Human origins: Out of AfricaIan Tattersall1

Division of Anthropology, American Museum of Natural History, New York NY 10024

Edited by Richard G. Klein, Stanford University, Stanford, CA, and approved July 16, 2009 (received for review March 23, 2009)

Our species, Homo sapiens, is highly autapomorphic (uniquely derived) among hominids in the structure of its skull and postcranialskeleton. It is also sharply distinguished from other organisms by its unique symbolic mode of cognition. The fossil and archaeologi-cal records combine to show fairly clearly that our physical and cognitive attributes both first appeared in Africa, but at differenttimes. Essentially modern bony conformation was established in that continent by the 200–150 Ka range (a dating in good agree-ment with dates for the origin of H. sapiens derived from modern molecular diversity). The event concerned was apparently short-term because it is essentially unanticipated in the fossil record. In contrast, the first convincing stirrings of symbolic behavior are notcurrently detectable until (possibly well) after 100 Ka. The radical reorganization of gene expression that underwrote the distinctivephysical appearance of H. sapiens was probably also responsible for the neural substrate that permits symbolic cognition. This exap-tively acquired potential lay unexploited until it was ‘‘discovered’’ via a cultural stimulus, plausibly the invention of language. Mod-ern humans appear to have definitively exited Africa to populate the rest of the globe only after both their physical and cognitivepeculiarities had been acquired within that continent.

Homo sapiens � evolution � fossil record � symbolic cognition

Africa is in a profound sense thefount of human evolution. Notonly did our zoological familyHominidae (Homo sapiens plus

its extinct close relatives, often nowa-days restricted to the subfamily Homini-nae; for the purposes of this article thedifference is merely notional) originatethere ca. 7 Ma (1), but over the past 2Ma the continent has regularly pumpedout new kinds of hominid into otherareas of the Old World (2). The genusHomo evolved in Africa at some timeca. 2 Ma [all older contenders to Homostatus are debatable (3, 4)], then rapidlyspread out of its natal continent to pop-ulate Eurasia for the first time (5, 6).The first truly cosmopolitan species ofHomo, Homo heidelbergensis, is firstknown from Africa at ca. 600 Ka (7),before appearing at sites in Europe andeastern Asia from ca. 500 Ka onward.The now-ubiquitous species H. sapiens,to which all living human beings belong,is initially documented in Africa as,somewhat later, is the first material evi-dence of the symbolic cognitive systemthat appears to be unique to humans.

Modern H. sapiens Is Highly Derived inIts OsteologyMorphologically, our living species H.sapiens is extremely distinctive. It is notunique among hominids in having a largebrain [averaging �1350–1400 mL in vol-ume (8)]; but it is unique in the propor-tions of the skull in which that brain ishoused, and in numerous smaller-scalecranial characteristics (4, 9). Among otherfeatures not found elsewhere, H. sapienspossesses a short, tall and more or lessglobular braincase, beneath the front ofwhich a small, anteroposteriorly short anddelicately built face is distinctly retracted(10). The orbits are surmounted by indi-

vidual supraciliary ridges; although notinvariably tiny, these are bipartite, with acentral portion separated from a lateralplate by an oblique crease (4, 11). In thelower jaw, the H. sapiens chin is not sim-ply a swelling at the external base of thesymphysis [which can be found elsewhereamong hominids (4, 12)]. Instead, it is acomplex structure in the form of an in-verted ‘‘T,’’ in which a vertical keelbounded by lateral depressions meets abasal transverse bar running between lat-eral tubercles (4, 12).

H. sapiens is equally derived in thestructure of its postcranial skeleton. Forexample, in sharp contrast to the re-cently reconstructed skeleton of Homoneanderthalensis (13), that of modernhumans is slender and delicately built;and although the Neanderthal rib cageis conical, tapering distinctly upwardsfrom a broad base that matches themarkedly flaring iliac blades of the pel-vis, in H. sapiens the thorax is barrel-shaped. It is relatively narrow and tapersinward at the bottom and at the top,while the relatively delicate pelvis belowit lacks lateral f lare and has notablymore vertical iliac blades. The dissimi-larity between the two species is strik-ing, and may have affected gait andexternal appearance (14). Nonetheless,until recently there was room for uncer-tainty over which thoracic/pelvic condi-tion was derived within the genusHomo.

The two best pelvic specimens re-ported for species of Australopithecus(15, 16) made it clear that a broad, f lar-ing pelvis is primitive for Hominidae;but whether pelvic f lare is also primitivefor the genus Homo [defined as thosehominids possessing essentially modernbody proportions (3)] was less evident.In contrast, the best skeleton of an early

Homo (KNM-WT 15000 from WestTurkana in Kenya) shows a weakly coni-cal thorax and, as reconstructed, only amodestly wide pelvis (17). Still, it be-longed to an immature and thus incom-pletely developed individual. At thesame time, the excellently preserved andwidely flaring adult pelvis (SH Pelvis 1)from the Sima de los Huesos at Atapu-erca in Spain (18) is that of a Neander-thal relative, and does no more thanconfirm that this pelvic conformation isprimitive for the Neanderthal clade.However, a recently reported adult pel-vis from Gona in Ethiopia (19), dated tobetween 1.4 and 0.9 Ma and attributedby its describers to the same species asthe Turkana specimen, joins more lim-ited materials described earlier (20, 21)in showing great robusticity and thebroadly flaring conformation. Availableevidence thus now strongly suggests thatthe wide, f lat, heavy pelvic morphologyis indeed primitive for the genus Homo,in which case, the basic body form of H.sapiens, as well as that of its skull, ishighly derived.

The same can also be said for theunique mode of cognition possessed byall living H. sapiens. Alone among or-ganisms, as far as can be told, our spe-cies exhibits symbolic mental processes.That is to say, its members deconstructthe world around them into a huge vo-cabulary of mental symbols. These theycombine and recombine in imaginationto describe alternate worlds and situa-tions, based on a capacity for generatinga potentially infinite array of meanings

Author contributions: I.T. designed research, performedresearch, and wrote the paper.

The author declares no conflict of interest

This article is a PNAS Direct Submission.

1E-mail: iant@amnh.org.

16018–16021 � PNAS � September 22, 2009 � vol. 106 � no. 38 www.pnas.org�cgi�doi�10.1073�pnas.0903207106

from a finite set of elements (‘‘discreteinfinity’’; see ref. 22). As a result, unlikeother organisms including such cogni-tively sophisticated ones as chimpanzees,we live in worlds that are largely of ourown making. And despite our manifoldphysical distinctions from our closestliving relatives, it is clearly this cognitivedifference that makes us feel so differentfrom them.

Fossil Record of H. sapiensHow and where, then, were the distinc-tive features of H. sapiens acquired? Therelevant fossil record is fairly thin, but itdoes firmly establish that as an anatomi-cally recognizable entity our speciesmade its first appearance in Africa (23,24). The ‘‘Out of Africa’’ hypothesis ofmodern human origins emerged in themid-1980s, when paleoanthropologistssuch as Gunter Brauer in Germany(e.g., ref. 25) and Chris Stringer in theU.K. (e.g., ref. 26) began to point outthat, sparse as they were, the earliestfossils that resembled members of ourspecies came from southern and easternAfrica. The resultant notion of a ‘‘singleAfrican origin’’ for modern humansstood in contrast to ‘‘multiregional’’ in-terpretations, in which the major mod-ern geographical groups of H. sapienswere seen as having extremely deeproots in time (27). The single Africanorigin notion received an enormousboost from molecular systematics whenDNA comparisons (28, 29) beganstrongly supporting earlier conclusionsbased on proteins (30) that Africa hadbeen the ultimate source of modern hu-man populations worldwide. Over thelast quarter-century, evidence on boththe molecular and the fossil fronts hasaccumulated to the point where therecan be little doubt that humankind ulti-mately originated in Africa.

Beyond this, however, the picture is alittle hazy. To some extent, the perspec-tive is complicated by context, for al-though Homo neanderthalensis, for ex-ample, was clearly a member of a largerclade of species united by a suite ofreadily identifiable characters, the dis-tinctive H. sapiens is largely isolated,bereft of evident close fossil relatives.This isolation was obscured for manyyears by paleoanthropologists’ habit ofdesignating as ‘‘archaic H. sapiens’’ avery motley assortment of relativelylarge-brained hominids that are geologi-cally fairly recent, but that are clearlynot of modern H. sapiens morphology.In Africa, fossils assigned at one time oranother to this meaningless wastebaskettaxon include the very distinctive craniafrom Ngaloba, Ndutu, Kabwe, Florisbadand Jebel Irhoud. In blurring the veryclear morphological boundaries of our

living species the artificial construct of‘‘archaic H. sapiens’’ served to obscure apicture of considerable morphologicaland presumably also taxonomic hominidcomplexity in the later Pleistocene, andit is fortunate that most paleoanthro-pologists seem now to have recognizedits counterproductive nature.

Morphological delineation of H. sapi-ens is more legitimately complicated bya group of hominid fossils, mostly fromSouth Africa, that possess all or nearlyall of the diagnostic skull characters ofH. sapiens, except for the bipartite browand/or the chin. Among such specimensare Border Cave 5, Boskop, Fish Hoek,Klasies River Mouth (except for AP6222), and maybe Cave of Hearths (4,31). Most of these hominids are poorlydated; but some, at least, are probablyquite recent, and all have traditionallybeen accepted as modern H. sapiens.More broadly, this group also includessome very early fossils, such as the OmoKibish 1 specimen from southern Ethio-pia recently redated to 195 Ka (23), andmost likely also the �160 Ka adult cra-nium from the northern Ethiopian siteof Herto (24) and the �130 Ka Singapartial cranium from Sudan (32). Noneof these fossils fulfills all of the moststringent morphological criteria applica-ble to living H. sapiens, but the Hertoand Kibish fossils in particular demon-strate quite clearly that the unusual ba-sic morphology evident in our specieswas established in Africa in the periodfollowing 200 Ka. If there is any justifi-cation at all for recognizing a form wemight call ‘‘archaic H. sapiens,’’ this iswhere we find it.

Given the radical departure in cranialanatomy of these early African H. sapi-ens from other hominids known in thesame time range, from the same conti-nent (there is very little relevant post-cranial record, if any), one reasonableconclusion is that the new morphologyalready exemplified at Herto and Kibisharose in a single change in gene regula-tion, with cascading developmental ef-fects throughout the body. Significantly,there is no reason to suppose that theramifications of this event were limitedto the bony modifications observable inthe fossils themselves.

Evidence of BehaviorOutside Africa there are no convincingclaimants to H. sapiens status before ca.93 Ka, the date of a clearly modern hu-man skeleton (Qafzeh 9) from the Le-vantine site of Jebel Qafzeh (33). Thisfind significantly predates any equivalentevidence from eastern Asia and Europe,and molecular dating supports the sug-gestion that the Qafzeh occurrence rep-resents an initial foray of anatomical H.

sapiens out of Africa that was ultimatelyfoiled by climatic vicissitudes (34). Whatmakes the Qafzeh fossil particularly rel-evant to the origin of modern humanbehavior and cognition is that, despiteits modern morphology, it is associatedwith Middle Paleolithic stone tools moreor less identical with those manufac-tured by penecontemporaneous Nean-derthals in the same region (35).

A recurrent pattern in hominid his-tory has been a temporal disconnectbetween innovation in the anatomicaland behavioral realms. Thus, stone toolmaking was apparently invented by aus-tralopiths, whereas the achievement ofmodern body form was not accompa-nied by any detectable advance in mate-rial culture (36). The same patternseems to have held in the case of H.sapiens: Modern human anatomy ap-pears in the record significantly earlierthan any evidence of symbolic behav-iors. Of course, because behavior doesnot itself fossilize it has to be inferredfrom material proxies, many of whichare very arguable. Based on varioustechnological and economic indicators(blade production, use of grindstones,pigment processing, hafting) it has beenproposed that hints of ‘‘modern’’ humanbehaviors can be detected as far back asseveral hundred thousand years ago, andthat a sequential accumulation of suchindicators reflects a slow, gradual attain-ment of ‘‘modern’’ behavior patternsover the past 300 to 500 Ka (37, 38).There is no doubt that in such indicatorswe are seeing evidence of increasingbehavioral complexity (even as averagehominid brain sizes were tending to in-crease, albeit not demonstrably in a linearfashion); but it can also convincingly beargued that no strictly technological aspectof Paleolithic behavior can be taken byitself as prima facie evidence of symbolic(as opposed to more generally cognitivelycomplex) behaviors. In this perspective itis legitimate to conclude that symboliccognitive processes, and their correlates,can be reliably inferred only from the pro-duction of overtly symbolic objects (asopposed to those that are merely techno-logically complex).

Even if we do not adopt the morestringent requirement for inferring theappearance of human symbolic cogni-tion, we can nonetheless remark on theremarkably crude technological contextsin which the first anatomical H. sapiensare found. The few stone tools reportedalong with the Omo Kibish 1 craniumhave been described as ‘‘unremarkable’’(39), whereas those from the same de-posits as the Herto cranium are notablyarchaic, consisting of some of the latestrecorded African handaxes, plus someMiddle Stone Age elements, approxi-

Tattersall PNAS � September 22, 2009 � vol. 106 � no. 38 � 16019

SP

EC

IAL

FE

AT

UR

E:

PE

RS

PE

CT

IVE

mately equivalent to the productions ofNeanderthals (40). It is only very muchlater in time that we begin to pick upany artifacts whatever that may convinc-ingly be interpreted as the products ofsymbolic minds. Earlier potential indica-tors such as the pigment processing andshellfishing recently reported from thesite of Pinnacle Point on the southernAfrican coast at ca. 160 Ka (41) aresomewhat wishful markers for ‘‘modern’’behavior patterns, especially given thatboth activities are documented for thealmost certainly nonsymbolic Homo ne-anderthalensis (42). And the ‘‘symbolic’’organization of the living space reportedfor the �100 Ka sites of Klasies RiverMouth in South Africa (43) is necessar-ily inferential.

In light of all this, the most ancientobjects that many observers would ac-cept as clearly symbolic come from theyounger site of Blombos Cave, also onthe southern African coast (44). Recov-ered from Middle Stone Age layersdated to ca. 77 Ka, they consist of twosmall ochre plaques bearing engravedgeometric designs. The same depositsalso yielded small gastropod shells thatwere apparently pierced for stringing(45). Body ornamentation is widely con-sidered a reliable proxy for symbolicbehavior patterns, and the Blombos evi-dence is supported by similar ‘‘beads’’found at other African Middle StoneAge sites, including the 82 Ka Grottedes Pigeons in Morocco (46). Interest-ingly, a possible occurrence of similarkind has been reported outside Africa atthe �100 Ka Israeli site of Skhu� l (47).This adds a potential hint of complexityto the picture, although biogeographershave generally considered the Levant anextension of the African continent (48).Still, in the record as currently known,the first fully mature expressions of thehuman capacity do not appear until ca.35 Ka (49–51) in Europe, when an ex-

traordinary artistic f lowering testifies tolives that were drenched in symbol.However, there is no reason to expectthat all of the dimensions of the newhuman symbolic capacity should havebeen exploited at once, and it is at thevery least plausible that what we arewitnessing in the African Middle StoneAge are the first stirrings of a long pro-cess of cognitive discovery that is stillcontinuing today.

It is worth noting that shortly afterBlombos times southern Africa experi-enced an episode of aridification thatmay have largely or entirely depopulatedthe area for an extended period (43),implying that early symbolic expressionin this region may not have been lin-early ancestral to later such expressionselsewhere. Nonetheless, whatever thedetails of the evidently complex acquisi-tion of symbolic cognition in our speciesmight have been, it seems clear that theAfrica provided the stage on which thisradically new mode of processing infor-mation initially evolved.

ConclusionsEvidently, then, ‘‘becoming human’’took place in two separate stages. First,the distinctive modern human morphol-ogy became established, very clearly inAfrica, and probably shortly after 200Ka. This event involved a radical depar-ture from the primitive Homo bodyform. Only ca. 100 Ka later, again inAfrica, and in a Middle Stone Age in-dustrial context, did modern symbolic be-haviors begin to be expressed, underwrit-ten by a new capacity that had mostplausibly been present but unexploited inthe first anatomical H. sapiens. In evolu-tionary terms this disconnect was entirelyroutine, for every new behavior has to bepermitted by a structure that already ex-ists: Birds, for example, had feathers formillions of years before coopting them for

flight, and tetrapods acquired their limbsin an aquatic context (52).

Symbolic reasoning appears to bequalitatively different from all otherforms of cognition, including its ownimmediate precursor. Its neural sub-strate continues to be strenuously de-bated (53, 54); but, whatever it was, thatstructural innovation was most plausiblyacquired as part and parcel of the radi-cal biological reorganization that gavebirth to H. sapiens as an anatomicallydistinctive entity. In which case (likethose feathers and limbs) it remainedunexploited, at least in the cognitivecontext, for a very substantial length oftime, until its new use was ‘‘discovered’’by its possessor. How this discovery wasmade remains a matter for conjecture,but a leading candidate for the necessar-ily cultural stimulus to symbolic process-ing of information is the invention oflanguage (55). Language is perhaps theultimate symbolic activity; and, in con-trast to theory of mind, the other lead-ing candidate for the role of releaser(56), it has the advantage of being acommunal rather than an internalizedattribute. The ability to use languagedepended, of course, on the presence ofthe vocal structures required to producespeech; but clearly these had alreadybeen exaptively acquired by the earliestanatomical H. sapiens.

Current evidence thus indicates thatH. sapiens as we know it today had adual origin: first as an anatomical entity,and only subsequently as a cognitiveone. The clear signal of both the fossiland archaeological records is that bothinnovations occurred in Africa, fromwhich the first fully modern humans ex-panded relatively recently to populatethe rest of the world.

ACKNOWLEDGMENTS. I thank R.G.K. for his ini-tiative in organizing this significant PNAS SpecialFeature on a critical aspect of human evolutionand for inviting me to contribute to it.

1. Gibbons A (2006) The First Human: The Race to Dis-cover Our Earliest Ancestors (Doubleday, New York).

2. Tattersall I (2003) Out of Africa again…and again? SciAm 13:38–45.

3. Wood BA, Collard M (1999) The human genus. Science284:65–71.

4. Schwartz JH, Tattersall I (2005) The Human FossilRecord, Vol 4: Craniodental Morphology of EarlyHominids (Genera Australopithecus, Paranthropus, Or-rorin), and Overview (Wiley-Liss, New York).

5. de Lumley H, et al. (2002) Datation par la methode40Ar/39Ar de la couche de cendres volcaniques(couche VI) de Dmanissi (Georgie) qui a livre desrestes d’hominides fossils de 1.81 Ma. C R Palevol1:181–189.

6. Swisher CC, III, et al. (1994) Age of the earliest knownhominids in Java, Indonesia. Science 263:1118–1121.

7. Clark JD, et al. (1984) Palaeoanthropological discover-ies in the Middle Awash Valley, Ethiopia. Nature307:423–428.

8. Holloway RL (2000) Brain. In Encyclopedia of HumanEvolution and Prehistory, eds Delson E, Tattersall I, VanCouvering J, Brooks A (Garland, New York), 2nd Ed.

9. Schwartz JH, Tattersall I (2005) The Human Fossil Record,Vol. 1: Terminology and Craniodental Morphology ofGenus Homo (Europe) (Wiley-Liss, New York).

10. Lieberman DE, McBratney BM, Krovitz G (2002) Sphe-noid shortening and the evolution of modern humancranial shape. Proc Natl Acad Sci USA 99:1134–1139.

11. Stringer CB, Hublin JJ, Vandermeersch B (1984) In TheOrigin of Modern Humans: A World Survey of the FossilEvidence, eds Spencer F, Smith F (Alan R. Liss, NewYork), pp 51–135.

12. Schwartz JH, Tattersall I (2000) The human chin revis-ited: What is it and who has it? J Hum Evol 38:367–409.

13. Sawyer GJ, Maley B (2005) Neanderthal reconstructed.Anat Rec 283B:23–31.

14. Rak Y (1990) On the differences between two pelvisesof Mousterian context from Qafzeh and Kebara Caves,Israel. Am J Phys Anthropol 81:323–332.

15. Schmid P (1983) Eine Rekonstruktion des Skelettes vonA.L. 288–1 (Hadar) und deren Konsequenzen. FoliaPrimatol 40:283–306.

16. Kibii JM, Clark RJ (2003) A reconstruction of the StW431 Australopithecus pelvis based on newly discoveredfragments. South African J Sci 99:225–226.

17. Walker A, Leakey REF (eds) (1993) The NariokotomeHomo erectus Skeleton. (Harvard University Press,Cambridge MA).

18. Arsuaga JL, et al. (1999) A complete human pelvisfrom the Middle Pleistocene of Spain. Nature399:255–258.

19. Simpson SW, et al. (2008) A female Homo erectus pelvisfrom Gona, Ethiopia. Science 322:1089–1092.

20. Rose MD (1984) A hominine hip bone, KNM-ER 3228,from East Lake Turkana. Am J Phys Anthropol 63:371–378.

21. Rosenberg KR, Zune L, Ruff CB (2006) Body size, bodyproportions, and encephalization in a Middle Pleisto-cene archaic human from northern China. Proc NatlAcad Sci USA 103:3552–3556.

22. Hauser MD, Chomsky M, Fitch WT (2002) The faculty oflanguage: What is it, who has it, and how did it evolve?Science 298:1569–1579.

23. McDougall I, Brown FH, Fleagle JG (2005) Stratigraphicplacement and age of modern humans from Kibish,Ethiopia. Nature 433:733–736.

24. White TD, et al. (2003) Pleistocene Homo sapiens fromMiddle Awash, Ethiopia. Nature 423:742–747.

16020 � www.pnas.org�cgi�doi�10.1073�pnas.0903207106 Tattersall

25. Brauer G (1984) In The Origins of Modern Humans: AWorld Survey of the Fossil Evidence, eds Smith F, Spen-cer F (Alan R. Liss, New York), pp 327–410.

26. Stringer CB, Andrews P (1988) Genetic and fossil evi-dence for the origin of modern humans. Science239:1263–1268.

27. Wolpoff MH, Wu X, Thorne AG (1984) In The Origins ofModern Humans: A World Survey of the Fossil Evidence,eds Smith F, Spencer F (Alan R. Liss, New York), pp 411–483.

28. Cann RL, Stoneking M, Wilson AC (1987) MitochondrialDNA and human evolution. Nature 325:31–36.

29. Stoneking M, Sherry ST, Redd AJ, Vigilant L (1993) Newapproaches to dating suggest a recent age for thehuman mtDNA ancestor. Phil Trans R Soc Lond Ser B337:167–175.

30. Nei M, Roychoudhury AK (1974) Genic variation withinand between the three major races of man, Caucasoids,Negroids and Mongoloids. Am J Hum Genet 26:421–443.

31. Tattersall I, Schwartz JH (2008) The morphological dis-tinctiveness of Homo sapiens and its recognition in thefossil record: Clarifying the problem. Evol Anthropol17:49–54.

32. Schwartz JH, Tattersall I (2005) The Human FossilRecord, Vol. 1: Craniodental Morphology of GenusHomo (Africa and Asia) (Wiley-Liss, New York).

33. Valladas H, et al. (1988) Thermoluminescence datingof Mousterian ‘‘Proto-Cro-Magnon’’ remains fromIsrael and the origin of modern man. Nature331:614 – 616.

34. Fagundes JRF, et al. (2007) Statistical evaluation ofalternative models of human evolution. Proc Natl AcadSci USA 104:17614–17619.

35. Vandermeersch B (1981) Les Hommes Fossiles deQafzeh (Israel) (CNRS, Paris).

36. Tattersall I (2004) What happened in the origin ofhuman consciousness? Anat Rec (New Anat) 267B:19–26.

37. McBrearty S, Brooks AS (2000) The revolution thatwasn’t: A new interpretation of the origin of modernhuman behavior. J Hum Evol 39:453–563.

38. Ambrose SH (2001) Paleolithic technology and humanevolution. Science 291:1748–1753.

39. Klein RG (1999) The Human Career (University of Chi-cago Press, Chicago), 2nd Ed.

40. Clark JD, et al. (2003) Stratigraphic, chronological andbehavioural contexts of Pleistocene Homo sapiensfrom Middle Awash, Ethiopia. Nature 423:747–752.

41. Marean CW, et al. (2008) Early human use of marineresources and pigment in South Africa in the MiddlePleistocene. Nature 449:905–909.

42. Stringer CB, et al. (2008) Neanderthal exploitation ofmarine mammals in Gibraltar. Proc Natl Acad Sci USA105:14319–14324.

43. Deacon H, Deacon J (1999) Human Beginnings in SouthAfrica: Uncovering the Secrets of the Stone Age (DavidPhilip, Cape Town).

44. Henshilwood C, et al. (2003) Emergence of modernhuman behavior: Middle Stone Age engravings fromSouth Africa. Science 295:1278–1280.

45. Henshilwood C, d’Errico F, Vanhaeren M, van NiekerkK, Jacobs Z (2004) Middle Stone Age shell beads fromSouth Africa. Science 304:404.

46. Bouzouggar A, et al. (2007) 82,000-year-old shell beadsfrom North Africa and implications for the origins ofmodern human behavior. Proc Natl Acad Sci USA104:1964–1969.

47. Vanhaeren M, et al. (2006) Middle Paleolithic shellbeads in Israel and Algeria. Science 312:1785–1788.

48. Tchernov E (1998) In Neandertals and Modern Humansin Western Asia, eds Akazawa T, Aoki K, Bar-Yosef Y(Plenum, New York), pp 77–90.

49. Clottes J (2008) Cave Art (Phaidon, New York).50. Conard NJ (2009) A female figurine from the basal

Aurignacian of Hohle Fels Cave in southwestern Ger-many. Nature 459:248–252.

51. Conard NJ, Malina M, Munzel SWC (2009) New flutesdocument the earliest musical tradition in southwest-ern Germany. Nature 460:737–740.

52. Shubin N (2008) Your Inner Fish (Pantheon, New York).53. Lieberman P (2007) The evolution of human speech: Its

anatomical and neural bases. Curr Anthropol 48:39–66.54. Coolidge FL, Wynn T (2005) Working memory, its exec-

utive functions, and the emergence of modern think-ing. Cambr Archaeol J 15:5–26.

55. Tattersall I (2008) An evolutionary framework for theacquisition of symbolic cognition by Homo sapiens.Comp Cogn Behav Revs 3:99–114.

56. Dunbar R (1996) Grooming, Gossip and the Origin ofLanguage (Faber and Faber, London).

Tattersall PNAS � September 22, 2009 � vol. 106 � no. 38 � 16021