. New York: Basic Books. spurt. Hum. Biol. 54:53–70. Prader A (1984) Biomedical and endocrinological as- Weirman ME and Crowley WF Jr (1986) Neuroendo- pects of normal growth and development. In J Borms, crine control of the onset of puberty. In F Falkner and R Hauspie, A Sand, C Susanne, and M Hebbelinck JM Tanner (eds.): Human Growth, 2nd ed., vol. 2. New (eds.): Human Growth and Development. New York: York: Plenum, pp. 225–241. Plenum, pp. 1–22. Weisner TS (1987) Socialization for parenthood in sib- Rogoff B, Seller MJ, Pirrotta S, Fox N, and White SH ling caretaking societies. In JB Lancaster, J Altmann, (1975) Age assignment of roles and responsibilities of AS Rossi, and LR Sherrod (eds.): Parenting Across the children: A cross cultural survey. Hum. Dev. 18:353– Life Span: Biosocial Dimensions. New York: Aldine de 369. Gruyter, pp. 237–270. Scammon RE (1930) The measurement of the body in Weisner TS (1996) The 5–7 transition as an ecocultural childhood. In JA Harris, CM Jackson, DG Paterson, project. In A Samaroff and M Haith (eds.): Reason and and RE Scammon (eds.): The Measurement of Man. Responsibility: The Passage Through Childhood. Chi- Minneapolis: University of Minnesota Press, pp. 173– cago: University of Chicago Press, pp. 295–326. 215. Weisner TS and Gallimore R (1977) My brother’s keeper: Schultz AH (1960) Age changes in primates and their modification in man. In JM Tanner (ed.): Human Child and sibling caretaking. Curr. Anthropol. 18:169– Growth. Oxford: Pergamon, pp. 1–20. 190. Shea BT (1989) Heterochrony in human evolution: The White SH (1965) Evidence for a hierarchical arrange- case for neoteny reconsidered. Yrbk Phys. Anthropol. ment of learning processes. In LP Lipsitt and CC 32:69–101. Spiker (eds.): Adv. Child Dev. Behav., vol. 2. New York: Short RV (1976) The evolution of human reproduction. Academic, pp. 187–220. Proc. R. Soc. Lond. [B] 195:3–24. Whiting B and Edwards C (1988) Children of Different Simons EL (1989) Human origins. Science 245:1343– Worlds. The Formation of Social Behavior. Cam- 1350. bridge, MA: Harvard University Press. Smail PJ, Faiman C, Hobson WC, Fuller GB, and Whiting B and Whiting JWM (1975) Children of Six Winter JS (1982) Further studies on adrenarche in Cultures. A Psycho-cultural Analysis. Cambridge, MA: nonhuman primates. Endocrinology 111:844–848. Harvard University Press. Smith BH (1991a) Age at weaning approximates age of Winter JSD (1978) Prepubertal and pubertal endocrinol- emergence of the first permanent molar in non-human ogy. In F Falkner and JM Tanner (eds.): Human primates. Am. J. Phys. Anthropol. Suppl. 12:163–164 Growth, vol. 2, Postnatal Growth. New York: Plenum, (abstract). pp. 183–213. Smith BH (1991b) Dental development and the evolu- Worthman CM (1986) Later-maturing populations and tion of life history in Hominidae. Am. J. Phys. Anthro- control of the onset of puberty (abstract). Am. J. Phys. pol. 86:157–174. Anthropol. 69:282.