sign in sign up
<<
Home , Human head

Speculations About the Selective Basis for Modern Craniofacial Form

The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters

Citation Lieberman, Daniel E. 2008. Speculations about the selective basis for modern human craniofacial form. Evolutionary Anthropology 17(1): 55-68.

Published Version doi:10.1002/evan.20154

Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:3716643

Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA Evolutionary Anthropology 17:55–68 (2008)

ARTICLES

Speculations About the Selective Basis for Modern Human Craniofacial Form

DANIEL E. LIEBERMAN

The last few decades have seen an explosion of knowledge about the time and Although it is always difficult to test place of origin of our species, Homo sapiens. New fossils, more sites, better hypotheses about natural selection dates, modern and fossil DNA, and scores of analyses have mostly disproved the from the fossil record, this problem is multiregional model of human evolution. By and large, the evidence generally sup- especially compounded for ports some version of the out-of-Africa model, according to which humans first because of the detailed nature of the evolved in Africa at least 200,000 years ago and then migrated to other parts of questions we ask, the complexity of the world. Remaining debates about human origins primarily address if and how human behavior, and our limited much hybridization occurred between modern humans and taxa of archaic Homo ability to conduct controlled experi- such as H. neanderthalensis. ments that test relationships among form, function, and fitness. Thus, the hypotheses presented here are diffi- Ironically, despite a growing con- these sources of information, the cult or perhaps impossible to test. sensus about when and where H. presents some especially inter- Another problem is that hypotheses sapiens first evolved, we know little esting problems for several reasons. about natural selection in human evo- more today than we did 20 years ago First, the skull is more derived in lution carry much emotional bag- about why we evolved. A key unre- modern humans than is the postcra- gage. This is particularly true for hot- solved question is: What were the nium,1 and is thus a focal point for button topics such as and selection pressures that favored the defining H. sapiens as a species and cognition, in which deeply held pre- evolution of modern humans in testing hypotheses about phyloge- conceptions about what it is to be Africa around 200 Ka? We remain far netic relationships within the genus human lead to a priori viewpoints and from a definitive answer to this ques- Homo.2–5 Second, the skull partici- strenuous objections to alternative tion, in part because of the substan- pates in so many critical functions, hypotheses. At one end of the contin- tial challenges of testing hypotheses among them cognition, vocalization, uum are scholars who insist that about selection using archeological, respiration, diet, and thermoregula- humans are special (for example, that genetic, and paleontological data. Of tion, that it is a valuable source of in- only humans can have language); at formation for making inferences the other end are scholars who about behavior. And finally, the human believe that all large-brained homi- Daniel Lieberman is Professor of Biologi- nids are essentially human (for exam- skull is remarkably odd. To name just cal Anthropology at Harvard University. ple, that Neanderthals are ‘‘human’’ His research uses both comparative and a few of our unusual craniofacial apo- too). Nevertheless, it is occasionally experimental approaches to address morphies, we are the only extant pri- why the looks the way it useful to speculate on controversial mate with an external , no snout, a does. Particular interests include the ori- topics. What follows, then, is a review gins of human craniofacial form and the spherically shaped (globular) - evolution of bipedal walking and running. of some previous ideas and some fur- He is just completing a book, The Evolu- case, a highly flexed cranial base, a ther speculations about why the mod- tion of the Human (Harvard Univer- that is retracted almost entirely ern human skull may be unique. A sity Press). beneath the braincase, fur on just the Departments of Anthropology, and Orga- few caveats. First, I admit that many nismic and Evolutionary Biology, Harvard top of the head, a , and so on. of the ideas presented here really are University, 11 Divinity Avenue, Cambridge These and other features beg the ques- MA, 02138, USA. E-mail: danlieb@fas. just speculations that will be difficult harvard.edu tion of what selective forces were at to test definitively. Second, I have work to make our skull so especially entirely avoided the topic of sexual derived in so many unique respects. selection not because it was unimpor- Key words: Homo sapiens; evolution; adaptation; This essay is therefore an effort to tant in human evolution, but because function; skull; cranium review and speculate about some of I have no idea how to test for its the potential selective advantages that occurrence. Finally, none of the fol- VC 2008 Wiley-Liss, Inc. may be responsible for the origin of lowing ideas are dear to my heart; I DOI 10.1002/evan.20154 modern human craniofacial form. I will be happy to see them all dis- Published online in Wiley InterScience (www.interscience.wiley.com). undertake it with some trepidation. proved. 56 Lieberman ARTICLES

First, however, it will be useful to tures that may be part of the same interacting factors influence neuro- consider a few key pitfalls in trying to morphological complex. cranial shape, including the size and infer adaptation from morphological While laudable, the spandrelist shape of the brain itself, as well as features. approach also has its problems, myriad interactions between the because to prove rather than assume brain, its dural folds, the basicra- that a feature is a spandrel, one must nium, the face, and the neurocranium SOME PROBLEMS WITH first prove that it is not an adaptation. (for review, see Lieberman, Ross, and 15 TESTING HYPOTHESES ABOUT Such tests are especially difficult Ravosa ). Consequently, multiple when applied to the fossil record or to factors potentially can cause the ADAPTATION IN THE SKULL species such as humans and other braincase to become more spherical. One resurgent problem with testing primates for which there are substan- In mice, for example, independent hypotheses about possible adapta- tial experimental constraints. In real- mutations that shorten either the face tions in the skeleton is the issue of ity, to assume that a feature is either or the basicranium both lead to a rel- setting up and rejecting appropriate an adaptation or a spandrel should be atively wider, more spherical neuro- null hypotheses. Like many biolo- equally burdensome in terms of cranium and a more flexed cranial gists, students of human evolution rejecting a null hypothesis. base.16 Without knowing how these have tended to fall into one of two A second problematic issue with test- factors interrelate, it would be diffi- extreme camps. At one extreme is the ing functional hypotheses about hu- cult to know which were targets of ‘‘adaptationist’’ perspective, critiqued man origins from the fossil record has selection and which were spandrels. most famously by Gould and Lewon- been the assumption of independence Finally, just as there is a tendency tin.6 This perspective typically as- among features. The H. sapiens skull, to ‘‘atomize’’ the skull as a set of fea- sumes a causal, adaptive relationship tures that are assumed to be inde- between form and function. As noted pendent, there is also a tendency to by Lauder,7 the adaptationist per- In reality, to assume that atomize selective hypotheses. We may spective is often characterized by a identify many bona fide adaptations in kind of ‘‘reverse engineering’’ ap- a feature is either an the skull that improve performance in proach in which one first identifies adaptation or a spandrel different functional roles such as novel features such as or should be equally speech, locomotion, cognition, and rounded cranial vaults and then as- thermoregulation. Yet these adapta- sumes that these features might have burdensome in terms of tions may not be independent at the been adaptive—that is, a heritable rejecting a null level of the organism’s behavior. Thus, feature that has been selected the selective forces behind the evolu- because it improves survival and/or hypothesis. tion of the human skull need to be con- differential reproductive success. sidered in their general behavioral and According to this approach, the null cultural contexts: how early humans hypothesis for a feature is that it is an foraged, hunted, used the landscape, adaptation unless proven otherwise. for example, is typically considered a and interacted with each other. Put dif- This logic has led to many proposed mosaic of many independent or par- ferently, good hypotheses about why adaptations in the human skull, tially independent features, such as a humans have distinctively shaped including that the flexed cranial base large brain, a flexed cranial base, a need to consider the selective is an adaptation for speech8 and that , highly curved cranial advantages of the skull’s various a chin is an adaptation for reducing vault , and so on. Many devel- derived features in the context of a wishboning of the .9 Since opmental studies, however, indicate larger behavioral and selective sce- Gould and Lewontin,6 however, there that these traits are not independent, nario. has been a reaction against adapta- but instead are highly integrated. Given these challenges, a useful tionism, leading to an opposite Over-atomization of traits is a prob- strategy for thinking about selective ‘‘spandrelist’’ extreme in which the lem not only for phylogenetic analy- hypotheses that may have driven the null hypothesis for a morphological ses, but also for testing hypotheses origin of modern human craniofacial feature is that it is a spandrel unless about natural selection. Skulls are form is to consider the skull as an proven otherwise. Lauder,7 for exam- complex, strongly integrated struc- integrated unit, then to work back- ple, proposed a demanding set of cri- tures characterized by high levels of ward to generate hypotheses about teria with which to test such evolu- covariation among multiple struc- particular functions and how these tionary arguments from design, tures, even in different regions such functions may be related. Formalized, namely that proposed adaptations are as the face, basicranium, and neuro- this is a three-step, largely inductive derived, that they are tested experi- cranium.10–14 Thus, when two or process. The first step is to determine mentally in terms of their effects on more features covary, it is no simple to what extent and how the features performance, that alternative func- matter to determine which were tar- that changed at a given speciation tions are also tested, and that the fea- gets of selection versus byproducts of event are integrated; in other words, tures have not been defined too selection. Consider, for example, the what changed with what? The second strictly to exclude other related fea- shape of the neurocranium. Many step is to identify the functional ARTICLES Speculations on Modern Human Craniofacial Form 57 effects of these shifts and consider how they would have affected the function of other proposed modular shifts. The third step is to generate testable hypotheses about how selec- tion might have acted, trying to reject simultaneously the alternative null hypotheses that a given feature is an adaptation and a spandrel.

MAJOR DERIVED FEATURES OF THE MODERN HUMAN SKULL As argued, in order to test hypothe- ses about the functional implications of modern human cranial form it is first useful to ask what are the major integrated units of the skull that changed together in the transition from archaic to modern humans. Since various studies have already addressed this problem in depth, I will only briefly summarize their results here. Most notably, Lieberman and colleagues12,15,17,18 have shown that many of the derived features of the H. sapiens skull are not independ- ent, but instead covary substantially both within adult and cross-sectional ontogenetic samples. Three inte- grated shifts between archaic Homo and modern humans account for a large proportion of the derived fea- tures of H. sapiens. First, the modern human skull has an absolutely and relatively smaller face, mostly in terms of anteroposterior length and superoinferior height. In Figure 1a, note, for example, that the orbits are Figure 1. Comparison of modern H. sapiens and archaic Homo cranial form. A. Lateral relatively more rectangular in modern and anterior views of Cro Magnon I (left) and Broken Hill (right), scaled to same height. than archaic humans). Second, the Note the relatively longer face and orbits of the archaic human (as shown by white modern human neurocranium, al- bars), which generates a more square-shaped orbit; the modern human face is also though similar in volume, is more larger and more projecting, and the neurocranium is more spherical. B. graph of centroid spherical than that of archaic Homo, size versus the first principal component of cranial shape in Pleistocene modern H. sapi- leading to more rounded contours of ens (stars) and archaic Homo (circles), H. erectus (squares), and H. habilis (pentagon) fol- lowing a Procrustes superimposition of the landmarks. Note that for this component of all the cranial vault bones. Third, the shape variation, which explains 32% of the variance, archaic Homo is largely a size- smaller modern human face is sub- scaled version of H. erectus, but that H. sapiens falls off the allometry between size and stantially retracted relative to the shape. Line drawings to left and right show mean form along PC1 of modern and ar- basicranium and neurocranium, con- chaic humans, which comprises neurocranial globularity. Relative facial size and facial tributing most obviously to a reduc- projection. See Lieberman and Bar-Yosef20 for details of sample and analysis. tion in browridge size.19 Geometric morphometric (GM) shifts (all evident in Figs. 1 and 2) back of the face is always roughly per- analyses that compare archaic and appear to be particularly important. pendicular to the anterior cranial modern human skulls17,18,20 corrobo- First, the modern human cranial base base.21–23 The combined effect of these rate these results, suggesting that a is about 158 more flexed. This flexion constraints is that a more flexed cra- relatively small number of develop- has several implications for skull nial base reorients the entire face as a mental shifts may be responsible for shape, in part because the top of the block more ventrally beneath the an- the integrated shape differences I face is the base floor of the anterior terior cranial fossa and decreases the have summarized. Three structural cranial base and also because the length of the nasopharynx behind the 58 Lieberman ARTICLES

tion, chewing, locomotion, respira- tion, and speech.

Cognition and Brain Shape One longstanding topic of interest is whether modern humans had any cognitive advantages over archaic humans, perhaps related to the ability to create more complex technologies such as the Upper Paleolithic, gener- ate more sophisticated language, and so on. However, it has been a chal- lenge to find or test for any evidence of such adaptations. Most of the rele- vant archeological evidence is a re- cord of absence (that is, the lack of Upper Paleolithic technologies and Figure 2. Midsagittal view of a, Skhul V (H. sapiens) and b, Broken Hill (archaic Homo). art made by archaic Homo) and does Note that in this particular comparison the angle of the cranial base is approximately 188 not explain why the archeological res- more flexed in the modern human. The Broken Hill scan is by courtesy of the Natural His- tory Museum. idues associated with modern and ar- chaic humans are essentially identical for more than 150,000 years. Fossil palate. Cranial base flexion may also Homo. As Figure 1b illustrates, large- evidence of neurological differences contribute to neurocranial sphericity brained taxa from the Middle Pleisto- has also been hard to discern. Abso- by decreasing the area of the platform cene, such as H. heidelbergensis, are lute and relative brain size in modern on which the brain sits.24,25 A second in many way scaled-up versions of H. and archaic humans is approximately major structural change is that in erectus, different primarily in terms equal.27 Also, there are few, if any, modern humans the anterior cranial of relatively larger and .20 reliable endocranial features that base is relatively longer and wider, In contrast, modern human skulls point to major differences in brain especially the more posterior part have an overall different architectural structure, such as Broca’s or Wer- comprised of the middle cranial configuration. Put differently, the evo- nicke’s areas, between archaic and fossa.17,19,26 Elongation of the ante- lutionary transformation from a H. modern humans.28 However, another rior cranial base probably contributes erectus skull into a H. heidelbergensis potential source of information is mostly to facial retraction by displac- skull can mostly be explained by whether the contrasts in overall neu- ing anteriorly the upper (superior) changes in the relative size of the rocranial shape reviewed earlier re- part of the posterior margin of the brain and face. In contrast, the evolu- flect some differences in the relative face, known as the posterior maxil- tionary transformation from an ar- size of, and hence, selection on par- lary (PM) plane. Because the PM chaic Homo skull into a modern ticular regions of the brain. The basis plane is always nearly perpendicular human skull is more complex, requir- for such speculations is that the to the top of the face, this displace- ing a relatively longer middle cranial braincase and basicranium grow ment rotates the face as a whole ven- fossa, a more flexed cranial base, and around the brain, largely in response trally (clockwise when viewed from a smaller face in terms of height and to stimuli triggered by the growth of the right).21,22 The third major shift is length. the brain itself.17,29,30 Thus, varia- that the modern human face is rela- tions in the relative size of specific tively shorter both anteroposteriorly parts of the brain could generate dif- and superoinferiorly, but not nar- POTENTIAL FUNCTIONAL CONSE- ferences in the shape of the cranium rower. Facial reduction, which has QUENCES OF MODERN HUMAN as a whole. occurred more gradually than cranial One such hypothesis is that the CRANIAL APOMORPHIES base flexion and anterior cranial base modern has a relatively elongation, contributes primarily to We can now speculate (and I larger frontal lobe.31 Despite its facial retraction, but may also be emphasize that what follows are spec- appeal, this hypothesis has been dis- implicated in cranial base flexion (for ulations, not tests) about how the pre- proved by comparative studies in discussion see Lieberman, Ross, and ceding shifts might be adaptive in hominoids. Measurements of the neo- Ravosa15). terms of their performance effects on cortex and other structures in the Note also that from a developmen- craniofacial function. Of the many brain of extant primates using mag- tal perspective the differences be- varied functions that a skull per- netic resonance imaging (MRI) have tween modern and archaic Homo forms, five appear to be most relevant shown that the frontal, parietal, and skulls appear to be greater than those to neurocranial sphericity, and/or occipital lobes in humans are about between H. erectus and archaic facial shorting and retraction: cogni- the same relative size as in other ARTICLES Speculations on Modern Human Craniofacial Form 59

brain relative to cranial base length influences cranial base flexion.15,39,40 While overall brain size cannot be the cause of more cranial base flexion in modern humans, a relatively larger temporal lobe is a possible cause, since it is the one portion of the brain that lies directly on the site where the cranial base actually flexes during the first few postnatal years.41 To reiterate, the hypothesis that rel- ative increases in the temporal lobe were a factor in the evolution of mod- Figure 3. A. Geometric morphometric analysis of cranial shape differences between ar- ern humans remains speculative with- chaic and modern humans, from Lieberman, McBratney, and Krovitz.17 The red outline is out further evidence. As initial tests of an average of two archaic Homo crania (Broken Hill, Bodo) that have been warped to the hypothesis, it would be useful to fit the same landmarks on four male modern humans (black outline). Grid deformations determine whether the MCF and tem- indicate the thin plate spline. Key differences, indicated by arrows, are the greater poral lobe volume are correlated, and degree of cranial base flexion, the relatively longer anterior cranial base, and the rela- whether humans with larger temporal tively smaller face. B. Outline of lateral view of the temporal lobes relative to the cranial base. Note that the temporal lobes sit above the sphenoid, which is major site of cranial lobes have more flexed cranial bases. base flexion. Also shown (in red) is the 908 constant angle between the back of face If so, however, then the hypothesis (the posterior maxillary plane) and the neutral horizontal axis (NHA) of the orbits. Because has intriguing implications for selec- this angle is constrained to be approximately 908 and the anterior cranial base is tightly tion. The temporal lobe has many correlated with the NHA, flexion of the cranial base causes the entire face to rotate complex functions, including cogni- underneath the anterior cranial fossa. tive roles relevant to the organization of sensory input. These include lan- guage and various kinds of memory, hominoids after scaling is taken into coincident with the origin of modern such as the recognition of words, consideration.32–35 In contrast, the humans. sounds, and visual images. In addi- human cerebellum is proportionately Such a shift might also help explain tion, the temporal lobe has been smaller than in apes, by approxi- some of the major derived aspects of shown to be important during intensely mately 20%, and the temporal lobe is modern human cranial shape. The spiritual and religious thoughts; for proportionately larger, by as much as anterior portion of the MCF com- example, stimulation of the temporal 25%. Other, more fine-scale differen- prises part of the anterior cranial lobe during surgery can induce spirit- ces may also exist within lobes. base and, as noted earlier, the mod- ual emotions even in self-described Within the occipital lobe, the primary ern human anterior cranial base is atheists.42 It is thus interesting to visual striate cortex is relatively relatively longer than that of archaic speculate that religious and spiritual smaller in humans36; within the fron- Homo, contributing to facial retrac- behaviors, so prevalent in all human tal lobe, the prefrontal cortex is about tion (more of the modern human face societies, are a derived, emergent 6% larger, possibly because of more is tucked below the anterior cranial property of selection on other capa- white matter.35,37,38 fossa).19 Both GM data from fossil bilities such as memory and lan- Of these differences, evidence that crania and the comparative MRI data guage. the temporal lobe became relatively from primate brains summarized - larger at some point in human evolu- lier suggest that the increases in tem- Bite Forces and Cooking tion is the most intriguing because of poral lobe size are responsible for this the differences I have noted in cranial anterior cranial base elongation. One intensely studied aspect of the shape between modern and archaic Another major factor that influences relationship between form and func- humans. In particular, geometric both neurocranial globularity and tion in the skull is the generation and morphometric comparisons of the facial retraction is the angle of the resistance of bite forces. Thus, it is cranial base (summarized in Fig. 3) cranial base, which is about 158 more not for lack of looking that most indicate that in modern versus ar- flexed in modern than in archaic researchers have found little reason chaic humans the middle cranial Homo. This difference, in conjunction to suspect that the modern human fossa (MCF), which houses the tem- with evidence of temporal lobe expan- face has any particular derived adap- poral lobes, is approximately 15%– sion, raises an interesting possibility. tations for generating or withstand- 20% longer and 20% wider relative to As Figure 3 illustrates, the temporal ing chewing forces. Nevertheless, the overall cranial size.17,18,26 If MCF size lobe sits on the center of the cranial human face presents an interesting correlates strongly with temporal lobe base, just above the spheno-occipital paradox, as shown in Figure 4, which volume (an untested hypothesis), synchondrosis, the major site of cra- compares, in lateral view, the esti- then it is possible that a proportional nial base flexion. Various lines of mated mechanical advantages of bite- increase in temporal lobe volume was evidence indicate that growth of the force production in the vertical plane 60 Lieberman ARTICLES

from nonhuman primates and other mammals with a somewhat tubular rostrum in which the maxillary tooth row is positioned well in front of the plane of the middle and upper face. Nevertheless, finite element modeling studies,51 as well as experimental data from mammals with retracted postcanine teeth, such as the rock hyrax (Procavia capensis), suggest that we can make at least two reason- able predictions.52,53 First, human fa- cial strains probably resemble those of other mammals in following a strong gradient, with strains highest near the occlusal plane and regions of muscle attachments and diminishing away from these sites, with only tiny magnitudes in the upper face and browridge.48,49 In addition, unilateral mastication probably generates some Figure 4. Mechanical advantages of the masseter (top) and temporalis (bottom) in a combination of twisting around an chimpanzee (a), H. sapiens (b), and a composite archaic Homo skull (c, a combination anteroposterior axis plus shearing in of the Broken Hill cranium and the Mauer mandible). Mechanical advantage (MA), the ratio of the lever to load , is calculated from the estimated average resultant of the coronal plane. masseter and temporalis for a chew on M2. Note that resultant is an approximate esti- As noted above, one of the most dis- mate and will vary depending on factors such bite location, food hardness, and other tinctive aspects of the Homo face is such factors. In spite of such variations, the modern human will have a higher MA for the lack of a rostrum, which in non- both muscles than did the archaic Homo, largely because of facial retraction and dimi- human primates and mammals such nution. as swine partly functions to dissipate twisting and bending strains away from the rest of the face.54,55 Both ar- of the masseter and temporalis mu- smaller cross-sectional areas for some chaic and modern humans therefore scles for a chew on the second molar muscles44 and relatively fewer fast- presumably resist such strains in an in a chimpanzee, a modern human, twitch (Type I) muscle fibers.45 Con- alternative manner by having a tall, and a representative archaic Homo sequently, most research on this topic wide, flat face in which most of the (Broken Hill). Although the orienta- has focused on how derived aspects mass of the face is distributed in the tions of the muscle resultants are esti- of midfacial projection in the Nean- coronal plane.52,56 If so, then one can mated and approximate, it is evident derthal skull may have been adaptive. hypothesize that the face in archaic that by retracting and shortening the But the other side of the coin is to Homo is larger, and especially taller face, the human skull has a relatively consider not the generation of bite than that of modern humans in order shorter load for both the mass- force, but its resistance. As shown by to withstand more and/or higher eter and the temporalis, leading to an Hylander and colleagues,46–49 chew- magnitude strains from twisting and approximately 20%–40% higher me- ing generates strains that are not only shearing. Put differently, the modern chanical advantage in modern versus similar in magnitude to those gener- human face may be superoinferiorly archaic humans. ated in the limbs during locomotion, shorter not to generate higher occlu- Although the human skull is more but also highly repetitive. A typical sal forces, but because it has to with- efficiently designed for producing bite chimpanzee spends about 50% of the stand less . Indeed, in terms of forces, it would probably be unwise day chewing.50 Thus, a key function its surface area in the coronal plane, to hypothesize that the human face of the face is to withstand mastica- the face is approximately 30%–50% was selected to become shorter in tory strains, either through adding smaller in modern than archaic order to increase masticatory effi- mass to decrease their magnitude or humans, with substantial decreases ciency. There is simply no corroborat- through repair mechanisms. having occurred in the last 12,000 ing evidence. Humans have small and How the human face withstands years.57 Moreover, some proportion thinly enameled tooth crowns that chewing forces is incompletely under- of this size diminution is epigenetic have become smaller rather than stood, in part because we know rela- and can be attributed to the mastica- larger in the genus Homo during the tively little about the nature of chew- tion of softer, more processed last few hundred thousand years.43 In ing-induced strains in the uniquely foods.53,58,59,106 addition, humans do not produce flat, short, retracted human face. This hypothesis needs to be tested high bite forces relative to our body Most in-vivo experimental data on but, if true, then one likely compo- size for various reasons, including craniofacial strain patterns comes nent of the selective equation behind ARTICLES Speculations on Modern Human Craniofacial Form 61

that volume and mass are propor- tional). However, as I will discuss, a smaller face may have other func- tional advantages related to locomo- tion, respiration, and vocalization. So small faces cannot be explained by diet alone.

Locomotion Although the role of the skull in locomotion is not often considered, Figure 5. Mechanical advantage (MA) of the nuchal muscles for a chimpanzee (a), H. one hypothesis to explore is that de- sapiens (b), and a composite archaic Homo skull (c, a combination of the Broken Hill creases in facial size played a unique cranium and the Mauer mandible). The center of gravity of each skull (bullseye) was esti- mated from the area centroid; the resultants of the nuchal and COG are based on role in humans to improve head sta- unpublished experimental data. The combination of a relatively long nuchal plane, a bilization capabilities during running. centrally located foramen magnum, and a nearly balanced skull gives the nuchal In all mammals, there is a tendency muscles a higher MA in modern humans than in the other two species. for the head to pitch in the sagittal plane during locomotion from a com- bination of vertical and horizontal facial size reduction (why big faces gins of the genus Homo, almost cer- accelerations (Fig. 5). These pitching were no longer adaptive) is most tainly involved a change in diet, with forces need to be controlled in order likely related to cooking and/or food increased reliance on meat, which is to stabilize the gaze and avoid falling. preparation. As Wrangham and col- easily and effectively tenderized by This is not a trivial problem in inher- leagues60,61 noted, cooking is not only pounding. Thus, the regular addition ently unstable, long-legged bipeds a universal human behavior, but also of meat to the diet, combined with such as humans. Although minimal may be a biological necessity for the adoption of simple food process- in walking, pitching forces increase humans, even modern ‘‘raw-foodists’’ ing technologies such as pounding (a several-fold in running gaits, espe- who intensively soak, grind, and oth- good use for Oldowan spheroids?), cially at , when high erwise process highly domesticated, may have led to the initial diminution ground reaction forces, the heel-strike high-quality foods. Cooking, of course in tooth size, but not to smaller faces. transient (HST), travel almost instan- has many advantages, including in- Later, cooking evolved in the Middle taneously up the legs and the spine, creasing nutrient accessibility and Paleolithic, favoring further reduc- delivering an impulsive load to the digestibility, inactivating toxins and tions in tooth size, as well as reduc- head that causes it to pitch forward parasites, and prolonging storage tions in face size. In addition, rapidly.70 times.62 Cooking also makes food eas- although teeth have become smaller Controlling head pitching forces is ier to chew. Cooking softens and ten- in both modern and archaic Homo much more of a challenge for bipedal derizes plants by breaking down cel- since the Middle Paleolithic,107 facial humans than for most quadrupedal lulose; it makes meat stiffer and less size decreases may have been less in mammals, which are well adapted for tough, and thus easier to fracture.63 Neanderthals for other reasons such running (cursorial specialists). When There is much debate, however, about as their expanded internal nasal cav- walking and running, mammals typi- when cooking became prevalent in ities and their reliance on anterior cally hold their with their human evolution. On the basis of dental loading.56,68,69 directed forward and their horizontal tooth size reduction, Wrangham and Note that the invention or elabora- semicircular canals within 208 of earth coworkers60 and Lucas63 have sug- tion of cooking technologies in the horizontal.71,72 Quadrupedal cursors gested that cooking first arose with Middle Paleolithic/Middle Stone Age effectively stabilize their heads during the genus Homo. However, according may help account for why there was trotting and galloping by flexing and to archeologists, the oldest evidence no longer a selective advantage to extending the , which has some of cooking in the form of hearths is maintaining a large face, but it pro- degree of horizontal orientation.73 from the Middle Paleolithic/Middle vides no selective advantage for This simple mechanism of stabiliza- Stone Age, about 250,000 years ago,64–67 reducing facial size. So any selective tion, however, is not available to approximately the same time as the scenario needs to consider why a large-brained bipedal hominids, whose evolution of H. sapiens (and the smaller, lighter face might be an ad- short vertical arise near the Neanderthals). aptation. The modern human face is center of the skull. In this respect, Further evidence is needed to shorter both superoinferiorly, reduc- human running is a bit like being on resolve this problem, but one hypo- ing its area in the coronal plane, and a pogo stick, because we have no thetical possibility is that tooth size anteroposteriorly. The combined mechanisms to control vertical fluctu- decreases and facial reduction consti- effect is to make the modern face ations of the head’s center of gravity tuted a two-stage process. The first absolutely less massive in modern (COG) and to dissipate the HST. stage, which occurred with the ori- versus archaic humans (assuming Instead, hominids must stabilize the 62 Lieberman ARTICLES head against pitching forces by flex- performance capabilities, better than and the rate of flow. In a circular tube ing or extending the head almost those of most mammals, particularly with laminar flow (characterized by entirely at the atlanto-occipital joint over long distances in hot, arid envi- no turbulence), resistance to flow, R, (AOJ). Yet even during running, ronments. These capabilities derive is quantified by the Hagen-Poisseuille humans usually manage to keep from numerous features, some of equation: angular rotations of the head below which improve running performance 1008/sec following the HST, well but are irrelevant to walking, in R ¼ 8nl=pr4 below the 2008 threshold at which the which pendular mechanics are funda- vestibulo-ocular reflex, which mentally different from the mass- where n is the viscosity of the gas, l is pitching movements and coordinates spring mechanics of running. When the tube’s length, and r is the tube’s 81 them with movements, becomes ER capabilities arose is difficult to radius. Because air viscosity is not saturated.74,75 pinpoint, but they were probably species-specific, and the nasal and As Bramble and Lieberman noted,76 present in Homo and may have oral cavities are neither wider nor humans have several features that are played an important role in acquiring taller in modern humans (it is impos- not present in apes and possibly are meat by scavenging, hunting, or sible to estimate the radius of the oro- also absent in australopithecines, both.76 It follows that if ER capabil- pharynx in fossils), modern humans which may help stabilize the head. ities improved modern human hunt- probably had slightly higher airflow These features include more sensitive ing performance, then morphological resistance during conditions of lami- anterior and posterior semicircular changes such as smaller faces that nar flow. However, most airflow canals,77 which accelerations in improved head stabilization might through the human pharynx is not the pitching plane, and a nuchal liga- have been a selective advantage. laminar, but turbulent, characterized ment, which may act as a passive, by circulating vortices with no partic- elastic mechanism to stabilize the ular orientation relative to the tube’s 78,79 Respiration head. To this list we should prob- walls or the average direction of flow. ably add a more balanced head, Having a smaller face also has Turbulence increases enormously dur- because it reduces the moment arm potential functional effects on the ing vigorous aerobic exercise. There of the pitching forces that need to be role of the pharynx in respiration, is no simple equation to estimate R in counteracted by head extension. Even particularly in terms of thermoregula- turbulent flow, but turbulence itself is though more balanced heads prob- tion during vigorous activities such as a function of the gas (or fluid’s) Reyn- ably arose in Homo in part because of running. Superoinferior and antero- old’s number, Re: bigger brains, which extend more posterior shortening of the face mass behind the AOJ, a smaller face decreases the length of several com- Re ¼ 2rvd=n also improves head stabilization per- ponents of the pharyngeal airway formance in bipeds with short vertical (Fig. 6), including the nasal cavity where r is the radius of the tube, v is necks, as shown in Figure 5. In mod- and the nasopharynx above the soft the average fluid velocity of the gas, d ern humans, the head’s COG is palate, the oral cavity, and the oro- is the density of the gas, and n is the approximately 1 cm anterior to the pharynx. If we assume that neck viscosity of the gas. As before, d and AOJ.80 Although it is difficult to calcu- length has not changed much in the n are determined by the environment, late precisely the location of the COG genus Homo, then the modern human and r is probably not much different in fossil crania, preliminary estimates pharynx is significantly shorter, possi- in archaic versus modern Homo. using the area centroid as an approxi- bly by about (12%–15%) relative to However, the fluid velocity, v, is prob- ably a little higher in modern humans mation suggest that facial diminution body mass than that in archaic because oxygen needs and lung vol- in modern humans may be responsi- Homo. This difference has several ume scale isometrically with body ble for moving the COG approxi- implications for the pharynx’s role in mass.82 Consequently, the same vol- mately 1 cm closer to the AOJ than in thermoregulation. The epithelial lin- ume of air per unit time (flow rate) archaic Homo. As shown in Figure 4, ing of the nasal cavity and pharynx is flows though a shorter tube, leading this small shift would halve pitching highly vascularized and covered by to a higher average velocity. This dif- moments and increase the mechani- hydrophilic mucosa. This lining func- ference in length, hence velocity, mat- cal advantage of the head extensors tions to humidify and warm air to ters considerably during aerobic exer- by about 50%. In other words, mod- approximately 75%–80% humidity and cise. At optimal walking speed, about ern humans may have been more 378 C, respectively; during expiration, 5 kph humans breathe about 2.5 L/ effective than archaic humans at sta- the same tissues can act to recapture min. During moderate jogging, 15 bilizing the head because of a smaller some of this heat and moisture. These kph, oxygen demands increase more face. thermoregulatory functions are partic- than seven-fold to about 18 L/mi- Of course the hypothesis that facial ularly important in the nasal cavity, nute,83 requiring both faster breath- diminution was an adaptation for which has a much higher ratio of epi- ing rates and larger volumes per head stabilization during running is thelial surface area to volume. breath, from about 500 mL to 1,500 untested. However, as Bramble and A shorter pharynx affects these mL. Even faster flow rates are needed Lieberman noted,76 humans have functions by altering two interrelated at higher speeds or higher altitudes. spectacular endurance running (ER) parameters: the degree of turbulence Consequently, turbulence in the mod- ARTICLES Speculations on Modern Human Craniofacial Form 63

Figure 6. Midline view of the pharynx and supralaryngeal vocal tract (SVT) in a modern human (a) and two alternative reconstructions of an archaic Homo (b,c). SVTh is shown in red and SVTv in blue. The ratio of SVTv to SVTh in the modern human is 1:1. In version b, the archaic Homo has been reconstructed with an intranarial larynx and an intraoral ; in version c, the archaic Homo has been reconstructed with a descended (nonintranarial) larynx and a rounded human tongue. In both cases, the length of the SVTv necessary to have a 1:1 SVTv:SVTh, is indicated by the dashed line. ern human pharnyx, already very modern humans.84 Because such a of two relative to velocity, P = kV2, high during conditions of exercise, boundary decreases effective heat thus generating much higher pres- must be higher for modern than ar- and moisture exchange, it is not sur- sures. One well-known effect of this chaic humans. prising that the nose in Homo has trade-off is that resistance is so high Turbulence and the resistance it many adaptations to increase turbu- in the human nasal cavity during vig- causes lead to an important trade-off lence, including an external nasal ves- orous exercise that the pressures can in thermoregulation. The advantage tibule, which is first evident in early be painful. Because airflow in the of turbulence is that it increases air- Homo,85 inferiorly directed nostrils, is much more laminar than flow contact with the epithelium in and a nasal valve.86 However, the dis- in the nose (r is an order of magni- the pharynx, especially the nose. advantage of turbulence is that it gen- tude bigger), this may explain why Within a cross-section, there is a ve- erates considerable extra work for humans are the only known mammal locity gradient in laminar flows, with lungs. In laminar flow, pressure is a that is an obligate mouth breather flow rates approaching zero near the linear function of velocity, P = kV (k during vigorous exercise. wall of the tube, creating a boundary is a constant). In turbulent flow, how- Putting together the evidence, a zone estimated to be 0.25 mm wide in ever, pressure increases to the power shortened pharynx in humans may be 64 Lieberman ARTICLES an adaptation for thermoregulation harmonics, known as formant fre- Thus, considerable effort has been in the hot and arid environments in quencies. Importantly, each vowel devoted to estimating the position of which modern humans appear to has a unique, stable, distinct, and the hyoid and larynx in fossil homi- have evolved. hence quantal combination of form- nids in order to make inferences During normal activity levels such ant frequencies that derive from the about the evolution of speech capabil- as walking, a shorter pharynx would shape of the SVT.89 According to the ities. Most efforts have been dis- help increase turbulent airflow in the quantal theory of speech,90,91 these proved. Originally, it was suggested nasopharynx, thereby increasing the formant combinations are stable over that the angle of the cranial base, efficiency of the respiratory epithe- a range of tongue positions, requiring both internal and external, was lium, especially for recovering mois- less muscular coordination because related to the position of the hyoid, ture. Moreover, during vigorous activ- of two key properties of the SVT (Fig. based on evidence that the larynx ity such as endurance running, when 6). First, the SVT has two tubes of descends postnatally relative to the humans uniquely switch to oral breath- similar length: the horizontal tube, as the cranial base 8,97,98 ing, a shortened oral cavity may help SVTh, extends from our to the flexes. This hypothesis, however, to improve the efficiency with which back of the oropharynx; the vertical was disproved by evidence that there 87,88 we dump heat. One cost of dump- tube, SVTv, extends from the soft pal- is no correlation during human on- ing heat, however, is a greater reli- ate to the vocal folds. Second, move- togeny between laryngeal descent and ance on water. ments of the tongue and can cranial base angulation.41 In another modify the cross-sectional area of effort, Arensburg and colleagues99 each tube independently by a ratio of suggested that, based on its human- approximately 10:1.92 like , the Neanderthal hyoid Vocalization Although other mammals vocalize, had a human-like position. This infer- A final function to consider is vocal- only humans have an SVT with 1:1 ence, however, is unjustified because ization. I do so, however, with appre- proportions and in which the tongue there is no evidence to link hyoid hension because the evolution of can modify their cross-sections inde- shape with hyoid position in humans human speech is among the most pendently. In nonhuman mammals or other species. contentious subjects in paleoanthro- with snouts, the SVTh is relatively One possibility, however, that re- pology. One problem is that there is long because the lower face is long; mains untested is that an anteropos- no agreement on what null hypothe- the SVTv is short because the hyoid is teriorly shorter face does improve sis to reject. To some, language and positioned relatively high relative to speech performance capabilities in fully modern speech is a unique the mandible. As Negus93 noted, this modern versus archaic humans by human apomorphy unless proved configuration permits the epiglottis, reducing SVTh relative to the SVTv.In otherwise; others assume that large- which is suspended from the hyoid, a normal human with an SVT of brained hominids such as Neander- to contact the soft palate, forming a approximately 17 mm, the length of thals must have possessed human- ‘‘tube within a tube.’’ The advantage each portion of the SVT is about 8.5 like linguistic abilities unless proved of this configuration is that air can cm. In a modern human with a nor- otherwise. Given that neither speech travel directly between the nasophar- mally long neck, this places the hyoid nor the tissues that produce it fossil- ynx and the trachea through the por- about 2–3 cm below the lower margin ize, these are tough null hypotheses tion of the pharynx in which food and of the mandible94 and the larynx to reject. liquids also travel, thereby minimiz- about the level of the 6th cervical ver- That said, let’s throw caution to the ing the risk of asphyxiation or aspira- tebra. If the human SVTh were about wind and consider how a shorter face tion. But during human postnatal on- 2 cm longer, as is the case for a Nean- and/or more flexed basicranium togeny, the hyoid descends relative to derthal then the SVTv would also might influence speech production. the soft palate, losing contact by have to be 2 cm longer to maintain a To do so, it may be useful to begin about 3 months. The SVTh/SVTv ratio 1:1 SVTh/SVTv. This presents a poten- with a short review of the acoustical is 1.4 in a human infant, but reaches tial anatomical problem because it and anatomical bases of speech. 1.1 by approximately 7–8 years.94 As would require a correspondingly lon- Speech sounds are essentially puffs of the ratio approaches 1.0, speech ger neck with taller pressurized air from the lung that becomes more quantal and percep- to prevent the larynx from being posi- pass through the vocal folds of the tion errors decrease.95,96 Because tioned in the chest below the 7th cer- larynx, which controls the vibrations. descent of the larynx leads to a trade- vical . Such a position would The frequency of the vibrations, typi- off between intelligible speech and prevent infrahyoid muscles such as cally between 250 and 4,000 Hz, risk of asphyxiation, the low human the sternothyroid from being depres- determines the pitch, also known as hyoid is evidence of selection for sors of the hyolaryngeal complex. the fundamental frequency (F0); the speech capabilities at the expense of While it is theoretically possible that amplitude of the vibrations deter- other fitness costs.92 archaic humans had relatively longer mines the volume. The sound waves Note that a 1:1 SVTh/SVTv ratio is necks than modern humans do, are subsequently filtered by the air- not necessary for articulate, intelli- measurements of Neanderthal cervi- way, technically known as the supra- gible speech, but influences the cal vertebrae indicate that they had laryngeal vocal tract (SVT), to create degree to which speech is quantal. necks that were similar or possibly ARTICLES Speculations on Modern Human Craniofacial Form 65 even slightly shorter than those of cranial fossa size, as the latter is the ning may have played in past hunter- modern humans.100,101 only way to infer temporal lobe vol- gatherer behavior and the ways in In short, it is reasonable to ume from fossils. Along the same which aspects of modern human skel- hypothesize that two derived aspects lines, it would be useful to test if hu- etal morphology would have improved of the modern human cranium would mans with relatively larger or smaller ER performance. have improved modern human per- temporal lobes have more or less A fourth problem is speech, which formance capabilities for producing flexed cranial bases. may be the most intractable of all more quantal vowels that would be A second issue to test is the extent problems related to modern human more stable and less susceptible to to which decreases in facial size may origins. The biggest immediate prob- errors of perception. First, the mod- be related functionally and develop- lem is to find some reliable way to ern human oral cavity is absolutely mentally to changes in food prepara- estimate the proportions of the SVT shorter by at least a centimeter. Sec- tion techniques. Testing this hypothe- using skeletal data. The preceding ond, the modern human cranial base sis more thoroughly will require more analysis is not only informal and is more flexed, thereby rotating the information on how strains are gener- unquantitative, but also avoided mak- whole face including the back of ated in the uniquely shaped human ing actual estimates of SVTv length. the palate ventrally underneath the face and the extent to which pound- Instead, it examined the effects of anterior cranial base, decreasing the ing versus cooking influences the bio- possible maximum SVTv length on length of the pharyngeal space behind mechanics of force resistance in the SVT proportions. Moreover, such the palate by another centimeter.22 face. In addition, we need more data analyses only address one aspect of The combined effect of these two on the evidence of cooking versus speech, quantal vowel production, shifts is to shorten SVTh by approxi- other forms of food preparation from and have little to tell us about the mately 2 cm. Of course a longer SVTh the archeological and fossil records. other cognitive aspects of language and a non-1:1 SVTh/SVTv ratio does Another key source of evidence lies in that cannot be addressed using fossil not allow one to infer that archaic teeth, the size of which may affect oc- data. Although the archeological re- humans couldn’t speak or even lacked clusal performance differently during cord is frequently used to make infer- sophisticated language. But, given the eating of cooked, pounded, and ences about language capabilities, it constraints on larynx position in the raw food.63 must be emphasized that such infer- neck, combined with the risks of A third issue that requires further ences are fraught with complications. asphyxiation that presumably increase analysis is endurance running, as this Evidence of symbolism appears to be with greater separation of the epiglot- behavior may relate to both head bal- coincident with complex cognition, tis and soft palate (a hypothesis that ance and pharyngeal shortening. In but only up to a point. The absence of needs to be tested), it is not unreason- addition to understanding more evidence of symbolic art should never able to speculate that a shorter face about the biophysics of airflow dur- be used as evidence of the absence of might have been selected for speech ing respiration and its thermoregula- abstract cognitive tasks such as lan- performance. To test this hypothesis, tory effects, we need to understand guage. Such logic might also lead us however, we need better data on neck more thoroughly the challenges that to conclude falsely that industrial length in Neanderthals and other ar- the heel-strike transient poses for modern humans have more complex chaic humans, as well as more reli- head pitching during running, and cognitive capabilities than did Paleo- able indicators of the position of the how the unique configuration of the lithic hunter-gatherers or Neolithic hyoid and larynx in fossil humans. human head, neck, and upper body farmers. Also, estimates of SVT proportions do counteract these forces. In addition, Finally, with regard to all the pre- not address the various other neuro- more data are needed on the impor- ceding hypotheses, there is the persis- logical bases of language and speech tance of running in human evolution. tent problem of adaptations versus that must also have been targets of This is not an easy issue to address spandrels. Many of the features I have selection at some point in human evo- because, since the invention of the discussed are related to more than lution. bow and arrow about 20,000 years one function, and they have complex ago, running has almost certainly developmental bases. Thus, a persis- become less important in human for- tent problem is to figure out ways to FINAL SPECULATIONS aging societies. Modern human distinguish between spandrels and As noted, the various hypotheses I hunter-gatherers such as the Hadza adaptations. As noted, one key issue have discussed are only speculations; and the Bushmen do run occasion- is choosing appropriate null hypothe- much research is necessary to test ally,76,102,103 but ER is no longer as ses to reject when testing whether them. The list of needed tests is long necessary as it once may have been. features are adaptations, spandrels, and challenging, but a few are espe- Nonetheless, human ER perform- or neither. cially important. First, in order to test ance, which is as impressive as or sur- In spite of all these caveats and if temporal lobe volume is relatively passes those of the best cursors, and problems, let us conclude by assum- larger in modern than archaic hu- relies on largely different biome- ing, blithely, that the foregoing specu- mans, data are needed on the strength chanics than walking, demands some lations are corroborated. How, then, of the relationship, if any, between explanation. We thus have to think might they fit into a broader evolu- temporal lobe volume and middle creatively about the roles that run- tionary scenario to explain how selec- 66 Lieberman ARTICLES tion might have operated in the origin doubtful that modern hunters could human skull form that we first see in of H. sapiens some 200,000 years ago spend a large proportion of the day modern H. sapiens about 200,000 in Africa? Opinions on this question hunting and still satisfy their nutrient years ago,104 and which appears more are likely to be varied, but the various needs with a chimpanzee-like diet or less coincidentally in the archeo- major derived features of the H. sapi- upon returning to camp. Cooking logical record.105 Testing this hypoth- ens skull all indicate some combina- thus enables modern human hunters esis and its components will be a seri- tion of changes in three aspects of and their families to rely on meat, ous but enjoyable challenge. behavior: improved running capabil- which has a lower probability of suc- ities, including better head balancing cessful acquisition but a higher and thermoregulation; increased reli- potential yield, and still be able to eat ACKNOWLEDGMENTS ance or dependence on cooking; and enough when they get back to camp I am grateful to many colleagues some suite of cognitive shifts in terms empty-handed. with whom I have discussed these of speech and possibly other aspects ideas, especially Ofer Bar-Yosef, Den- of complex reasoning, perception, nis Bramble, Philip Lieberman, Peter and/or knowledge (admittedly, this is the various major Lucas, Rob McCarthy, David Pil- highly speculative). To me, these derived features of the beam, David Raichlen, and Richard shifts all point to a way of life, still Wrangham. Thanks also to John evident in hunter-gatherer societies, H. sapiens skull all Fleagle and Richard Leakey for the that may have differed between mod- indicate some invitation to speculate without the ern and archaic Homo. Notably, mod- usual constraints. ern human foragers rarely face star- combination of changes vation because they rely on social net- in three aspects of works and cooperative relationships REFERENCES within a large group of nonkin that behavior: improved involves considerable sharing of 1 Pearson OM. 2000. Postcranial remains and running capabilities, the origin of modern humans. Evol Anthropol meat, along with a diversity of other 9:229–247. foods. Interestingly, many of the fac- including better head 2 Howells WW. 1989. Skull shapes and the map: tors that underlie this overall strategy balancing and craniometric analyses in the dispersion of mod- are linked, for several reasons, by ern Homo. Peabody Museum Papers 79:1–189. thermoregulation; 3 Lieberman DE. 1995. Testing hypotheses endurance running, cooking, and lan- about recent human evolution from skulls. Curr guage. increased reliance or Anthropol 36:159–197. First, until the invention of modern 4 Lahr MM. 1996. The evolution of modern dependence on human cranial diversity: a study in cranial vari- technologies such as the bow and ation. Cambridge: Cambridge University Press. arrow, endurance running was prob- cooking; and some suite 5 Schwartz JH, Tattersall I. 2003. Craniodental ably useful in order to hunt and scav- of cognitive shifts morphology of genus Homo (Africa and Asia). enge successfully with low risk of New York: Wiley-Liss. 6 Gould SJ, Lewontin R. 1979. The spandrels of physical harm. Second, while hunting San Marco and the Panglossian paradigm: a is critical for modern humans who critique of the adaptationist programme. Proc require very high-quality diets relative Finally, this sort of way of life R Soc Lond B. 205:581–598. to body mass, hunting is also a low- requires language and other cognitive 7 Lauder GV. 1996. The argument from design. skills in order to maintain social rela- In: Rose, M, Lauder GV, editors. Adaptation. percentage endeavor; it cannot be New York: Academic Press. p 55–91. tionships and to foster among nonkin accomplished without food sharing 8 Lieberman P, Crelin ES. 1971. On the speech and cooking. Without the existence of the cooperation necessary for food of Neanderthal man. Linguistic In 2:201–222. reliable sharing, hunters who return sharing. These cognitive skills, doubt- 9 Daegling DJ. 1993. Functional morphology of the human chin. Evol Anthropol 1:170–177. to camp empty-handed would suffer less, were also advantageous for many other key tasks such as tracking 10 Cheverud JM. 1982. Phenotypic, genetic, reduced fitness, as would their fami- and environmental integration in the cranium. lies. In addition, there is a good argu- animals and monitoring resources in Evolution 36:499–516. ment to be made that cooking is an complex changing habitats. 11 Cheverud JM. 1995. Morphological integra- In short, although archaic humans tion in the saddle-back tamarin (Saguinus fusci- important complementary strategy to collis) cranium. Am Nat 145:63–89. may have done everything that mod- food sharing. Recall that chimpan- 12 Lieberman DE, Pearson OM, Mowbray KM. zees, like many mammals, spend at ern humans did (hunt, speak, share, 2000. Basicranial influence on overall cranial shape. J Hum Evol 38:291–315. least half their day eating, largely and even cook), it is interesting 13 Gonzalez-Jose R, Van Der Molen S, Gonza- because they need to fill their stom- to speculate that modern humans lez-Perez E, Hernandez M. 2004. Patterns of achs regularly (about every two may have invented a more intense, phenotypic covariation and correlation in mod- socially cooperative, and technologi- ern humans as viewed from morphological inte- hours) with relatively high-fiber-con- gration. Am J Phys Anthropol 123:69–77. tent food. As Wrangham and Conklin- cally based way of using these skills. 14 Ackermann RR. 2005. Ontogenetic integration Brittain61 noted, cooking not only This combination of hunting, shar- of the hominoid face. J Hum Evol 48:175–97. improves the nutrient content of ing, cooking, communicating, and 15 Lieberman DE, Ross CF, Ravosa MJ. 2000. The primate cranial base: ontogeny, function, food, but also decreases digestion cooperating may have provided the and integration. Yearbook Phys Anthropol time. Without cooked food, it is behavioral context for selection on 43:117–169. ARTICLES Speculations on Modern Human Craniofacial Form 67

16 Hallgrı´msson B, Lieberman DE, Liu W, primate temporal lobe. J Hum Evol 42:505– essing on masticatory strain and craniofacial Ford-Hutchinson AF, Jirik FR. 2007. Epigenetic 533. growth in a retrognathic face. J Hum Evol 46: interactions and the structure of phenotypic 35 Schenker NM, Desgouttes AM, Semendeferi 655–677. variation in the cranium. Evol Dev 9:76–91. K. 2005. Neural connectivity and cortical sub- 54 Thomason JJ, Russell AP. 1986. Mechanical 17 Lieberman DE, McBratney BM, Krovitz G. strates of cognition in hominoids. J Hum Evol factors in the evolution of the mammalian sec- 2002. The evolution and development of cranial 49:547–569. ondary palate: a theoretical analysis. J Morphol form in Homo sapiens. Proc Natl Acad Sci USA 36 Holloway RL, Broadfield DC, Yuan MS, 189:199–213. 99:1134–1139. Schwartz JH, Tattersall I. 2004. The human fos- 55 Rafferty KL, Herring SW, Marshall CD. 18 Lieberman DE, Krovitz G, McBratney-Owen sil record, vol. 3, brain endocasts: the Paleon- 2003. Biomechanics of the rostrum and the role BM. 2004. Testing hypotheses about tinkering eurological evidence. New York: Wiley. of facial sutures. J Morphol 257:33–44. in the fossil record: the case of the human 37 Semendeferi K, Lu A, Schenker N, Damasio 56 Demes B. 1987. Another look at an old face: skull. J Exp Zool (Mol Dev Evol) 302B:284–301. H. 2002. Humans and great apes share a large biomechanics of the Neanderthal facial skeleton 19 Lieberman DE. 2000. Ontogeny, homology, frontal cortex. Nature Neurosci 5:272–276. reconsidered. J Hum Evol 16:297–303. and phylogeny in the hominid craniofacial 38 Schoenemann PT, Sheehan MJ, Glotzer LD. 57 Lieberman DE. 1998. Sphenoid shortening skeleton: the problem of the browridge. In: 2005. Prefrontal white matter volume is dispro- and the evolution of modern human cranial O’Higgins P, Cohn M, editors. Development, portionately larger in humans than in other pri- shape. Nature 393:158–162. growth and evolution: implications for the mates. Nature Neurosci 8:242–252. 58 Carlson DS. 1976. Temporal variation in study of hominid skeletal evolution. London: 39 Ross CR, Ravosa MJ. 1993. Basicranial flex- prehistoric Nubian crania. Am J Phys Anthro- Academic Press. p 85–122. ion, relative brain size, and facial kyphosis in pol 45:467–484. 20 Lieberman DE, Bar-Yosef O. 2005. Apples nonhuman primates. Am J Phys Anthropol 59 Carlson DS, Van Gerven DP. 1977. Mastica- and oranges: morphological versus behavioral 91:305–324. tory function and post-Pleistocene evolution in transitions in the Pleistocene. In: Lieberman 40 McCarthy RC. 2001. Anthropoid cranial base Nubia. Am J Phys Anthropol 46:495–506. DE, Smith RJ, Kelley J, editors. Interpreting the architecture and scaling relationships. J Hum past: essays on human, primate and mammal 60 Wrangham RW, Jones JH, Laden G, Pilbeam Evol 40:41–66. evolution. Boston: Brill Academic Publishers. D, Conklin-Brittain NL. 1999. The raw and the p 275–296. 41 Lieberman DE, McCarthy RC. 1999. The on- stolen: cooking and the ecology of human ori- togeny of cranial base angulation in humans gins. Curr Anthropol 40:567–594. 21 Enlow DH. 1990. Facial growth, 3rd ed. Phil- and chimpanzees and its implications for adelphia: WB Saunders. 61 Wrangham RW, Conklin-Brittain NL. 2003. reconstructing pharyngeal dimensions. J Hum Cooking as a biological trait. Comp Biochem 22 McCarthy RC, Lieberman DE. 2001. The Evol 36:487–517. Physiol 136A:35–46. posterior maxillary (PM) plane and anterior 42 Persinger MA. 2001. The neuropsychiatry of cranial architecture in primates. Anat Rec 264: 62 Wandsnider L. 1997. The roasted and the paranormal experiences. J Neuropsychol Clin 247–260. boiled: food composition and heat treatment Neurosci 13:515–524. with special emphasis on pit-hearth cooking. 23 Bastir M, Rosas A. 2004. Facial heights: evo- 43 Brace CL, Smith SL, Hunt KD. 1991. What J Anthropol Archeol 16:1–48. lutionary relevance of postnatal ontogeny for fa- big teeth you had grandma! size, cial orientation and skull morphology in humans 63 Lucas PW. 2004. How teeth work. Cam- past and present. In: Kelley MA, Larsen CS, edi- and chimpanzees. J Hum Evol 47:359–381. bridge: Cambridge University Press. tors. Advances in dental anthropology. New 64 James SR. 1989. Hominid use of fire in the 24 Biegert J. 1963. The evaluation of character- York: Wiley-Liss. p 33–57. istics of the skull, , and feet for primate Lower and Middle Pleistocene: a review of the 44 Demes B, Creel N. 1988. Bite force, diet and taxonomy. In: Washburn SL, editor. Classifica- evidence. Curr Anthropol 30:1–26. cranial morphology in fossil hominids. J Hum tion and human evolution. Chicago: Aldine de 65 Brace CL. 1995. The stages of human evolu- Evol 17:657–670. Gruyter. p 116–145. tion. Englewood Cliffs: Prentice-Hall. 45 Stedman HH, Kozyak BW, Nelson A, Thesier 25 Ross CF, Henneberg M, Richard S, Ravosa 66 Rowlett RM. 2000. Fire control by Homo DM, Su LT, Low DW, Bridges CR, Shrager JB, MJ. 2004. Curvilinear and geometric modeling erectus in Africa and East Asia. Acad Anthropol Minugh-Purvis N, Mitchell MA. 2004. Myosin gene of basicranial flexion in a phylogenetic context: Sinica 19:198–208. mutation correlates with anatomical changes in is it adaptive? is it constrained? J Hum Evol the human lineage. Nature 428:415–418. 67 Goldberg P, Weiner S, Bar-Yosef O, Xu Q, 46:185–213. Liu J. 2001. Site formation processes at Zhou- 46 Hylander WL, Johnson KR. 1984. Jaw mus- 26 Spoor F, O’Higgins P, Dean C, Lieberman koudian, China. J Hum Evol 41:483–530. cle function and wishboning of the mandible DE. 1999. Anterior sphenoid in modern humans. during mastication in macaques and baboons. 68 Anto´ n SC. 1994. Biomechanical and other Nature 397:572. Am J Phys Anthropol 94:523–547. perspectives on the Neandertal face. In: Corruc- 27 Ruff CB, Trinkaus E, Holliday TW. 1997. 47 Hylander WL, Johnson, KR, Crompton AW. cini R, Ciochon R, editors. Integrated pathways Body mass and encephalization in Pleistocene 1987. Loading patterns and jaw movement dur- to the past, Festschrift in honour of F.C. Howell. Homo. Nature 387:173–176. ing mastication in Macaca fascicularis: a Saddle River NJ: Prentice-Hall. p 677–695. 28 Holloway RL. 1985. The poor brain of Homo strain, electromyographic and cineradiographic 69 O’Connor CF, Franciscus RG, Holton NE. sapiens neanderthalensis: see what you please. analysis. Am J Phys Anthropol 72:287–314. 2005. Bite force production capability and effi- In: Delson E, editor. Ancestors: the hard evi- 48 Hylander WL, Picq PG, Johnson KR. 1991. ciency in Neandertals and modern humans. Am dence. New York: A. Liss. p 319–324. Masticatory-stress hypotheses and the supraorbital J Phys Anthropol 127:129–151. 29 Opperman L. 2000. Cranial sutures as intra- region of primates. Am J Phys Anthropol 86:1–36. 70 Lieberman DE, Raichlen DA, Pontzer H, membranous bone growth sites. Dev Dyn 219: 49 Hylander WL, Picq PG, Johnson KR. 1992. Bramble DM, Cutright-Smith E. 2006. The 472–485. Bone strain and the supraorbital region of pri- human gluteus maximus and its role in run- 30 Wilkie AO, Morriss-Kay GM. 2001. Genetics mates. In: Carlson DS, Goldstein SA, editors. ning. J Exp Biol 209:2143–2155. of craniofacial development and malformation. Bone biodynamics in orthodontic and ortho- 71 Graf W, de Waele C, Vidal PP. 1995. Func- Nat Rev Genet 2:458–468. paedic treatment. Ann Arbor: Craniofacial tional anatomy of the head-neck movement sys- 31 Deacon TW. 1997. The symbolic species: the Growth Series, Center for Human Growth and tem of quadrupedal and bipedal mammals. J co-evolution of language and the brain. New Development, Vol. 27. p 315–349. Anat 186:55–57. York: WW Norton. 50 Wrangham RW. 1977. Feeding behaviour of 72 Strait D, Ross CF. 1999. Kinematic data on 32 Semendeferi K. 2001. Advances in the study chimpanzees in Gombe National Park, Tanza- primate head and neck posture: implications of hominoid brain evolution: magenetic reso- nia. In: Clutton Brock TH, editor. Primate ecol- for the evolution of basicranial flexion and an nance imaging (MRI) and 3-D imaging. In: ogy. London: Academic Press, p 503–538. evaluation of registration planes used in paleo- Falk D, Gibson K, editors. Evolutionary anat- 51 Gross MD, Arbel, G. Hershkovitz I. 2001. anthropology. Am J Phys Anthropol 108:205– omy of the primate cerebral cortex. Cambridge: Three-dimensional finite element analysis of 222. Cambridge University Press. p 257–289. the facial skeleton on simulated occlusal load- 73 Dunbar DC. 2004. Stabilization and mobility 33 Semendeferi K, Armstrong E, Schleicher A, ing. J Oral Rehabil 28:684–689. of the head and trunk in vervet monkeys (Cerco- Zillesn K, Van Hoesen GW. 2001. Prefrontal cor- 52 Ross CF. 2001. In vivo function of the cra- pithecus aethiops) during treadmill walks and tex in humans and apes: a comparative study of niofacial haft: the interorbital ‘‘pillar’’. Am J gallops. J Exp Biol 207:427–438. area 10. Am J Phys Anthropol 114:224–241. Phys Anthropol 116:108–139. 74 Demer JL, Viirre ES. 1996. Visual-vestibular 34 Rilling JK, Seligman RA. 2002. A quantita- 53 Lieberman DE, Krovitz GE, Yates FW, Dev- interaction during standing, walking, and run- tive morphometric comparative analysis of the lin M, St Claire M. 2004. Effects of food proc- ning. J Vestib Res 6:295–313. 68 Lieberman ARTICLES

75 Grossman GE, Leigh RJ, Abel LA, Lanska airflow dynamics in the . Am J 97 Lieberman P, Crelin ES, Klatt DH. 1972. DJ, Thurston SE. 1988. Frequency and velocity Hum Biol 16:625–638. Phonetic ability and related anatomy of new- of rotational head perturbations during locomo- 87 Ferrus L, Commenges D, Gire J, Varene P. born and adult humans, Neanderthal man and tion. Exp Brain Res 70:470–476. 1984. Respiratory water loss as a function of the chimpanzee. Am Anthropol 74:287–307. 76 Bramble DM, Lieberman DE. 2004. Endur- ventilatory or environmental factors. Resp 98 Laitman JT, Heimbuch RC, Crelin ES. 1979. ance running and the evolution of Homo. Physiol 56:11–20. The basicranium of fossil hominids as an indi- Nature 432:345–352. 88 Varene P, Ferrus L, Manier G, Gire J. 1986. cator of their upper respiratory tract. Am J 77 Spoor F, Wood B, Zonneveld F. 1994. Impli- Heat and water respiratory exchanges: compari- Phys Anthropol 51:15–34. cations of early hominid labyrinthine morphol- son between mouth and nose breathing in 99 Arensburg B, Tillier AM, Vandermeersch B, ogy for evolution of human bipedal locomotion. humans. Clin Physiol 6:405–414. Duday H, Schepartz LA, Rak Y. 1989. A Middle Pa- laeolithic human hyoid bone. Nature 338: 758–760. Nature 369:645–648. 89 Peterson GE, Barney HL. 1952. Control 78 Dimery NJ, Alexander RM, Deyst KA. 1985. methods used in a study of the vowels. J Acous- 100 Trinkaus E. 1983. The Shanidar Neander- Mechanics of the ligamentum nuchae of some tical Soc Am 24:175–184. tals. New York: Academic Press. artiodactyls. J Zool, Lond 206:341–351. 90 Stevens KN. 1989. On the quantal nature of 101 McCarthy RC, Strait D, Lieberman P, Yates 79 Mercer SR, Bogduk N. 2003. Clinical anat- speech, J Phonet 17:3–46. F. n.d. A recent origin for modern human speech capabilities. Science. omy of the ligamentum nuchae. Clin Anat 16: 91 Stevens KN. 1972. The quantal nature of 484–93. speech: evidence from articulatory-acoustic data. 102 Liebenberg L. 1990. The art of tracking: 80 Schultz AH. 1942. Conditions for balancing In: Denes PB, David EE Jr, editors. Human com- the origin of Science. Claremont South Africa: the head in primates. Am J Phys Anthropol munication: a unified view. New York: McGraw- David Philip. 29:483–497. Hill. p 51–66. 103 O’Connell JF, Hawkes K, Blurton-Jones 81 Vogel S. 2004. Comparative biomechanics. 92 Lieberman P. 2006. Towards an evolutionary NG. 1988. Hadza scavenging: implications for Plio-Pleistocene hominid subsistence. Curr Princeton: Princeton University Press. biology of language. Cambridge: Harvard Uni- Anthropol 29:356–363 82 Schmidt-Nielsen K. 1984. Scaling: why is versity Press. 104 McDougall I, Brown FH, Fleagle JG. 2005. animal size so important? Cambridge: Cam- 93 Negus V. 1949. The comparative anatomy Stratigraphic placement and age of modern hu- bridge University Press. and physiology of the larynx. New York: Hafner mans from Kibish, Ethiopia. Nature 433:733–736. 83 Margaria R, Cerretelli P, Aghemo P, Sassi G. Publishing. 105 McBrearty S, Brooks A. 2000. The revolu- 1963. Energy cost of running. Appl Physiol 18: 94 Lieberman DE, McCarthy RC, Hiiemae KM, tion that wasn’t: a new interpretation of the ori- 367–370. Palmer JB. 2001. Ontogeny of postnatal hyoid gin of modern human behavior. J Hum Evol 84 Hahn I, Scherer PW, Mozell MM. 1993. and larynx descent in humans. Arch Oral Biol 39:453–563. 46:117–128. Velocity profiles measured for airflow through 106 Corruccini RL. 1999. How anthropology a large-scale model of the human nasal cavity. 95 Lieberman P. 1980. On the development of informs the orthodontic diagnosis of malocclu- J Appl Physiol 75:2273–2287. vowel production in young children. In: Yeni- sion’s causes. Lewiston: Edwin Mellen Press. 85 Franciscus RG, Trinkaus E. 1988. Nasal Komshian G, Kavanagh J, editors. Child pho- nology, perception and production. New York: 107 Brace CL. 2000. Evolution in an Anthropo- morphology and the emergence of Homo. J logical View. Lanham: AltaMira Press. Hum Evol 75:517–527. Academic Press. p 113–142. 86 Churchill SE, Shackelford LL, Georgi JN, 96 Buhr RD. 1980. The development of vowels Black MT. 2004. Morphological variation and in an infant. J Speech Hearing Res 23:75–94. VC 2008 Wiley-Liss, Inc.

Articles in Forthcoming Issues

• The Primate Palette: The Evolution of Primate Coloration Brenda J. Bradley and Nicholas I. Mundy • The Semicircular Canal System and Locomotion: The Case of Extinct Lemuroids and Lorisoids Alan Walker, Timothy M. Ryan, Mary T. Silcox, Elwyn Simons, and Fred Spoor • Towards a Map of Capuchin Monkeys’ Tool Use Eduardo B. Ottoni and Patricia Izar • Timing Primate Evolution: Lessons From the Discordance Between Molecular and Paleontological Estimates M. E. Steiper and N. M. Young • All Roads Lead To ... Everywhere? Kenneth M. Weiss