Robert H. Gargett Middle Palaeolithic burial is not a dead ... · human origins, taphonomy, site formation Middle Palaeolithic burial is not a dead issue: the view from Qafzeh, Saint-Césaire,

Robert H. GargettArchaeology andPalaeoanthropology, School ofHuman and EnvironmentalStudies, University ofNew England, Armidale,N.S.W. 2351, Australia

Received 8 July 1997Revision received 7 January1999 and accepted17 February 1999

Keywords: MiddlePalaeolithic, Neandertal,ritual, mortuary,Mousterian,Chatelperronian,Castelperronien, modernhuman origins, taphonomy,site formation

Middle Palaeolithic burial is not a deadissue: the view from Qafzeh,Saint-Césaire, Kebara, Amud, andDederiyeh

Inferences of purposeful Middle Palaeolithic (MP) burial are almostuniversally accepted, despite published arguments that the pre-1960sdiscoveries are equally well explained by natural processes. In themodern human origins debate (perhaps the most hotly disputedquestion in palaeoanthropology) inferences of MP burial are crucialin arguments for an early Upper Pleistocene emergence of modernhumans. The present paper contributed to that debate byre-examining a number of post-1960s excavations of MP hominidremains. Because these were excavated with meticulous attention todepositional circumstances and stratigraphic context, most palaeo-anthropologists consider these inferences of purposeful burial to bebased on irrefutable evidence. This paper focuses on the reasoningbehind such claims, especially the assumption that articulated sketetalmaterial is prima facie evidence for deliberate burial. First it reviews arange of processes operating in caves and rockshelters that conditionthe probability of articulated skeletal material preserving withouthominid intervention. Processes such as deposition, decomposition,and disturbance are inherently more variable in caves and rock-shelters than is usually acknowledged. The first section concludesthat purposeful protection is not necessary to account for the preser-vation of articulated skeletal remains. The second part of the paperexamines the published record from Qafzeh, Saint-Cesaire, Kebara,Amud and Dederiyeh, where the majority of the remains claimed tohave been buried are fragmented, incomplete, and disarticulated.This re-examination suggests that in all of the post-1960s cases ofputative burial, the hominid remains occur in special depositionalcircumstances, which by themselves are sufficient to account for thepreservation in evidence at these sites. This conclusion severelyweakens arguments for purposeful burial at the five sites. More-over, the equivocal nature of the evidence in the more recent casesrenders even less secure the similar claims made for discoveries ofhominid skeletal remains at La Chapelle-aux-Saints, Le Mousterier,La Ferrassie, Teshik-Tash, La Grotte du Regourdou, Shanidar, andseveral others. Finally, by highlighting the equivocal nature of theevidence, this paper underscores the ongoing need for palaeoanthro-pologists to specify as wide a range of taphonomic processes aspossible when interpreting the archaeological record. This will aidin producing robust inferences, and will bring about increasinglyaccurate knowledge of when hominids became human.

? 1999 Academic Press

Journal of Human Evolution (1999) 37, 27–90Article No. jhev.1999.0301Available online at http://www.idealibrary.com on

0047–2484/99/070027+64$30.00/0 ? 1999 Academic Press

The report of my death was an exaggeration.

28 . .

(Mark Twain, cable from Europe to theAssociated Press: The Oxford Dictionary ofQuotations, 1979)

Occasionally, to the delight of paleon-tologists, the entire skeleton of a long-deadanimal is preserved, with every bone presentand in place. Such an event occurs only inspecial circumstances: the body of the ani-mal must have come to rest in a place whereit could lie undisturbed for a great manyyears; it must have been rapidly coveredwith a suitable enclosing matrix, and it musthave been preserved and strengthenedby percolating solutions of appropriatechemical composition. More usually, theanimal’s body is subjected to the destructiveinfluences that characterize any naturalenvironment. The skeleton becomesdisarticulated—broken down into its individ-ual parts, each of which has then to contendwith the attention of carnivores and with theforces of decay and disintegration.

(Brain, 1981:11)

Introduction

This paper examines recent claims for pur-poseful Middle Palaeolithic (MP) burial, tosee if natural processes can be ruled out asan explanation for preservation of the homi-nid remains. The first part considers a rangeof taphonomic processes that determinewhether or not a hominid’s remains will bepreserved, and in what condition. Thesecond part looks at published excavationreports from Qafzeh, Saint-Césaire, Kebara,Amud, and Dederiyeh. When the publishedobservations are examined from a tapho-nomic perspective, it becomes clear thatthere are perfectly plausible, indeed satisfy-ing, natural explanations for the hominidremains.

Background

Palaeoanthropologists are deeply divided onthe question of when hominids becamehuman (see e.g., Binford, 1985; Wolpoff

et al., 1988; Marshack, 1989; Mellars, 1989,1995; Lindly & Clark, 1990; McBrearty,1990; Dettwyler, 1991; Duff et al., 1992;Stiner & Kuhn, 1992; Gamble, 1993;Hayden, 1993; Schepartz, 1993; Stringer &Gamble, 1993; Trinkaus & Shipman, 1993;Wynn, 1993; Klein, 1994; Lieberman &Shea, 1994; Mithen, 1995, 1996; Willermet& Clark, 1995; Noble & Davidson, 1996).There is much disagreement about how tointerpret the archaeological record forwhat it can tell us about the cognitive abili-ties of MP hominids (i.e., both the archaicand the modern morphospecies thatmade the same kind of stone artefacts—Mousterian with or without the Levalloistechnique). In this debate inferencesof purposeful MP burial strongly supportarguments for an early emergence ofhumanness (i.e., before about 40 to 60 kaago), because most palaeoanthropologistsbelieve that hominids buried their deadas early as 100 ka ago, based on theinferences of purposeful burial at QafzehCave (Vandermeersch, 1981; Valladas et al.,1988).

Because it is widely accepted as goodevidence for humanness, purposeful burialhas become a crucial datum in the modernhuman origins debate. There has been muchdiscussion of the behavioural and cognitiveimplications (see e.g., Peyrony, 1921;Bouyssonie, 1954; Bergounioux, 1958;Binford, 1968; Vandermeersch, 1976;Harrold, 1980; Shackley, 1980; Chase &Dibble, 1987; Bar-Yosef, 1988; Gargett,1989a; Noble & Davidson, 1989, 1991,1993; Smirnov, 1989; Tillier, 1990;Belfer-Cohen & Hovers, 1992; Defleur,1993). Obviously the timing and emergenceof purposeful burial are important to thedebate. Yet despite the importance that

29

purposeful MP burial assumes, there hasbeen little critical scrutiny of the evidencefor it.

Claims of purposeful burial should becapable of withstanding examination,informed by knowledge of formation pro-cesses and taphonomy. After all, relatedquestions have spawned intense argument,and launched scores of archaeological,ethnoarchaeological, and ethologicalstudies—e.g., whether or not hominidshunted prior to about 40 ka ago. In contrastthere has been very little interest in thetaphonomy of hominid burial. Since theearly twentieth century, when archaeologistsbegan explicitly to argue that Neandertalremains had been purposefully buried (e.g.,Bouyssonie et al., 1908), there has beennearly universal acceptance that MP homi-nids buried their dead. Only in the lasttwenty years have doubts been raised aboutsome of the best-known claims, such asthe Shanidar flower burial and the Teshik-Tash circle of horns (e.g., Binford, 1981;Davidson & Noble, 1989; Gargett, 1989a;Noble & Davidson, 1989). However, before1989 there had been no strenuous criticism,and no one suggested that, taken together,the evidence for purposeful MP burial wasequivocal.

In 1989 I examined published claims forpurposeful burial at La Chapelle-aux-Saints(Bouyssonie et al., 1908, 1913; Boule,1909), Le Moustier (Peyrony, 1930), LaFerrassie (Capitan & Peyrony, 1909, 1910,1911, 1912a,b, 1921; Peyrony, 1934;Heim, 1968), Teshik-Tash (Okladnikov,1949; Movius, 1953), La Grotte duRégourdou (Bonifay, 1962, 1964; Bonifay& Vandermeersch, 1962), and Shanidar(Solecki, 1955, 1960, 1961, 1963, 1971;Solecki & Leroi-Gourhan, 1961; Leroi-Gourhan, 1975). I argued that their preser-vation was equally well explained by naturaldeposition (Gargett, 1989a,b). Further-more, I suggested that similar doubt couldbe cast on claims for purposeful burial at

Roc de Marsal (Bordes & Lafille, 1962),Kiik-Koba (Bonch-Osmolovskij, 1940;Klein, 1966), La Quina (Martin, 1923),Amud (Suzuki & Takai, 1970) and Tabun(Garrod & Bate, 1937). To judge by thecontent of university textbooks, most schol-arly works, and the popular press in the last10 years, my conclusions have been widelyignored. Seven principle criticisms of myearlier paper bear reviewing here, becausethey are the basis on which the work hasbeen judged. The present paper is in partaimed at addressing some of these perceivedshortcomings.

First, Gargett (1989a) dealt only withputative burials recovered in the early part ofthis century, which were not well excavatedor recorded. Some complained that this wasunscientific (e.g., Frayer & Montet-White,1989; Trinkaus, 1989). Yet, if shortcomingsare evident in the most often cited examples,constructive criticism and plausible alterna-tive explanations ought to be consideredworthy of publication.

Second, I chose to critique only claimsfor Neandertal burial, and not those of skel-etally modern MP hominids. I now addressthat omission, by examining a number ofmore recent Neandertal discoveries, and bybroadening the scope of my examination toinclude the discoveries from Qafzeh.

Third, some scholars question my auth-ority, and argue that the excavators and theirconclusions are above reproach (Bricker,1989; Hayden, 1993). It may be true thatsome archaeologists are above reproach.Nevertheless, archaeological inferencesshould be able to stand on their own, andnot depend purely on reputation.

Fourth, a number of commentators dis-agreed with the alternative explanations Iproposed, or asserted that I had misinter-preted the literature. My replies are a matterof record (Gargett, 1989a:184–188,b:326–329).

Fifth, I did not explain the absence ofarticulated remains before about 100 ka ago

30 . .

1Vermeersch et al. (1998) describe an articulatedhominid in the open air, in what is argued to be an MPopen-pit chert mine. However its depositional circum-stances are just what I would expect if natural sedi-mentary processes were to have been responsible forpreserving a hominid intact in an open-air site. Theremains are buried in what is argued to be an undis-turbed MP ‘‘extraction dump,’’ the undisturbed natureof which is inferred on the basis of the presumedstratigraphic integrity of what the authors claim are midand mid- to late-MP stone artefacts. A 6000- to 9500-year-old aridsol has developed on the surface at the site,which the excavators use to infer a Pleistocene age forthe remains. I leave it up to the reader to scrutinizeVermeersch et al.’s arguments in light of the criteria Ibring to bear in this paper.

(Gilman, 1989). As I discuss later, thismay simply imply that, in earlier times,hominids spent little time in caves and rock-shelters, and were concomitantly less likelyto have been preserved (cf. Binford & Ho,1985).

Sixth, many noted that I ignored UpperPalaeolithic and later burials, many of whichwould not pass my stringent criteria (e.g.,Belfer-Cohen & Hovers, 1992). If, in theMP, one is dealing with the earliest evidencefor purposeful burial, even a cursory glanceat the Upper Palaeolithic evidence revealsdifferences that obviate the need for a com-parison between the two. In the UpperPalaeolithic, for example, one often seessomething more than might be expected tooccur naturally (cf. Stringer & Gamble,1993:160; Noble & Davidson, 1996:208–209). Moreover, with the exception of therecent claims from Taramsa Hill, Egypt(Vermeersch et al., 1998) only fully modern,Upper Palaeolithic humans have beenfound articulated in open-air sites—e.g.,Lake Mungo, Sunghir, Dolní Vestonice(Stringer & Gamble, 1993:160), or withunequivocal evidence for mortuary treat-ment, in the form of extreme flexion orcomplete extension, or cremated, or withcopious, unmistakable grave goods, andso on.1

Seventh, and most important, more cre-dence is given to claims of MP burial madein the last quarter-century (e.g., Farizy &

Masset, 1989; Gamble, 1989; Mellars,1989; Trinkaus, 1989; Bar-Yosef &Meignen, 1992; Garrard, 1992; Roe, 1992;Rosenberg, 1992; Schepartz, 1992; Smith,1992). Indeed, there is a widespreadconviction that the recently excavated MPhominids from Qafzeh (Vandermeersch,1981), Kebara (Arensburg et al., 1985;Bar-Yosef et al., 1986, 1988, 1992), Saint-Césaire (Lévêque et al., 1993; Lévêque& Vandermeersch, 1980), Dederiyeh(Akazawa et al., 1993; Akazawa et al.,1995a,b), and Amud (Rak et al., 1994;Hovers et al., 1995) are indisputable evi-dence for purposeful burial. This belief hasperpetuated acceptance of any and all claimsfor MP burial, and nurtured claims for pur-poseful disposal of the dead even earlier, inthe Middle Pleistocene. For example, recentspectacular discoveries of early archaicHomo from the Sima de los Huesos inAtapuerca, northern Spain, have bredspeculation that the disarticulated remainswere purposefully disposed of (see e.g.,Bahn, 1996; Gore, 1996). However, thepresence of nonhominid remains in thesame natural trap undermines any appeal toarcane mortuary ritual as an explanation ofthe hominid remains. That is, unless onewere to argue that the animals, too, wereplaced in the cave by the archaic hominids.Clearly the widespread acceptance of claimsfor purposeful MP burial is encouragingsuch fanciful behavioural inferences.

Because of the faith placed in the morerecent discoveries as irrefutable evidence ofpurposeful MP burial, there is an urgentneed to examine them for what their depo-sitional circumstances can reveal. I hope toshow that in spite of the superior care withwhich they were recovered, the more recentclaims still suffer from the same inferentialshortcomings as those from La Chapelle-aux-Saints, Le Moustier, La Ferrassie,Teshik-Tash, Régourdou, and Shanidar(see Gargett, 1989a,b). Ironically, if myinferences about the recent discoveries are

31

more robust, it will be because of the qualityof the data that resulted from the meticulousexcavations.

The taphonomy of hominid burial

The primary critical stance I assume in thiswork is that natural processes must be ruledout before invoking human or hominidbehaviour to explain archaeological occur-rences. If purposeful burial were indeed theonly means by which articulated hominidskeletal remains could be preserved, therewould be no need to argue that it hadoccurred in the MP, and there would be nopoint in questioning the inference. But, asthe epigraph from Brain (1981) reminds us,purposeful burial is not the only way thatarticulated skeletal remains can be pre-served. Indeed, processes that have nothingto do with hominid behaviour have alwaysbeen at work creating a fossil record thatincludes articulated skeletons. Specialdepositional circumstances are all that isneeded—i.e., natural processes that com-bine to preserve the skeletal elements intheir anatomical relationships, and (somewould say) to mimic purposeful burial.Thus, because natural processes are capableof preserving articulated skeletons, naturalprocesses need to be ruled out before humanor hominid agency is inferred. To accom-plish this, archaeologists need adequatemethodological tools. In this case they needknowledge of the range and variability ofsite-forming processes to enable them todistinguish between what is likely to benatural and what could only be produced byhominid behaviour.

Explaining the archaeological occurrenceof articulated hominid remains should be nodifferent from that of interpreting similarlypreserved animal bones. Zooarchaeologistshave given considerable thought to the epis-temology of interpreting animal bones fromarchaeological sites (see e.g., Binford, 1981;Gifford, 1981; Gifford-Gonzalez, 1989,

1991; Haynes, 1990). The consensus seemsto be that stronger inferences result whenmultiple, convergent lines of evidence areemployed, and when present-day processesare the source for strong analogies.Strong—i.e., relational—analogies enablemore accurate modelling of past processes(cf. Wylie, 1989; Gifford-Gonzalez, 1991).Relational analogies are those which dependon recognizing relevant causal relationshipsbetween observed formal attributes and theprocesses that created them. Therefore, inconstructing robust inferences from the evi-dence of animal bones, archaeologists needto take into account the presently observablerange, and the variability inherent in pro-cesses that can affect animal skeletons,both at the time of death, and after—i.e., thebiostratinomic and diagenetic processes(Gifford, 1981).

This part of the paper presents an over-view of the site-formation processes thatcan play a part in preserving hominid skel-etal remains. I propose that claims for pur-poseful MP burial rest ultimately on thepresence of articulated skeletal remains,which is assumed to preclude natural burial.This is unwarranted. Five questions andtheir implications form the core of myargument.( 1) What constitutes evidence of purpose-

ful protection of the corpse?( 2) What is the probability of natural

burial in caves and rockshelters?( 3) What is the prior probability of preser-

vation under any circumstances?( 4) What is the importance of articulation?( 5) What is the variability in decompo-

sition rates, disarticulation sequencesand likelihood of disturbance?

Before turning to these five issues, I wishto dispense with two lines of evidencethat are often employed by proponentsof MP burial, but which cannot supportclaims for purposeful burial: position ofthe corpse and so-called grave goods orofferings.

32 . .

Position of the corpse. As one of manycategories of mortuary ‘‘style,’’ diachronicchanges in body position have been usefulin making inferences of modern humancultural change. Position of the corpsealso figures prominently in the argumentsfor Neandertal burial. Bouyssonie et al.(1913:630), for example, proposed it as oneof five criteria for inferring purposeful burialof the La Chapelle-aux-Saints hominid, say-ing that the individual was flexed, ‘‘as if insleep’’ (my translation). Since that earlytwentieth-century discovery, a number ofexcavators have employed the position of thecorpse as primary evidence for burial. Amud1 is another good example of this (Suzuki &Takai, 1970). Archaeologists have usedbody position as a means of inferring MP‘‘culture’’ change and geographic ‘‘cultural’’variation, in studies modelled on modernhuman mortuary analyses (but see alsoBelfer-Cohen & Hovers, 1992; and mostnotably Smirnov, 1989; Defleur, 1993). Inthese cases the a priori assumption is thatpurposeful burial had occurred. Analyses ofMP corpse position, therefore, depend onthe presumption that purposeful burial hasoccurred, and thus that the position of theskeleton is the result of conscious choices onthe part of MP hominids. On the model oftightly bound human corpses readied forburial in very small places, position meanssomething. Present-day humans make con-scious choices about how they are going to‘‘lay out’’ a corpse. However, if one isindeed interested in evidence useful fordetermining if purposeful burial hasoccurred, it makes little sense to employcategories of evidence that assume that ithas occurred. There are several reasons whythe position of the corpse is an unhelpfulcriterion for recognizing purposeful MPburial.

First, its use implies that mortuary treat-ment has occurred, even though it is beingused to argue that purposeful treatment hasoccurred. This is a circular argument.

Second, use of body position has lackeddescriptive precision. MP skeletons’ pos-itions are always described using some vari-ant of ‘‘flexed,’’ with no attention paid to theimprecision the term carries with it. Whenone looks at the body positions of claimedMP burials one is struck by the varietyencapsulated by the term ‘‘flexed,’’ from thealmost extended Qafzeh 9 (Vandermeersch,1981) to the nearly ‘‘foetal’’ position ofAmud 1 (Suzuki & Takai, 1970). Thus, assupport for an argument in favour of pur-poseful MP burial, the criterion of bodyposition is virtually meaningless, becauseany position a hominid corpse could assumeon its own could be described as ‘‘flexed.’’Only in purposeful burials of modernhumans does one see skeletons ‘‘fullyextended’’ (in the archaeological sense oflimb elements parallel to the long axis of thebody) and ‘‘fully flexed’’ skeletons (i.e., withthe heels a few centimetres from ischium,knees nearly touching ribs, and carpalswithin a few centimetres of proximalhumerus). Such positions have never beendescribed in MP contexts. The term‘‘flexed’’ thus masks a great deal of varia-bility, variability that is important to answer-ing the question of whether or not MPhominids buried their dead.

Third, using body position in argumentsfor MP burial presumes that a flexed corpsecould not occur naturally. Many workershave claimed that the degree of flexion ofsome of the putative MP burials precludesnatural death and entombment (e.g., Villa,1989:325), or that rigor mortis and post-mortem bloating would in all likelihood havecaused even a very flexed, unburied corpseto extend. However, there is nothingunnatural about the positions of MP homi-nids with very flexed lower limbs. Newell(1984), for example, describes the exca-vation of a pre-European-contact Inuithouse whose occupants were asleep whenpack ice over-rode the beach berm behindwhich their house stood, killing and

33

entombing them. These individuals hadobviously not been prepared for burial, butwere nevertheless very flexed—every bit asflexed as Amud 1 (Suzuki & Takai, 1970).To preserve in the death position onlyrequires the right environmental conditions.

Grave offerings. Archaeological materials‘‘associated’’ with claimed MP burials areoften called grave offerings or grave furni-ture (cf. Binford, 1968; Chase & Dibble,1987; Noble & Davidson, 1989). Oftenthese are scraps of animal bone, or stonetools, or bedrock clasts. The Law of Associ-ation (Worsaae, 1843), which developed outof mortuary archaeology, refers only toobjects that are indisputably buried with thecorpse at the time of interment. In noclaimed MP burials has there been anydemonstration of the existence of suchassociation. Close proximity is not, in anysense, association. For example, the infer-ence of flowers buried with the Shanidar 4individual (Solecki & Leroi-Gourhan, 1961;Leroi-Gourhan, 1975) suffers from thismethodological shortcoming. Sedimentsamples were taken, almost as an after-thought, ‘‘from about the same level on whichthe skeleton lay’’ (Solecki, 1971:247) (myemphasis), without any effort to demon-strate that the samples were part of a‘‘burial’’ context. Most MP cave and rock-shelter sites are places where there are plentyof stone artefacts and animal bones. Thus,one has to argue nimbly, and ultimatelydistort the intent of the Law of Association,to propose that the proximity of hominidskeletons to other archaeological sedimentsstands as evidence of anything other thantheir presumptive contemporaneity (andthere is no guarantee that this would be thecase). Until such time as MP burial is dem-onstrated, and not argued, an object inproximity to a hominid skeleton in theabsence of stratigraphic evidence for a gravecannot stand as evidence of purposefulburial. As long as there are no clearlydefined MP burial strata (a point to which I

turn in a moment), there can technically andpractically be no such thing as ‘‘associated’’grave goods.

I now return to the five questions posedearlier, with the aim of developing somecriteria that can be useful when examiningthe depositional circumstances of MP homi-nid remains. The first of these is protectionof the corpse.

What constitutes evidence of purposefulprotection of the corpse?Purposeful protection of the corpse is alwaysinferred where putative MP burials are con-cerned, but this category of evidence has yetto be observed in an MP context. In a strictstratigraphic sense, the most secure evidencefor purposeful burial is the artificial stratumcontaining the remains, created at the timeof interment (Drucker, 1972:5; Harris,1979:95). To be distinguished from natu-rally accumulating sediments, and thus tostand as evidence for burial, the new stratummust be distinct from those upon which orwithin which it occurs, and from those accu-mulating above it and the original surface. Itis not enough, for example, to recognize adepression or a low spot within which theremains occur, and which might be inter-preted as a ‘‘burial pit.’’ Unless a new stra-tum can be distinguished, there is no logicalway to argue that the remains were purpose-fully protected. The reason for this is simple:if the overlying sediments are part of a moreextensive deposit that includes the ‘‘fill’’ ofthe ‘‘pit,’’ this greatly weakens the argumentthat the overlying sediments were the resultof purposeful burial. It goes without sayingthat the new stratum must not be an artefactof natural erosion that has simply removedthe upper portion of a more extensive,continuous stratum containing the remains.

A new stratum is the key to discerningunequivocally that purposeful burial hasoccurred. In the absence of unequivocalsupport, other lines of evidence must beemployed, which need to be grounded in an

34 . .

understanding of the range of processes thatcan affect preservation.

Where the question of MP burial isconcerned, failure to distinguish a newstratum—and thus to demonstrateunequivocally that purposeful burial hasoccurred—leaves the inference open to chal-lenge, and to the need for alternatives toexplain the deposition of hominid remains.

For example, at La Chapelle-aux-Saintsthe excavators identified a depression con-taining Neandertal remains. Yet the pub-lished profile does not distinguish betweenthe sediments that accumulated in thedepression from those that accumulatedabove it (Bouyssonie et al., 1908:516).Thus, one is left to conclude that the samesedimentary process covered the corpse andfilled the entire cave to a depth of 0·5 m.One cannot rule out the possibility that theNeandertal was placed in an oversized pitand covered by such an extensive new stra-tum. However, it does seem rather unlikely.In any case, one cannot rule out the possi-bility that the skeleton was naturally buried,without relying on a number of auxiliaryhypotheses, as I now explain.

In arguing that the La Chapelle-aux-SaintsNeandertal was buried, the excavators reliedheavily on the inference that the depressionwas purposefully created, and on the pres-ence of an articulated skeleton. Otherwise,the argument goes, it could not have pre-served in the way it did. Later I examine theassumption that disturbance would inevi-tably have occurred, and question the war-rant for the assumption that an articulatedskeleton is evidence of purposeful burial. Fornow I simply suggest that the presence of apit is not evidence of purposeful burial.Whether the depression was purposefullycreated or not, there is no evidence that itwas filled at the time the Neandertal remainswere deposited. Thus, the inference of pur-poseful burial at La Chapelle-aux-Saintsrelies on creaky archaeological inferencebased on weakly warranted assumptions.

Having suggested the stringent require-ment of recognizing a new stratum to inferburial unequivocally, I would hasten to addthat, archaeological reality being what it is,discoveries of skeletal remains will notalways be in sediments that lend themselveseasily to distinguishing a new stratum.Indeed, even in what are clearly modernhuman cemeteries, burial strata as I havedefined them are not always evident (a pointmade by many commentators on my 1989apaper). However, the need to observe a newstratum when arguing that purposeful burialhas occurred cannot be overemphasized.Without it, to support a claim for MP burial(or to be critical of one) archaeologistsmust rely on models and probabilisticinferences based on their understanding ofthe processes that condition preservation.Accurately inferring the taphonomic historyof a hominid skeleton will depend on thequality of the models that are brought tobear. In the case of MP burial, makingrobust inferences depends on specifying asbroad a range as possible of the processescapable of producing the archaeologicallyrecovered materials, while avoiding relianceon simplistic models of site formation.

Developing a nuanced understanding ofnatural formation processes will involveconsidering multiple, independent lines ofevidence, which can produce a mutuallyconstraining and mutually supportive evi-dentiary basis for inferences (see e.g., Wylie,1989; Gifford-Gonzalez, 1991). With theobject of building up a battery of modelsof site formation and taphonomy, the nextsection begins with a brief overview ofthe depositional environment in which MPhominid remains are most commonlyfound—i.e., calcareous caves and rockshel-ters. The depositional environment in cavesand rockshelters determines the probabilitythat a hominid corpse will preserve natu-rally. This is such a crucial issue that Ipresented the relevant aspects in my earlierexamination of the evidence for Neandertal

35

burial (Gargett, 1989a). More comprehen-sive treatments of karst geomorphology andcave sedimentology may be found in Ford(1976); Jennings (1985); Dreybrodt (1988)and White (1988) (see also Courty et al.,1989).

What is the probability of natural burial incaves and rockshelters?Archaeological deposits in caves and rock-shelters contain more kinds and sizes ofsediment than any other depositionalenvironment (Jennings, 1985:163). Whenthey develop in calcareous bedrock, cavesare also (usually) places where bone pre-serves better than elsewhere, for a varietyof reasons, including sedimentation rate,mode of deposition, pH, and microclimate.All things being unequal, therefore, bonesin caves and rockshelters enjoy a greaterchance of survival than those lying out in theopen. This is why caves and rockshelters arethe pre-eminent European Palaeolithicarchaeological sites.

Investigations of MP burial mustacknowledge that each cave’s and rock-shelter’s depositional history is complex andunique. Even the idea that one coulddevelop a set of criteria by which to evaluatethe inference of purposeful burial is ren-dered suspect by the complexity of interpret-ing such sites. For this reason alone it isdifficult to be critical of the inferences thatemerge. To prime the reader for the criti-cism I make later, I briefly introduce themajor characteristics of sedimentation incaves and rockshelters (for further elabor-ation refer to Gargett, 1989a:158–161).

Autochthonous sediments. The bedrockwithin which caves and rockshelters developis often a prime source of sediments. Incalcareous bedrock, dissolution of carbon-ates can result in gradual accumulation ofwhatever insoluble material is left behind,usually sediments in the clay size range.Freeze-thaw and wet-dry cycles near theopen air cause angular and subangular

blocks of all sizes to break away from thebedrock on the walls and ceiling. Inaddition, episodic, sudden bedrock break-down contributes sediments that can rangefrom dust- to house-sized, which are likelyto be poorly sorted to unsorted, and may ormay not be clast supported. Many caves androckshelters that contained putative burialswere more susceptible to sudden bedrockbreakdown because they occurred intectonically active areas (e.g., Shanidar,Qafzeh). One could imagine that a rockfallresponsible for killing and preserving ahominid need not be voluminous, norinclude very large fragments. Therefore,before other processes are invoked, suddenbedrock breakdown must be ruled out asan explanation for the presence of some ofthe claimed MP burials. Unfortunately, dis-tinguishing the sediments of a catastrophicbedrock collapse from those of more gradualprocesses, such as exfoliation, has provenexceedingly difficult, because of thevariability inherent in breakdown products(Farrand, 1985:28).

Solecki (1960:613), for example, makesmuch of what he observed as ‘‘a cluster’’ of‘‘smaller stones’’ ‘‘superimposed upon alayer’’ of ‘‘larger stones’’ lying immediatelysuperior to the Shanidar I remains. He infersthat

‘‘survivors of the rockfall returned . . . and,seeing what had happened, heaped someloose stones, the closest at hand, over theunfortunate’s remains.’’

Farrand (1985:26, Table 1) suggests a wayin which previously deposited sedimentsmight be altered by catastrophic collapseand thereby act as witness to such an event:‘‘Crushed Debris (of all kinds)’’. Thus theinterpretation of bedrock collapse restsalmost completely on interpretations of whatconstitutes crushed debris. Following fromFarrand’s suggestion, it would seem thatcrushed hominid skeletal material would beone potential indicator of rockfall, as would

36 . .

broken stone artefacts. As it happens,both categories of evidence occur atSaint-Césaire, about which I comment laterin this paper.

Allochthonous sediments. In addition to thepotential for a great variety of autoch-thonous sediments, caves and rocksheltersinvite occupation by many kinds of animals,some of which (including hominids) canintroduce all manner of allochthonous sedi-ments, among them bones. Fuel for fires orfor bedding, and faeces are also commonconstituents of such sites, as are stone arte-facts in hominid habitations. A number ofsouthwestern Asian sites contain largequantities of ash.

Depositional environment. In contrast withmost depositional environments, for longperiods caves and rockshelters can experi-ence little erosion of their deposits, thusincreasing the net rate of deposition. How-ever, even though protected from somekinds of erosion, reactivation of karstic sub-terranean stream activity can cause flushingout. Moreover, sinkholes can form in pre-viously stable sediments creating, effectively,a drain through which material passes tolower parts of the cave system. Thus,although the rate of deposition in caves androckshelters can be on average more rapidthan in most subaerial contexts, makingthem, effectively, ‘‘sediment traps,’’ the netrate of deposition is unpredictable. And,since the deposition rate is almost never aconstant, there is no theoretical warrant forestimating annual rates based on overallaccumulation, even when there is a chrono-metric estimate of the length of time adeposit took to form. For example, a 1-m-deep deposit that takes 10 ka to accumulatebuilt up at a theoretical average rate of 1 mmevery 10 years. However, it might be thecase that 90% of the deposit accumulated inthe first ten years—i.e., a real rate of 90 mma year. In this case the theoretical rate is atvariance with the actual rate. There is alwaysthe possibility that within an apparently

slow-to-accumulate deposit, relatively rapidsedimentation will occur, regardless ofwhether the main source of sediments isautochthonous or allochthonous.

Deposition in caves and rockshelters alsovaries according to many characteristics ofthe bedrock and local hydrology. In particu-lar, the timing and periodicity of bedrockbreakdown events would vary according tobedding thickness and dip. For example,caves developed in thinly bedded, steeplydipping sedimentary rock will tend to breakdown at a much higher rate, and in muchsmaller individual clasts, than those in mas-sive, horizontally bedded, sedimentary rock.Thus the rate of deposition can be faster orslower, depending on the unique conditionsoperating in the site, and can change inresponse to alterations in those conditionsresulting from the site’s morphological lifehistory, or from climatic or other environ-mental change. As an example of the hugevariability in rates of cave deposition, andthe impossibility of predicting such charac-teristics, two roughly contemporary caves indifferent parts of the world, La Chapelle-aux-Saints and Shanidar, accumulated sedi-ments at different rates: just over a metre hasbuilt up at La Chapelle since the Neandertalwas deposited, while approximately 15 mhas accumulated in what is assumed to bemore or less the same period at ShanidarCave. The difference is due entirely to dif-ferences in the bedrock and geomorphicenvironment, resulting in vastly differentdepositional histories. At Shanidar Cave,for example, bedrock from the ceiling hascollapsed repeatedly over the millennia,probably encouraged by active tectonism,creating an enormous vault above the exca-vated area from which all of the hominidmaterial was recovered.

Microdepositional variability. Added to theinherent variety of sediments and depo-sitional rates in caves and rockshelters is thestrong likelihood that different microdepo-sitional environments will exist within a

37

given site, creating locations with differentkinds and rates of deposition and differentbone preservation potential. The archaeolo-gist faced with interpreting the likelihoodof hominid skeletons preserving in suchcircumstances needs to be aware of thepossibilities.

One example of a microdepositionalenvironment with an above-average poten-tial to preserve skeletal material is that ofthe low spot (including closed depressions).These come in two types: (1) erosionalunconformities (after Harris, 1979), wheresediments are deleted in a confined space bya variety of processes (including, but notrestricted to, purposeful excavation—seebelow), and (2) topographic low spotsthat are simply the fortuitous product ofdepositional circumstances. Crucial to thisdiscussion of MP burial is the well-knownsedimentological phenomenon that it takesmore energy to transport a sediment out of adepression or a low spot than it took to get itthere. However they are formed, low spotsare ‘‘sediment traps’’—they will accumulatesediments more rapidly than surroundingsurfaces. In addition, any vertebrate carcassthat came to rest in a depression or a lowspot in a cave or rockshelter would stand agreater chance of preservation, because itwould be more rapidly buried than similarmaterial in a more easily eroded environ-ment, and thus would be more rapidly pro-tected from trampling and other destructiveprocesses.

In light of the depositional characteristicsof low spots, and the likelihood that closeddepressions containing hominid remains willbe interpreted as burial pits, it is importantto understand that closed depressions andother low spots can form naturally in avariety of ways. In calcareous sediments,for example, solution cavities can form inbedrock, which I have suggested as an expla-nation for the ‘‘graves’’ at Kiik-Koba, andfor the two depressions reported at LaChapelle-aux-Saints, only one of which

contained hominid remains (Bouyssonie,1954:108) (in addition to inferring that thelarge depression at La Chapelle-aux-Saintswas deliberately excavated to bury ahominid, the contents of a much smallerdepression near the entrance to the cave,presumably formed in the same way, wasinterpreted as a ritual feature). In a lime-stone cave, water dripping from the ceilingcan form metre-sized circular depressions inunconsolidated floor sediments (personalobservation). Inside caves and rockshelters,developing talus cones can create low spotswhere they adjoin walls, which is apparent atRégourdou (in a similar vein, I have sug-gested that La Ferrassie 1 and 2 may havebeen rapidly buried because they came torest on a near-wall surface that was slopingdownward towards the wall). When watercommunicates between the upper and lowerportions of a cave that has been verticallybisected by talus, it can produce hourglass-like erosional cavities in the unconsolidatedsediments lying against bedrock walls, whichcan be closed off by accumulating sedimentsand begin to act as sediment traps (alsoapparent at Régourdou). Resting animalscan produce hollows in unconsolidatedsediments where they habitually lie (e.g.,kangaroos and dogs), and some animals areknown to excavate such places intentionally(e.g., cave bears). Most importantly, mod-ern human foragers are known to prepareshallow basins in unconsolidated sediments,when they wish to be comfortable, or tokeep warm when sleeping on the ground(e.g., Gould, 1977:41; Hayden, 1979:173).Regardless of how they formed, closeddepressions might easily be interpreted asburial features if they were to contain homi-nid remains. A small number of MP homi-nid remains have been recovered either atthe contact with, or in sediments immedi-ately superior to, closed depressions. Theseare usually inferred to represent burial orritual contexts. However, none of these rare‘‘pits’’ has been found filled with a new

38 . .

stratum, thus precluding unequivocalinterpretation of such features as burial pits.

In general, the location of skeletal remainsin caves and rockshelters can be assumed tohave a major determining role in bone pres-ervation, although the data to support thiscontention have been rather unsystemati-cally collected, and are still, unfortunately,sparse. Many cavers, cave archaeologists,and palaeontologists have alluded to theways in which depositional circumstancesand ethology can combine to contribute topreservation (e.g., Koby, 1941, 1953; Ford,1976; Kurtén, 1976; Sutcliffe et al., 1976;Courty et al., 1989; Kerbis-Peterhans, 1990;Lam, 1992; Theunissen et al., 1998). Myown analysis of behaviourally meaningfulspatial patterning in a cave bear faunafrom Pod Hradem Cave, Czech Republic(Gargett, 1996), is the first systematic studyof spatial patterning in the traces left byorganisms other than modern humans.From such accounts, the following expec-tations can be derived. Against the wall, oramong boulders, bones will tend to collect,because whether they are consciously orunconsciously moved around, they are lesslikely to be moved subsequent to their com-ing to rest in out-of-the-way places. Inaddition, once against the wall or in a natu-ral niche, they are protected from tramplingand comminution. Thus, delicate bones, aswell as articulations between bones, willtend to survive better if they are protectedfrom disturbance by fortuitous location inout-of-the-way places, or if they are buriedby rockfall. Although at present there are nodata to support the proposition, the inherentcomplexity and microdepositional variabilityin caves and rockshelters might lead to dif-ferential preservation of different bones ofan individual skeleton, should the carcassspan two areas with different depositionalrates and thus different preservationpotential—e.g., the boundary between atalus slope and the floor; the boundarybetween the floor and a depression in the

floor; a discontinuous or uneven roof col-lapse, which would leave parts of a skeletonnearer the new surface, and thus more sus-ceptible to trampling or other disturbance.Moreover, subsequent erosion of the sedi-ments that encase previously buried skeletalmaterial will most likely lead to differentialdegradation of the exposed portions. All ofthese hypotheses are empirically investi-gable. This is a neglected area of researchand is in need of much attention. In additionto the mechanical determinants of preser-vation, one could add the differential effectsof sediment chemistry and diagenesis, whichalso act variably to degrade or preserve bone(Courty et al., 1989; Weiner et al., 1993,1995; Schiegl et al., 1996).

In sum, caves and rockshelters are inher-ently variable, inherently complex depo-sitional environments, which obviate the useof simplistic models of site formation andbone preservation.

What is the prior probability of preservationunder any circumstances?In the knowledge that caves and rocksheltersare capable of preserving bone better than inthe open air, and that the microdepositionalvariability in caves could lead to differentialpreservation of bone, one can begin to seehow articulated skeletal remains might attimes preserve naturally. But how frequentlydo articulated remains preserve? How likelyis it? Complete, articulated hominid skel-etons are extremely rare. Most of the severalhundred documented MP hominid individ-uals are known only from fragments of skel-etal elements. Often these are the moredurable parts of the skeleton, such as thecranium, mandible and dentition. Thus, theoverwhelming majority of the hominids thathave contributed parts of their skeletons tothe MP fossil record were obviously notspared the ignominy of destruction of largeportions of their skeleton. Some, like theTeshik-Tash boy, were gnawed by carni-vores. Moreover, most of the relatively few

39

MP hominids claimed to have been buriedare incomplete. In fact, most of the putativeburials scrutinized later in this paper areincomplete! The fragmentary nature of theMP hominid fossil record has led mostworkers to conclude that articulated homi-nid remains are anomalous, and in need ofspecial explanation. The question arises asto whether the anomaly is real or imagined,and whether or not purposeful burial is theonly possible explanation.

The fossil record demonstrates that MPhominid skeletons had a low probability ofbeing preserved in such a way that burialwould be suspected. However, it is simplyunwarranted to assume that disturbanceprocesses would always have resulted indisarticulated and fragmentary hominidremains. There is, in truth, no good way ofassessing the prior probability of disturbancein the case of a hypothetical MP hominid. Atbest, one can get an idea of the posteriorprobability of preservation from the recorditself, and then only on a site-specificbasis. For example, suppose that the 23Neandertal specimens from Kebara Cave(Bar-Yosef et al., 1992) (a) were not pur-posefully buried, that (b) each represents theremains of a single individual, and (c) thatthey represent every Neandertal that everdied in the cave. On that evidence, the 23hominids apparently had about a one in tenchance of reasonably intact preservation.On the other hand, if one assumes (a)that purposeful burial was occurring atKebara, and (b) that all ages and individualsstood an equal chance of being buried upontheir death, then purposeful burial didnothing to improve the likelihood ofpreservation.

Either way one looks at the MP fossilrecord, the more or less complete andarticulated hominid remains claimed asburials are rare occurrences. Yet, to argueon that basis alone that their preservationmust be the result of purposeful burial is toignore the almost continuous range of com-

pleteness that exists in the fossil record,from the predominant isolated teeth andcranial fragments, to whole elements andarticulated body segments, to disarticulatedpartial skeletons, and to complete or nearlycomplete articulated or nearly articulatedskeletons. The most complete individualsare therefore only slightly more ‘‘anoma-lous’’ than the next-best preserved remains,and so on. Where does one draw the lineand say that such-and-such a portion musthave been buried and that another less well-preserved portion was not? This is not merequibbling; such perceptions are crucial tothe question of MP burial (cf. Binford,1968; Harrold, 1980). As I point out in thenext section, the presence of articulated por-tions of hominid skeletons is ultimately whatleads archaeologists to infer purposefulburial.

Yet while purposeful burial is a sufficientexplanation for articulated hominid remains,it is not a necessary explanation. Often, theinference of purposeful burial is not enoughto adequately explain the condition of thefinds. More often than not, auxiliary hypoth-eses are required. At times, some very tenu-ous auxiliary hypotheses have been neededto prop up the inference of purposeful burialfor partial skeletons—e.g., that the remainsare those of a secondary burial, or they resultfrom post-interment ritual—usually withoutany supporting evidence beyond argumentfrom want of evident alternatives.

On this point, it is worth reiterating thatonly a few of the relatively small number ofMP hominids claimed to have been buriedare technically complete. Most are partialskeletons, and some are quite incomplete(e.g., Kebara 2 is missing the right lowerlimb, most of the left lower limb, and thecranium (Bar-Yosef et al., 1992). It isreasonable to suggest that equal weight begiven to alternative explanations thataccount for the presence of articulatedskeletal elements, and that did not requirefanciful auxiliary hypotheses to explain

40 . .

the differential preservation of skeletalparts.

Finally, if MP hominid mortality was likethat of most species, including modernhumans before the advent of modern medi-cine, one would expect high numbers of veryyoung in most natural death assemblages.Despite the high probability that infantremains would succumb to destruction at ahigher rate than adult skeletons, infants dohave teeth and petrous portions of theirtemporal bones, both of which are extremelydense, and might be expected to preserve atleast as well as some parts of adults. Thisshould mean that, all things being equal,parts of juveniles should still preserve in highnumbers in localities where they, along withadults, were dying. The fossil record bearsthis out. For example, at Kebara Cave thereare more immature specimens than adult:eight under 1 year; five between 1 and 10years; one between 10 and 20; eight adults(Bar-Yosef et al., 1992). Thus, it is clear thatjuveniles, at Kebara at least, are being pre-served in some form as often as adults.Similarly high numbers of juveniles occur atQafzeh (Vandermeersch, 1981).

The question remains, however, ‘‘Howare the more fragile parts of immature homi-nids being preserved, and how does oneexplain the occasionally well-preservedinfant?’’ The simple answer is that the depo-sitional circumstances are as important tothe preservation of infant remains as they arefor adult skeletons. While infant remains areoverall more susceptible to destruction, thesame processes that can preserve adultremains have an almost equal capacity topreserve the fragile remains of infants.Thus, there is no need to view preservedimmature remains as anomalous. Theirpreservation depends on special depositionalcircumstances, no less than that of adultremains.

In sum, given the reality of deposition andpreservation in caves and rockshelters as Ihave outlined it, it should come as no

surprise that all of the Neandertal speci-mens whose depositional circumstances Ipreviously examined (Gargett, 1989a) werefound in circumstances that would beexpected to produce better than averagepreservation. For example La Chapelle-aux-Saints, several individuals at La Ferrassie,Kiik-Koba 1 and 2, and Roc de Marsal 1were discovered in depressions filled withsurrounding sediments. La Ferrassie 1 and 2were found on inward sloping sedimentsnear walls, where they would be less likely tosuccumb to trampling, and where theywould have been buried relatively rapidly.At Teshik-Tash the remains were found ina basin where water-borne sediments pre-served and obscured them. At Shanidarall nine individuals were discovered undercollapsed bedrock or in natural nichesformed by cave breakdown, which were sub-sequently buried by collapsing sediments.

I have argued, therefore, that somethingother than purposeful burial was preservingall of those remains. Clearly the depositionalenvironment has a powerful influence on theprior probability of hominid remains pre-serving. Add to this that the depositionalcircumstances are variable within individualcaves and rockshelters, and vary a great dealfrom site to site. Thus, it is virtually impos-sible to say what the prior probability ofpreservation would have been for a singlesite, much less for the entire MP. Certainlythere is no warrant for assuming that depo-sition would always have been too slow tonaturally bury a hominid corpse. There istherefore no point in arguing solely on thebasis of the fragmentary condition of mosthominid fossils, that the rare, articulatedspecimens must therefore have been pur-posefully protected. The presence of all theclaimed burials in depositional circum-stances that lend themselves to protectingand preserving skeletal material should alertpalaeoanthropologists to the possibility thatall of the so called burials are the result ofspecial natural depositional circumstances,

41

and not, therefore, the result of MPmortuary practices.

What is the importance of articulation?If a vertebrate decomposes fully withoutbeing buried, natural disarticulation takesplace—with the soft tissue gone, the skeletalelements become separate, and can be actedon by gravity and other disturbance pro-cesses. Logically, if burial did not occurprior to decomposition and disarticulation,and no other disturbance occurred, such askeleton would eventually be found with allof the elements present, and more or less inrelation to their anatomical neighbours, butnot articulated. Thus, for skeletal elementsto preserve in articulation, burial of somekind must occur prior to soft tissue decom-position. There can be no doubt that forthose rare hominids discovered with articu-lated portions, the remains must have beenburied prior to soft tissue decomposition.The question is whether or not that burialwas purposeful.

For the majority of palaeoanthropologiststhe inference of purposeful burial bestexplains articulated hominid remains.Indeed, the mere discovery of articulatedremains is viewed by many as prima facieevidence of purposeful burial. For example,in their treatment of Natufian and MPburial, Belfer Cohen & Hovers (1992:464)state: ‘‘. . . skeletal articulation remains thesingle unchallenged criterion for intentionalburial.’’ Their assertion, too, has goneunchallenged, which leads one to suspectthat this argument has many adherents (cf.Rak et al., 1994). However, implicit inBelfer-Cohen and Hovers’s position is theunwarranted assumption that in all casesinvolving MP hominids, natural burialwould be too slow to bury a carcass before itdecomposed and disarticulated naturally, orwas disturbed in some other way.

To examine this assumption, one needs tolook at the fossil record of other animals, tosee if articulated preservation occurs, and

under what circumstances. Natural burialoccurs in the fossil record, and occasionallyan articulated skeleton comes to light. Allthat is needed is special depositional circum-stances (cf. Brain, 1981:11). Often one findsthe articulated skeletal material of miredanimals, or those that have been buried incatastrophic floods. Articulated MP homi-nid remains are always found in specialdepositional circumstances, even when, asin the case of the Taramsa Hill hominid(Vermeersch et al., 1998), they may befound in open-air contexts. With that onepossible exception, the entire corpus ofputative MP burials comes from caves androckshelters. And, as I outlined above, cavesand rockshelters are above average in theircapacity to produce the kind of special depo-sitional circumstances that can lead to natu-ral burial. For example, in the FriesenhahnCave, two adult and one juvenile saber-toothed cat (Homotherium sp.) skeletonswere preserved, articulated, in Pleistocenebasin deposits (originally described inEvans, 1961; Marean & Ehrhardt, 1995).Friesenhahn Cave was an animal den, andnot simply a natural deadfall trap, where onemight expect to find abundant evidence ofcomplete (but not necessarily articulated)animal remains. The Homotherium speci-mens died where they had lived, a claimwhich is attested to by the presence of alarge quantity of juvenile mammoth boneinferred to have been the remains oftheir meals (Marean & Ehrhardt, 1995).(Students of MP burial will remember that asimilar depositional environment most likelyled to preservation of the, albeit disarticu-lated remains of the Teshik-Tash juvenileNeandertal.)

Without a complete review of the palae-ontological literature, it would be imposs-ible to say how many such examples ofarticulated vertebrate remains have beenrecovered, and how those would relatenumerically to the fragmentary ones,and whether or not that ratio differed

42 . .

significantly from the MP hominid fossilrecord. I do not intend to undertake such areview, or to produce such evidence in thispaper, because little would be achieved by it,even though many commentators have sug-gested its necessity. I decline to do sobecause it would be theoretically incapableof achieving any useful end. One cannot,after all, determine the prior probability ofoccurrence of articulated vertebrate skel-etons, given the uniqueness of every caveand rockshelter’s depositional and occu-pational history, and the uniqueness of everyspecies’ life history and behaviour. Let ussuppose, for example, that hominids spent80% of their time in caves, as against 50%for cave bears, and less for denning carni-vores. Those behavioural dissimilaritieswould be likely to result in a different priorprobability of preservation for eachspecies—the more time spent in such aplace, the higher the probability that deathwill occur there.

In any given depositional environment, allthings being equal, the species dying ingreater numbers would naturally preservein greater numbers. The probability of sub-stantial interspecific behavioural differencesrenders impotent any comparisons I couldmake on the basis of a review of thevertebrate fossil record. As mentioned ear-lier, I strongly suspect that interspecificbehavioural differences might explain thedearth of skeletal preservation similar to thatof MP hominids among the hominids ofearlier times, such as Homo erectus, whoseremains are sometimes found in caves.Rather than implying real cultural differ-ences, which would be the position held bymost proponents of MP burial, it maysignal nothing more than that H. erectusspent little time in caves (cf. Binford & Ho,1985), or that prior to the MP such spaceswere predominantly the domain of denningcarnivores that later became scarcethrough competition with more success-ful predators, such as the Neandertals

(Gamble, 1993). Thus, H. erectus remainswere more often than not subjected to sub-aerial conditions, leading to a much lowerlikelihood of survival. Commentators onmy 1989a paper who pointed out that theMP hominid record was unique in thenumber of articulated specimens may simplyhave failed to appreciate the complexity ofthe issue, including the importance tobone’s survival of the interplay of site type,depositional regime, and behaviour.

Thus, because of the complexities of siteformation and taphonomy, one could neverknow how often hominids might have beenburied naturally in cave contexts, no matterhow much is known about the rest of thefossil record. Notwithstanding our inabilityto assess the prior probability of burial, onecan still know something about otherprocesses which would have a bearing onwhether or not a hominid would be pre-served in articulation by natural sedimen-tation. From knowledge of such processesone can develop expectations for whatwould happen to a vertebrate after death ifit were allowed to decompose naturally in acave or rockshelter. Such expectations,based on knowledge of processes acting inthe present, go a long way towards ex-plaining the preservation of articulated MPhominid remains.

What is the variability in decomposition rates,disarticulation sequences, and the likelihood ofdisturbance?If vertebrates naturally preserve with ele-ments articulated, one has the beginningsof an empirical framework to assess theclaim that ‘‘skeletal articulation remainsthe single unchallenged criterion for inten-tional burial’’ (Belfer-Cohen & Hovers,1992:464). To take the framework further,however, one must look in more detail atthree processes that condition the prob-ability that a skeleton, or portion thereof,would remain intact while it was naturallyburied—i.e., decomposition, disarticulation,

43

and likelihood of disturbance. Implicit inBelfer-Cohen and Hovers’s statement is theassumption that natural burial of articulatedskeletons would be highly unlikely, if notimpossible. To examine that assumption Ibegin by specifying the range of variability indisarticulation rates.

Decomposition. Disarticulation occurs aspart of the process of decomposition. Tobetter understand the likelihood of a homi-nid skeleton preserving with its elementsarticulated, one must have an appreciationof the variables affecting decomposition, andthus the variation in the time it takes ananimal carcass to become disarticulated.When soft tissue connecting two skeletalelements decomposes, the elements canbecome disarticulated—i.e., they willbecome displaced from their anatomicalrelationship. However, if the organism doesnot decompose, or only decomposes partly,in the absence of disturbance it cannotbecome completely disarticulated. The rateat which decomposition occurs, as well asthe way that it proceeds, are influenced bymoisture and temperature (e.g., Micozzi,1986; Galloway, 1989).

As Micozzi (1997) points out,

‘‘postmortem change is essentially a compe-tition between decay and desiccation, and. . . temperature and humidity largely deter-mine the outcome.’’

The rate at which a dead organism dries outwill therefore have a bearing on whether ornot, and the degree to which disarticulationwill occur. Rapid desiccation will occur inarid environments, regardless of the tem-perature. Under circumstances of relativelyhigh temperature and humidity, decomposi-tion will prevail over desiccation. However,as the temperature drops, decompositionslows, and desiccation progresses to thedetriment of decomposition. The lower theambient temperature, the longer the carcasswill take to decompose and the lower will bethe degree of desiccation required to halt

decomposition and thus preclude disarticu-lation. For example, between 15) and 37)Cdesiccation must be rapid to halt decompo-sition; below 5)C bacterially induced putre-faction will cease, and only a minimal degreeof desiccation is required to preserve softtissue (Micozzi, 1997). Finally, freezing willpreserve a carcass indefinitely. When and ifthawing occurs the effects of anaerobic bac-teria are minimized (Micozzi, 1986)—andthus bloating in decomposing carcasses, andthe concomitant straightening of limbs, areless likely to occur. If a carcass remains dry,skeletal elements can stay articulated for aconsiderable time. Just how long is difficultto estimate, although in the arid environ-ment of the American southwest, mummified organic materials can survive forhundreds of years (Galloway, 1989). At theother end of the moisture spectrum, as theHomotherium remains from FriesenhahnCave (mentioned above) demonstrate,deposition in wet sediments can also pro-long decomposition and ensure preservationof articulated skeletons.

Although temperature and humidity arein general determined by climate, climaticconditions are not the only source ofvariability in ambient temperature ormoisture in, especially, caves. Temperaturein caves and rockshelters varies accordingto aspect—e.g., south-facing caves in thenorthern hemisphere could be expected toremain warmer longer than those facingnorth. And even with today’s equable worldclimate, north-facing shelters in the north-ern hemisphere tend to stay cold even whenthe ambient temperature is relatively warm.Humidity, too, is subject to variation. Evencaves developed in karst can become drywhen the wet karstic conditions that createdthe cave cease. For example, I have observeddesiccated, articulated, large mammalianremains in a karstic cave in Israel, where,clearly, the wet conditions that created thecave had given over to a much drier regime.Thus, the chance that temperature and

44 . .

humidity combined to preserve the fewarticulated hominid remains may have beenvanishingly small, yet it still could haveproduced the specimens claimed to havebeen purposefully buried.

In spite of being aware of variability indecomposition rates, in practice one cannotsay for certain if a given hominid was likelyto have undergone a protracted decompo-sition. This is because no one can say withany certainty what conditions prevailed atthe time the hominid died, nor how muchthis may have contributed to its preservationas an articulated skeleton. One could con-tend that such events were rare in the MP.However, in the fossil record one is alreadydealing with the results of rare events. Andone can only guess at how rare the knownfossil hominids are in relation to all thosethat died without leaving a trace.

Even if one cannot say for certain that agiven specimen was preserved because ofenvironmental factors, it is safe to say thatsome of the southwestern Asian specimenswere likely to have been subjected to morearid conditions than the Neandertals ofwestern Europe, whose corpses were in turnlikely to have been subjected to extremelycold, often periglacial conditions. Ambienttemperatures sufficient to freeze a hominidcorpse would be expectable in rocksheltersand cave entrances under periglacial con-ditions or, for that matter, during most mid-to high-latitude winters during interglacials.Moreover, even in more or less humid,present-day coastal localities, Pleistoceneair and sea temperatures would have waxedand waned in combination with marineregression and transgression cycles creatingfavourable preservation conditions at sometimes and not at others. As I pointed outabove, retarding decomposition to the pointwhere disarticulation was precluded wouldrequire some combination of ambient tem-perature below 5)C and only relatively lowhumidity. At 4)C bacterial growth isextremely slow, and at 0) no desiccation

would be required for preservation to occur(Micozzi, 1997). Any carcass deposited ina cave or rockshelter under the right con-ditions could have been naturally buriedbefore it became partially or completelydisarticulated. In sum, one cannot simplyassume that purposeful burial is required topreserve an articulated skeleton.

From a taphonomic perspective, the fore-going discussion offers archaeologists a newline of evidence for inferring depositionalhistories: the degree of articulation canprovide evidence for the degree to which acarcass decomposed prior to being encasedin surrounding sediments. All thingsbeing equal, a hominid buried in aninstant—whether by purposeful burial or ina rockfall—should be preserved with moreelements articulated than one that hasundergone some pre-burial decomposition.Therefore, as I explain below, it shouldbe possible to assess the likelihood thata carcass has lain exposed for a time beforeburial, as well as the relative time it layexposed, by reference to its degree ofarticulation, and the species-specificdisarticulation sequence.

Disarticulation. Vertebrate skeletons donot disarticulate willy-nilly. A sequence ofdisarticulation that is more or less uniformacross a species allows predictions to bemade about which elements can be expectedto disarticulate in what order (Toots, 1965;Hill, 1979; Hill & Behrensmeyer, 1984;Weigelt, 1989). In a long-term study ofvertebrate decomposition and natural car-cass disarticulation, involving the EastAfrican bovid Damaliscus korrigum, Hill(1979) found that there is an anatomicallylogical, more or less regular sequence ofdisarticulation and decomposition. Hill &Behrensmeyer (1984) found a high corre-lation in the order of disarticulation acrossfour species of ruminant artiodactyls andone equid species. Micozzi (1986) was ableto find a similar sequence in laboratorymice. The sequence of disarticulation has

45

thus been well documented for some mam-malian groups. It appears to be equally trueof humans, although detailed studies similarto those documented for animals have yet tobe carried out on human subjects. The dataon the human sequence of disarticulationare thus coarse-grained (see e.g., Haglund,1997; McKeown & Bennett, 1995, on esti-mating time since death on the basis of toothloss; Micozzi, 1986 for an attempt to usemice as analogues for humans).

Although the disarticulation sequence ofother mammalian taxa might not be perfectmodels for hominid decomposition, anumber of anatomical similarities resultingfrom common evolutionary descent make itpossible to use them as analogues to modelthe order in which hominid skeletal ele-ments would come apart, if allowed todecompose without being encased in sedi-ments. The mammalian disarticulationsequences comes into play, when the Kebara2 discovery is examined below.

Disturbance. Decomposition is just one ofthe ways a hominid skeleton might becomedisarticulated. Scavenging carnivores andsubsequent cave occupants could also havehad a significant impact on unburied orpartly naturally buried hominid corpses.The Teshik-Tash Neandertal remains, forexample, were subjected to carnivore modi-fication (Movius, 1953), and were thusincomplete and disarticulated. While thereis good reason to suppose that carnivoredisturbance may have occurred in someplaces and at some times throughout theMP, there is no warrant for the assumptionthat it would always have threatenedhominid remains lying exposed.

One might readily infer that the majorityof MP hominid corpses did succumb tocarnivore and other kinds of disturbance;given that so much of the record consists offragments. The corollary of this is that thosefew articulated skeletons that survived did sobecause they sustained no carnivore or otherdisturbance, for reasons that are the subject

of this paper. However, there is great varia-bility inherent in the desirability of hominidremains to scavengers, as well as the poten-tial for great variability in the distribution ofpredators and scavengers. Desiccation willrender a carcass undesirable to scavengingcarnivores (Gifford, 1981), and Haynes(1982:268) reports that wolves ‘‘prefer freshmeat . . . to feeding very long on frozencarrion.’’ Therefore, at times hominidremains might not have been sought by suchanimals. In fact, absence of carnivore distur-bance might mean nothing more thanthat they were absent from the area atthe time the carcass was in an otherwisedesirable state. Gamble (1993:166), forexample, argues that carnivores were beingoutcompeted by hominids, thus explainingthe apparent absence of carnivores fromEuropean sites during periods of hominidoccupation. In a situation where, forexample, carnivores had been successfullyout-competed by MP hominids in a localarea, the prior probability of carnivore dis-turbance would have been reduced to zero.Straus (1982) argues that, usually, hominidand carnivore occupations are alternating,rather than contemporaneous, suggestingthat the two kinds of animal may haveavoided one another’s habitations (cf.Stiner, 1994). The length of time betweenalternating occupations might have hadan impact on the likelihood of a hominidskeleton preserving intact, but it is equallylikely that corpses in special depositionalcircumstances would in any case escape dis-turbance, whether or not temperature orhumidity or both had rendered the corpseundesirable—rockfall, for example, wouldinstantly have protected hominid remainsfrom subsequent disturbance, includingby carnivores, thereby reducing the priorprobability of disturbance to near zero(although it would always be possibleto argue that scavenging carnivores werecapable of digging up such naturally buriedhominids).

46 . .

Disturbance by subsequent occupantswould have been another potential threat tothe unburied corpse of a hominid. However,many of the hominid discoveries I examinedfor the 1989a paper, and several of those Icritique in the present work, would likelyhave escaped such subsequent disturbancequite easily, by having been buried, natu-rally, in bedrock breakdown events (e.g.,those in Shanidar Cave). Others ended upnear walls, or in natural niches, where tram-pling would have been less likely, and thuswould have had a higher prior probability ofbeing preserved. So, as with carnivore dis-turbance, one need not presume that contin-ued occupation or re-occupation of a caveor rockshelter subsequent to the death of ahominid would have disturbed the remains.Again, it must be assumed that the vastmajority of MP hominid remains suc-cumbed to some form of physical distur-bance, since there are so few completeskeletons. But postmortem disturbance, ofwhatever kind, need not, and I would arguedid not happen in every case.

SummaryThe vast majority of MP hominids that diedin caves and rockshelters must have had ahigh prior probability of being dispersedand fragmented prior to natural burial, orsubsequently, due to postdepositional dis-turbance. With all that can happen to avertebrate after death, survival of an articu-lated skeleton thus acquires the appearanceof being an event of low probability. On theevidence, survival of even fragments of MPhominids must have been a highly improb-able event. However, in reality one has noway of assessing, a priori, what that prob-ability was. All that can be known is that,a posteriori, it appears low.

In spite of the low probability of preser-vaton, hominid skeletal remains occasionallypreserve in a manner that suggests they werepurposefully buried. It was the presence ofarticulated hominid remains that led to the

first inferences of burial, and it continues tofuel claims for purposeful MP burial. Yet, asI have argued, purposeful burial is by nomeans the only way that hominid remainscan preserve intact and articulated. Thus,because hominid remains can be preservedin articulation by more than one means, theprimary interpretive task for one who sus-pects purposeful burial should be to specifythe range of natural processes that couldhave been responsible. When, and if, itbecomes clear that natural processes alonecould not account for the archaeologicaldiscoveries, then human or hominid agencymight reasonably be inferred.

Unfortunately, there exist no straight-forward means of assessing the credibility ofclaims for MP burial. Each case is unique.Thus there is no simple recipe forinterpretation—the taphonomic assessmentof hominid burial is by nature qualitative,not quantitative; inductive, not deductive;interpretive, not prescriptive. Under thecircumstances one would like to turn toempirically-based research for help inassessing claims of MP burial. However, todate there has been no explicit research into,or modelling of, the processes involvedin natural burial of articulated vertebrateskeletal material, and certainly none thataddresses the issue of whether or not MPhominids buried their dead.

One thing is certain, the wide range ofcompleteness and degree of articulation rep-resented in the MP fossil record demandsexplanation. Given the variability inherentin cave and rockshelter site formation pro-cesses it is tenuous at best to accept theunwarranted assumption that hominid skel-etal remains could only have preserved inanatomical connection due to purposefulburial, especially given that claimed burialsrepresent an inestimably small fraction ofthe MP hominids that lived and died. In thecase studies that follow, I present evidencethat claims for MP burial are greatly weak-ened when viewed in light of expectable

47

natural processes and the variability inherentin them.

I end this section where I began it, bysuggesting that to adequately assess claimsfor purposeful MP burial, taphonomichistories of the hominid skeletons need to beinferred. For each specimen, such a historywould necessarily include, but not berestricted to, the following observations.- Is there evidence of a new stratum con-

taining the remains created at the timethe remains were buried, and distinctfrom those upon which the newstratum was deposited, and from thoseaccumulating above?

- Is the skeleton complete? Missing or dis-placed elements could mean that theybecame disarticulated during a process ofnatural burial, or that they simply neverpreserved, due to particular circum-stances of deposition. Missing elementsmay also be due to postdepositional ero-sion, chemical decomposition (i.e., dia-genesis) or other disturbance, such asburrowing.

- Is the skeleton articulated? If so, areall elements articulated? If not, whatexplains the disarticulated portions?Gradual burial of a slowly decomposingcorpse could be indicated.

- Are the bones of the skeleton preservedequally? Unequal preservation could bethe result of microdepositional variation,identification of which could also help toexplain preservation of the remainingportions.

- Is the skeleton fragmented? Are thebreaks peri- or postmortem? Perimortembreaks may be the result of catastrophicrockfall, while postmortem breaks mayonly signal sediment compaction. Post-mortem breaks could equally well, how-ever, imply exposure to destructivepostdepositional sedimentation.

- What is the position of the skeleton?Could a similar pose have beenadopted during sleep or because of a

traumatic injury, or in extremis, or could itonly have occurred due to mortuarytreatment?

- Is there evidence of disturbance dueto trampling? Is there any evidence ofportions that could have been broken ordisplaced by trampling during a processof natural burial?

- What are the characteristics of the bed-rock? Is it likely to have broken downcatastrophically, or gradually, or both?

- Does the cave or shelter occur in a tec-tonically active area? If so, it may be thatepisodic catastrophic breakdown hasoccurred.

- What are the characteristics of the depo-sitional contexts in which the remains arefound? Are the remains in a low spot thatwould have accumulated sediments moreor less rapidly, or in some other micro-depositional environment that wouldnaturally bury or otherwise protect acorpse, such as near a wall, or in a naturalniche?

- Do the sediments themselves tell a story?Are there rodent burrows, erosionalchannels, crushed skeletal material, bro-ken stone tools, or any other evidence tosuggest disturbance of one kind oranother?

The foregoing is by no means an exhaustivelist of what needs to be considered whenexamining claims for purposeful burial inthe MP. Numerous natural processes andpossible combinations of natural processesare capable of preserving articulated homi-nid remains in caves and rockshelters. Thus,it is incumbent upon the archaeologist tolook beyond the unwarranted assumptionthat articulated remains presuppose pur-poseful burial, and to consider all the evi-dence that might have a bearing on whetheror not the remains could have been naturallydeposited.

In the following sections of the paper, Iexamine published accounts of the findingsfrom Qafzeh Cave, Saint-Césaire, Kebara

48 . .

Cave, Amud Cave, and Dederiyeh Cave. Ilook critically at the arguments for purpose-ful burial. In each case I approach the obser-vations and the claims with a view towardsruling out the natural processes that couldhave preserved articulated hominid remains.My conclusions should be regarded as alter-native models constructed to explain thearchaeological occurrence of well-preservedhominid remains, and are themselves sub-ject to revision upon presentation of newevidence. I begin by looking at the discover-ies from the Mousterian levels of QafzehCave, where a number of skeletally modernfossils were recovered in the 1970s.

Qafzeh Cave

In the 1970s a joint Israeli and Frenchexpedition recovered the skeletal remains ofas many as 18 MP hominids during severalepisodes of excavation at Qafzeh Cave, nearNazareth, Israel (Vandermeersch, 1966,1969, 1970, 1981). Of the 18, most werefragmentary, three were articulated partialskeletons and two were articulated, nearlycomplete skeletons. The presence of somany hominid fossils was already enoughto make Qafzeh an important palaeo-anthropological locality. However, it grew insignificance with the publication of thermo-luminescence dates of 90 to 100 ka BP forburned Mousterian flint artefacts inferred tobe contemporaneous with the fossil remains(Valladas et al., 1988). The Qafzeh fossilhominids, deemed to be skeletally modern,were thus contemporary with the plesiomor-phic Neandertals. Discovery of morphologi-cally modern Homo contemporary withNeandertals in the Near East and Europebrings into sharp relief the question of whenhominids became human, especially giventhe ongoing debate about what Mousterianstone tools imply about the cognitiveabilities of their makers. Thus, whetheror not any of the Qafzeh specimenswere purposefully buried has a direct

bearing on questions of modern humanorigins.

At the time of their discovery it wasbelieved that the Neandertals had buriedtheir dead. There was thus no need toquestion whether or not these remainswere purposefully protected, because ofthe expectation, common at the time, thatmorphologically modern hominids wouldhave been even more like modern humansthan the Neandertals. Nevertheless, there issufficient evidence to suggest that theQafzeh hominids could have been naturallyburied.

The catalogue of hominids from recentexcavations in Qafzeh Cave includes:- Qafzeh 8, a partial adult skeleton recov-

ered from Couche XVII, including a por-tion of the tooth-bearing lower face withsome teeth, a fragment of the right oscoxa, and fragments of the right upperlimb and lower right and left limbs;

- Qafzeh 9, an almost complete adultrecovered from Couche XVII;

- Qafzeh 10, a child’s skeleton recoveredfrom Couche XVII, near the distal limbsof Qafzeh 9;

- Qafzeh 11, the upper portion of aninfant, was discovered in Couche XXII;

- Qafzeh 12, a child’s cranium recoveredfrom Couche XVII;

- Qafzeh 13, an incomplete foetal cranium,found in Couche XVa;

- Qafzeh 14, a fragmentary child’s craniumrecovered from Couche XVII;

- Qafzeh 15, the upper portion of a child’sskeleton recovered from Couche XVII.

Of the eight individuals, six were recoveredfrom Couche XVII, a spatially limited sedi-mentary unit mostly comprising limestonerubble infilled with fines (Vandermeersch,1981:27). Qafzeh 11 was earlier in time;Qafzeh 13 was later. Whether judged byreference to the remaining sediments atQafzeh, or by comparison with most otherMP fossil localities, the contents of CoucheXVII would be an astonishing collection of

49

fossil hominids—six individuals in the spaceof approximately 3 m3 of sediment. Nowonder they are unquestioningly deemed tohave been purposeful burials—the density ofskeletal material in some ways resemblesthat of present-day cemeteries!

Protection of the corpseThe first place to look for evidence of pur-poseful burial is the immediate stratigraphiccontext of the skeleton. As with all other MPhominids, none of the Qafzeh specimenswas found in a new stratum created atthe time of its deposition. Stratigraphic evi-dence, therefore, cannot be used to say thatthe hominids were purposefully protected atthe time of their death.

Despite the absence of unequivocalstratigraphic evidence, one specimen wasrecovered in depositional circumstances thattantalized the excavators, and which appearto be, at least in part, artificial. Qafzeh 11was discovered in one end of a 20–25-cm-deep depression in the deepest depositsof the cave’s vestibule. Vandermeersch(1970:299) describes the niche as ‘‘altered’’bedrock, composed of calcareous rubble en-veloped in calcareous sand. The report goeson to suggest that two bedrock boulders, oneatop the lower limb of the 10-year-old, wereemployed to shore the putative grave’s sides,which were so unconsolidated that theycollapsed easily. Although the presence of adepression is indisputable, the inference thatit functioned as a grave is less secure.

As I mentioned previously, all kinds ofanimals, as well as people, are capable ofcreating such hollows as comfortable restingplaces, and not just as places to inter thedead. As with the other specimens atQafzeh, the absence of a distinct new stra-tum created at the time Qafzeh 11 wasburied makes the burial hypothesis lessrobust. Not even the presence of an antlerin the individual’s possession necessarilyimplies that it was a grave offering.Presumably such faunal remains occur in

Mousterian contexts from time to time. Pre-sumably they were being transported thereby hominids. And if Qafzeh 11 had been inpossession of the antler at the time of itsdemise, and it met with no subsequent dis-turbance, presumably it would have beenpreserved along with the skeleton.

Condition of the skeletal remainsAll of the Qafzeh skeletal material is uni-formly described as highly fragmented. Onlytwo specimens are completely preserved, ornearly so—Qafzeh 9 and 10. Qafzeh 8,although it is clearly an articulated skeleton,has been affected by postdepositional pro-cesses that destroyed the left (upper) side ofthe skeleton. Even so, Qafzeh 8’s survivingelements are for the most part crushed andoften incomplete. The remaining hominidspecimens are also less than perfectly pre-served and incomplete. Thus, there is evi-dence of considerable peri- or postmortemdestruction. If purposeful burial was occur-ring, it did a poor job of protecting theseindividuals from destructive processes.Unless one is to imagine that they werepurposefully buried in an incomplete state,one must explain how they came to be thatway. Natural burial by expectable depo-sitional processes provides a parsimoniousexplanation for the Qafzeh remains.

Whether purposefully buried or naturallyencased in sediments, one must alwaysexpect that in caves and rockshelters erosioncan occur. Whenever it rains in caves likeQafzeh, run-off from the surface above istransmitted through a chimney to the talusslope below, and then out through the hori-zontal opening. At such times surface ero-sion is unavoidable, and quite capable ofexposing portions of buried skeletons, whichcould lead to their eventual destruction.Erosion could easily account for the absenceof Qafzeh 8’s cranium, or the missing lowerportions of Qafzeh 11 and 15. However,from an excavator’s standpoint, evidence forsuch erosion might be elusive.

50 . .

If any of the Qafzeh hominids had beencaught in a shallow rockfall of variabledepth, the missing parts of individuals maysimply not have been covered by enoughsediments to ensure preservation. Thefragmentary state of several of the Qafzehhominids might therefore stand as proxyevidence that Couche XVII accumulatedrapidly, perhaps in an instant. Given thepossibility of rockfall, the presence of somany remains in the same location becomeseasier to explain—the brow of the cave is thebest-lit portion, with a vantage of the wadioutside. The cave mouth is also where mostbreakdown due to freeze–thaw and wet–dry cycles would be expected. Hominids(perhaps) habitually inhabiting an area ofinherently unstable ceiling bedrock, coupledwith the cave’s location in a tectonicallyactive area, meant that the ‘‘vestibule’’would have been a prime candidate forrepeated catastrophic preservation of homi-nid remains.

While erosion or incomplete burial couldexplain the absence of parts of some skel-etons, erosion cannot explain the crushingand deletion of portions of the elements thatdid survive. Qafzeh 8, for example, pre-served some of the right carpals complete,while others are crushed; some of thephalanges survived complete, while othersare crushed, and some are absent. Moreoverthere are other similar examples of differen-tial preservation from this specimen. Theuppermost portions of Qafzeh 8 are missingand were very likely affected by the processthat resulted in brecciation. However thisprocess cannot account for all of the missingportions. Diagenesis would probably haveaffected articulating elements, such as thecarpals, equally, and would not, therefore,have resulted in differential preservation ofsome of the wrist bones of the right hand.Thus some other explanation must besought. The evidence suggests rockfall.Those parts that survived intact mayhave been just those that escaped direct

impacts by falling rubble. Qafzeh 9 presentsperhaps the clearest evidence for such anevent.

Qafzeh 9 is virtually complete. However,Vandermeersch (1981:51) describes itscondition in this way:

In the grave, it lay buried on the left side andwas crushed by the weight of the sediments,to the point where the right and left ribs werein contact with one another. The almostcomplete absence of sediments in the spacebetween permits one to infer that the crush-ing must have occurred a short time afterthe inhumation. This crushing has brokenthe cranium into numerous fragments [mytranslation]

Besides the crushed elements, the crucialevidence given here is that there are nosediments between the left and right ribs,with the implication that damage to theribcage occurred before the soft tissuedecomposed. Otherwise, as the soft tissuedecomposed, sediment would have leakedinto the chest cavity and been present on thelowermost bones of the ribcage when theupper (left side) ribs collapsed. Further-more, there is the inference that Qafzeh 9’sribcage was crushed by the weight of sedi-ments accumulating above it after purpose-ful burial. Does the absence of sedimentbetween the ribs necessarily imply thatcrushing occurred ‘‘a short time’’ afterdeath, or can it imply that crushing occurredperimortem? One way to clear up the matterwould be to examine the break margins forevidence of peri- or postmortem damage.Columnar fractures and right angle breakmargins would imply that the damageoccurred well after the individual’s death,while spiral fractures and break margins withacute angles would more than likely occurperimortem. However, even if the bones areequivocal on the matter, the ribs supplyevidence that the crushing occurred close tothe time of Qafzeh 9’s death.

Qafzeh 9 occurs in a stratum that may beno more than 50 cm thick (see below for a

51

discussion of the cave’s stratification). Itwould be difficult to specify, but consider-able pressure must be required to collapse ahominid ribcage when the bone is fresh.Therefore, I question whether 50 cm ofsediments, placed carefully on the corpse ina purposeful burial, would have had suf-ficient energy to collapse the ribcage prior tosoft tissue decomposition. And, if greaterthan 50 cm of sediments was required tocrush the ribcage, by the time enough sedi-ment had accumulated above the buriedcorpse it seems highly likely that the corpsewould have decomposed to the point wheresediments were entering the chest cavity.Even if Qafzeh 9 had become desiccatedafter being purposefully buried, it seemsimprobable that no sediment would haveintruded between the left and right sidesbefore or during the collapse of its ribcage(although that possibility cannot be ruledout).

Thus, a more parsimonious explanationfor Qafzeh 9’s condition would be thatthe crushing weight of falling sedimentsoccurred at the moment of her death due tosudden bedrock breakdown. On the evi-dence, Qafzeh 10, the infant, succumbed tosimilar forces, which would be expected if ithad been in the care of Qafzeh 9 when theywere both killed by rockfall. Qafzeh 8,likewise, appears to have been crushed.

Finally, crushing is not restricted to thehominids in Couche XVII. The entireunderside of Qafzeh 11 is described as beingdamaged almost beyond recognition, andthe pelvis is completely crushed. Theribs, too, were ‘‘crushed against the rock’’(Vandermeersch, 1970:299). Moreover, theindividual was very probably hit sharplyon the right forehead (Vandermeersch,1970:299), which left a depressed fractureand a hole in the cranium. It is described asfollows:

On the right side of the frontal, about 3 cmbehind the frontal boss, there is a deepdepression, 2 cm across, which ends in a

hole 8 mm by 5 mm. On the back, the bonedisplays a regular bevel [biseau]

A break margin described as beveled, ratherthan at 90) to the surface, suggests perimor-tem, rather than postmortem damage. Thus,despite the presence of a shallow depressionin which the remains were discovered, con-tained in the description of Qafzeh 11’sdiscovery is evidence that this individualmay have been killed on the spot by fallingrubble. Altogether, the evidence fromQafzeh Cave strongly suggests natural burialin a destructive depositional environment.

Position of the corpseAll of the Qafzeh specimens are in poses thatcould have occurred naturally. However, theproximity and relative position of Qafzeh 9and 10 are used to postulate that the twocomprise the first multiple burial in thearchaeological record (Vandermeersch,1970:32). Qafzeh 9 is lying on her right side,with her right arm slightly adducted andslightly flexed at the elbow, and her righthand resting under her right femur; her leftupper arm is abducted, and slightly flexed atthe elbow. Her femora are flexed abouthalfway from the anatomical position, herknees are also flexed at about right angles tothe femora; her left foot is extremelyinverted and slightly plantarflexed; her rightfoot is slightly inverted, and in extremeplantarflexion. Qafzeh 10, the infant, wasdiscovered near the feet of Qafzeh 9. Theinfant’s right arm is adducted posteriorlyand flexed slightly at the elbow, the left isdifficult to make out in photographs. Theinfant’s lower limbs are both slightlyadducted, with slight flexion of the kneesapparent (Vandermeersch, 1990).

As I said before, basing an hypothesis ofburial on the position of skeletons is fraughtwith problems. Nothing about the positionof these individuals implies intentional prep-aration, or compels acceptance of the burialhypothesis. One can easily postulate what

52 . .

would have happened to a female carryingan infant if a small earthquake had loosenedseveral cubic metres of sediments from theceiling of this cave. If caught in such arockfall an upright adult female with aninfant in her arms could easily have droppedthe vulnerable subadult at her feet as bothperished. Multiple burial, as postulated byVandermeersch, seems on the evidence tobe only one possibility among many, and anunlikely one, at that.

ArticulationMost of the Qafzeh specimens were articu-lated. This should come as no surprise,because whether encased in a rockfall,gradually buried by aggrading sediments inthe cave during prolonged decomposition,or purposefully buried, one would expect askeleton to remain articulated.

Characteristics of the bedrockVandermeersch (1981) describes CoucheXVII, in the cave’s mouth, as being com-posed predominantly of calcareous rubblefilled with fine sediments. These depositsoccur at the distal end of an immense taluscone, and are evidence of a long history ofbedrock breakdown. On a visit to Qafzeh in1989 I observed the bedrock within whichthe cave developed. It is steeply dipping andfinely bedded calcareous rock that breaksdown into pebble- and cobble-sized clasts.Such parent material is intrinsically moresusceptible to breakdown than more mas-sively and horizontally bedded materialwould be (Jennings, 1985:164). Moreover,relatively constant breakdown of the friablesediments would have occurred throughoutits depositional history, especially near themouth of the cave, where diurnal tempera-ture gradients and wet/dry cycles wouldhave increased the rate at which the bedrockdegraded (Farrand, 1985). The witness pro-file left by the archaeological excavationsnear the mouth of the cave also testifies toQafzeh’s history of more or less continuous

breakdown during deposition of CoucheXVII. Finally, Qafzeh occurs in one of themore tectonically active areas of the world(i.e., at the northern end of the East AfricanRift). Such inherently unstable bedrockwould probably have produced a hailstormof potentially lethal particles and dust duringan earthquake.

Characteristics of the sedimentsThe hominids in Couche XVII are buried ina homogeneous layer of larger clasts infilledwith fine sediments. As described by itsexcavators, Couche XVII varies from about10 to about 50 cm deep (Figure 1), averagesabout 25 cm, and covers an area about 3 mby 4 m (Figure 2). It occurs near the bottomof a suite of distinctly stratified sedimentsdefined on the basis of granulometry andcolour: Couches XV to XVIII. However,although XVII is given a separate descriptivetaxon in the stratigraphic column, there issome ambiguity in Vandermeersch (1981)regarding its distinctiveness. Immediatelyabove and contiguous with it is a massiveaccumulation of brecciated sediments,named Couche XVI, described as part of thelarger group comprising XV to XVIII, anddiffering from the rest only in that it isbrecciated. Thus, Couche XVII mightsimply be the unaltered lower portion of amuch thicker stratum that built up as aresult of a uniform depositional process.However, Vandermeersch (1981:27), statesthat the brecciation masks the contactbetween XVI and XVII, which implies thatthe excavators did in fact recognize a strati-graphic break between XVII and theoverlying XVI. In addition, the publishedprofile shown in Figure 1 depicts a dottedline, which is presumably an inferred con-tinuation of the upper limit of CoucheXVII—i.e., a clear indication that theexcavators had decided Couche XVII wasa discrete depositional unit. In spite ofthe ambiguity introduced in its descrip-tion, there seems to be some warrant to

53

Figure 1. Qafzeh Cave. North-south stratigraphic profile, at western excavation limit. Most of the claimedburials come from stratum XVII (after Vandermeersch, 1981:28, Figure 6). Reproduced with permissionfrom Les Hommes Fossiles de Qafzeh (1981) Paris, C.N.R.S. Copyright retained by C.N.R.S., Paris.

identifying Couche XVII as a separatedepositional context.

Further evidence for Couche XVII’s dis-tinctiveness comes from the description ofits constituents. Couche XVII is unique in

this suite of deposits in being described ascontaining ‘‘nombreuses traces de foyers’’(Vandermeersch, 1981:27)—i.e., numeroustraces of hearths (my translation). Thisdescription distinguishes XVII from those

54 . .

Figure 2. Qafzeh Cave. Schematic representation of locations of claimed burials in plan (afterVandermeersch, 1981:33, Figure 8). Note that not all individuals are complete, nor are the poses of theskeletons uniform. Reproduced with permission from Les Hommes Fossiles de Qafzeh (1981) Paris,C.N.R.S. Copyright retained by C.N.R.S., Paris.

above and below it. It also implies that theevidence for hearths is not in the form ofhearths, per se, but something else that beliesthe onetime presence of hearths. One caninfer this from the description of the earlierCouche XXI, which is stated to havecontained ‘‘numerous hearths.’’ There isthus some question as to the nature of theevidence for hearths in XVII. It could,for example, be in the form of dispersedcharcoal, in which case it may have filtereddownward into an otherwise heterogeneousjumble of bedrock collapse material. Giventhat the evidence in XVII and XXI isdescribed differently, one is left to wonderwhether the ‘‘traces’’ in XVII are truehearths. Moreover, given the effort todescribe the sediments of XVII as distinctfrom the rest, including XXI, it is reasonableto conclude, at least tentatively, that XVII is

a distinct depositional unit, and that the sixhominid specimens within it were depositedin a horizontally and vertically restrictedspace. The absence of true hearth featureswould lend support to the suggestion thatCouche XVII accumulated rapidly, and inthe process buried a group of MP hominids.On the other hand, if complete hearths arepresent in Couche XVII, it might only implythat its deposition was not instantaneous,but rather, episodic. Such a scenario wouldnot preclude natural burial of the Qafzehhominids in Couche XVII.

A geomorphologist would have a difficulttime determining macroscopically if any ofthe Qafzeh sediments were the result of arockfall—i.e., whether they had accumu-lated rapidly enough to kill and bury thehominids found there. However, the crush-ing evident in all of the skeletal material is

55

strong support for the occurrence of rockfall(Ferrand, 1985).

SummaryDealing with what they thought were mod-ern humans, the excavators assumed thatthe Qafzeh hominids had been buried.There is thus little in the way of explicitargument in the published account towardswhich one can address critcism. In spite ofthis, it is possible to assess whether or notnatural processes could have been respon-sible for the fossil remains. The cave isdeveloped in friable, thinly bedded andsteeply dipping calcareous rock. There isample evidence that bedrock in the cave’svestibule has been breaking down more orless continuously throughout its history.Villa (1989:325) views as anomalous a simi-lar situation at Shanidar, in which severalindividuals died in the same part of the cave.Rather, I have argued (Gargett, 1989a) thatthose individuals were entombed by a time-transgressive phenomenon—i.e., localized,episodic bedrock breakdown beneath adeveloping ceiling vault. A similar fate mayhave befallen the Qafzeh hominids. In anyevent, given the sedimentary context, theposition, completeness and state of preser-vation of most of the specimens fromCouche XVII and XXII, natural processescannot be ruled out as an explanation for thehominid fossils at Qafzeh.

Figure 3. Saint-Césaire. Schematic stratigraphic profileillustrating granulometry. Hominid remains found atbase of stratum EJOP

superior (after Miskovsky &Lévêque, 1993:10, Figure 2.1). Reproduced withpermission from Context of a Late Neandertal: Impli-cations of Multidisciplinary Research for the Transition toUpper Paleolithic Adaptation at Saint-Cesaire, Charente-Maritime, France (1993) Madison, WI: PrehistoryPress. Copyright retained by F. Leveque.

Saint-Césaire (La Roche à Pierrot)

The disarticulated and highly fragmentedSaint-Césaire Neandertal specimen was dis-covered in 1979, in La Roche à Pierrot,a small rockshelter near Saint-Césaire(Charente-Maritime), France (Lévêque &Vandermeersch, 1980). It is a most import-ant find because burned flint artefacts fromthe same sedimentary context have yielded amean thermoluminescence date of about36,300&2700 BP (Mercier et al., 1991),placing it in Europe several thousand years

after the earliest dates for stone tools andother traces thought to be contemporaneouswith behaviourally modern hominids (e.g.,ApSimon, 1980; Melars, 1989). Thecrushed, partial and fragmentary adult skel-eton (>500 pieces) was encased in the upperportion of level EJOP (Miskovsky & Lévêque,1993) (Figure 3), a sedimentary unit con-taining stone and other artefacts ascribedto the Castelperronien industry. Suchassemblages are thought by many to repre-sent a ‘‘hybrid’’ of Middle and Upper Paleo-lithic industries, and are thus crucial toarguments about the cognitive capacities of

56 . .

the last Neandertals in Europe. Miskovsky &Lévêque (1993:10) define EJOP

superior bythe presence of ‘‘numerous angular pieces oflimestone’’in the upper ca. 10 cm of EJOP,which is an otherwise uniform matrix offines that are the result of bedrock dissol-ution filling in an earlier rockfall (Backer,1993:105).

The unusually numerous limestonepebbles and cobbles overlying the hominidskeleton are visible in the published photo-graphs (e.g., Miskovsky & Lévêque,1993:Figure 2.2), and are indeed suggestiveof a different sedimentary regime, and(at a minimum) a destructive depositionalenvironment. The 70-cm-diameter concen-tration of skeletal remains that comprise thepresumptive Neandertal burial occurs about1 m away from the wall of the shelter (Fig-ure 4, top left) (Backer, 1993:106, Figure9.1; Vandermeersch, 1993:130), at the baseof what the excavators refer to as ‘‘RockfallB.’’ Judged too friable to be excavated in thefield, the remains were removed en bloc, toenable careful study in the laboratory.There, the remains were found to be verti-cally compressed into a ‘‘few’’ centimetres,and all but the feet were representedby at least some identifiable fragments(Vandermeersch, 1993:129).

In spite of what would appear to begood evidence for natural deposition—in the form of sediments described as‘‘rockfall’’—in the following excerpt,Vandermeersch (1993:130) argues explicitlythat the Saint-Césaire specimen was apurposeful burial:

1. We observed no differences between theskeletal deposition area and the other partsof the layer. There was no apparent pit.2. The archaeological layer contained manylimestone blocks, but these blocks wereabsent from the skeletal area. They were veryclose to the skeleton, but not between thehuman bones.3. There was also a very low frequency oflithic artifacts directly associated with theskeletal block.

4. All the human bones were found within asmall area, almost circular, measuring about70 cm in diameter.5. All the bones were at the same level andthe human bone deposit was only a fewcentimetres thick.6. Some elements, but not all, were articu-lated. For instance, the maxillary and themandible were almost in occlusion andarticulated. Some fragments of the tibia andfibula diaphyses were side by side. Many ofthe hand bones were only slightly disturbedfrom their relative anatomical position.Unfortunately, we have not been able to fullyreconstitute the skeleton in place; too manybones are fragmented and the fragmentsdisplaced.7. By comparison we must remark thatno articulated animal skeleton or even anypart of any animal skeleton in anatomicalposition has been found in this layer.8. Finally, no cut marks have been observedso far on the preserved human bones.

If we consider all these arguments it seemsto me that they support the burial hypoth-esis. If it was not a burial it would be the firsttime, to my knowledge that fragments ofalmost all parts of a skeleton would havebeen found together in a Paleolithic site,outside of a burial context. But if we acceptthe burial hypothesis, another questionarises: it is indeed really difficult to under-stand how to fit a complete body in so smalla space. Therefore if we accept this hypoth-esis we should be led to consider anotherone: the possibility of a secondary burial.

These arguments for purposeful burialdeserve careful unpacking, because althoughthe points must be taken as containing fac-tual statements about the disposition of theremains, they nevertheless involve less thanparsimonious inferences.

Protection of the corpseThere was no discernible new stratum sup-porting the inference of purposeful protec-tion of a corpse—i.e., there was no grave.The absence of unequivocal evidence forpurposeful burial forces one to look furtherinto what can be inferred from other aspectsof the Saint-Césaire specimen’s conditionand the depositional context.

57

Figure 4. Saint-Césaire. Plan of excavation. Top left: hominid remains were recovered from squaresE4–F4 (after Backer, 1993:106, Figure 9.1). Top right: distribution and density of unretouched flakes¦2 cm (after Backer, 1993:109, Figure 9.3). Bottom left: distribution of naturally modified pieces (afterBacker, 1993:117, Figure 9.6). Bottom right: small flake surplus (after Backer, 1993:119, Figure 9.8).Reproduced with permission from Context of a Late Neandertal: Implications of Multidisciplinary Research forthe Transition to Upper Paleolithic Adaptation at Saint-Cesaire, Charente-Maritime, France (1993) Madison,WI: Prehistory Press. Copyright retained by F. Leveque.

Condition of the skeletonThe Saint-Césaire skeleton is incomplete(missing all the bones of the feet), highly

fragmented, and only some portions of theskeleton were found articulated. Moreover itwas found to be vertically compressed into a

58 . .

few centimetres of the deposit. This offerslittle support for the contention that theremains represent a primary interment, andthere is good reason to doubt that it isevidence for secondary burial.

If purposefully buried while the soft tissuestill held the joints in connection, one wouldexpect that an individual would be preservedwith most of the same joints in anatomicalposition once the soft tissue had decom-posed, no matter what insult befell thedeposits postburial. This, after all, is themajor supposition of the advocates of MPburial. If the Saint-Césaire individual hadbeen purposefully buried as an intact corpse,and postdepositional disturbance had notremoved the foot bones, it seems unlikelythat any amount of postdepositional sedi-ment compaction would have disarticulatedthe skeleton to such a degree that only afew parts were more or less where theywould have been in a living individual.Moreover, if the individual had been caughtin a single rockfall event it would probablyhave been protected from disarticulation aslong as it stayed buried. Saint-Césaire 1,however, is a disarticulated skeleton. Thisargues equally against burial in a single,catastrophic rockfall, and purposeful pri-mary interment.

However, while primary interment seemsunlikely, the hypothesis of secondary burialis only very weakly supported by the evi-dence. Secondary interments most oftenoccur once the soft tissue has decomposed,and would thus preclude preservation ofarticulated body segments. However, onecannot rule out secondary interment ofarticulated portions of a corpse not fullyskeletonized. Such a scenario could explainthe absence of foot bones, and the presenceof only a few articulated elements. Thehypothesis of secondary burial might explainthe presence or absence of certain parts, andthe differential articulation in evidence.Nevertheless the secondary burial of orig-inally intact skeletal elements might have

resulted in better preserved, intact elements,and might not, therefore, best explain thedegree of fragmentation or vertical compres-sion in evidence at Saint-Césaire. For sec-ondary burial to explain the disarticulated,fragmented, and vertically compressedstate of the Saint-Césaire skeleton it musthave involved secondary burial of alreadyfragmented skeletal elements. This seemsimprobable.

Even if one were to accept the possibilityof secondary burial of an already fragmentedand incomplete skeleton it would still meanignoring all of the evidence for a destructivedepositional environment at La Roche àPierrot. By being asked to accept thehypothesis of secondary burial, one is beingasked, ultimately, to accept that the skeletalfragmentation and the high-energy depo-sitional environment are unrelated—i.e.,merely coincidental—phenomena. Such arequest stretches the limits of credulity,because gradual burial in the destructiveenvironment of Saint-Césaire could accountfor the fragmentation and vertical compres-sion of the remains. Gradual, natural burialof a skeletonizing corpse might tend to com-press the corpse initially, and subsequent orrepeated insult would more than likely leadto further vertical compression and fragmen-tation of the skeletal elements, while main-taining, to some degree, the anatomicallocations of individual elements. Prolongedexposure could also explain the absence ofbody parts, such as the feet and half ofthe cranium, either through scavenging, ortrampling.

The likelihood of disturbanceOther animal remains were not recovered insimilar circumstances at Saint-Césaire. Iwould expect that, if this was a place thatNeandertals habitually used to shelter, otheranimals may simply have avoided it, or havebeen absent from the area (Straus, 1982;Gamble, 1993:166). Moreover, given thevicissitudes of decomposition, and the

59

impossibility of knowing if an exposedcorpse would have been attractive to scav-engers, the necessity of disturbance bycarnivores or other occupants is obviated.Under the circumstances, the absence ofevidence for other animals in the deposits isat best weak support for an argument forpurposeful burial.

Position of the skeletonThe position of the corpse is the next evi-dence proposed to support the inference ofpurposeful burial. Vandermeersch assertsthat the Saint-Césaire specimen could nothave been contained within a 70-cm circleunder circumstances of natural burial. Yetthere is some empirical evidence to suggestthat a 70-cm circle is sufficient to accommo-date a hominid body. Newell’s (1984)observation of the protohistoric Inuit habi-tation catastrophically destroyed by pack iceclearly shows that a human being, surprisedin sleep, could be encompassed in an area of70 cm. In the case of the Saint-Césairespecimen, whose feet are missing, 70 cmwould have been commodious. Thus, thehorizontal extent of the space withinwhich the Saint-Césaire remains were foundis at best weak evidence for purposefulburial.

Characteristics of the sedimentsThe sedimentary context of the Saint-Césaire specimen offers very little supportto arguments for its purposeful burial—especially given the excavators’ habit ofreferring to the sediments as ‘‘rockfall.’’EJOP

superior contains many limestonecobbles, any one of which could have causedthe demise of a sleeping Neandertal. Therewere bedrock blocks ‘‘very close’’ to theskeleton, but none ‘‘between the humanbones’’ (Vandermeersch, 1993:130). This isto be expected, for two reasons. First, theskeleton is described as lying at the lowerextreme of EJOP

superior—in other words atthe bottom of the layer of ‘‘rockfall.’’ Most

of the rockfall derives from the wall of theshelter, and was deposited during aninferred cold phase (Miskovsky & Lévêque,1993:14). Because the remains were lying atthe base of a layer of large rocks, one wouldnaturally expect to find no great amountof rubble beneath them. Second, giventhe variability inherent in the spatial distri-bution and particle size associated with evensmall bedrock collapse events, it is quiteexpectable that the limestone blocks ofinterest to Vandermeersch might have been‘‘very close to the skeleton, but not betweenthe human bones’’ (Vandermeersch, 1993:130). A plan of the site (Backer, 1993:106,Figure 9.1) clearly illustrates a non-uniformdistribution of limestone clasts. In sum, theabsence of large rubble in the 70-cm circlecontaining the skeletal remains cannot standas evidence for burial.

Nor do the allochthonous sediments pro-vide much support for the hypothesis ofpurposeful burial. The few vertical centi-metres containing the skeletal remainsyielded ‘‘a very low frequency of lithic arti-facts’’ (Vandermeersch, 1993:130). Lowerfrequencies of sharp stones in the vicinity ofthe remains is just what one would expect ifthe individual had died in its sleep and waslater buried in the shelter’s destructive depo-sitional environment. Any stone tools thataccumulated above the remains wouldhave been removed in the process of exca-vation prior to the skeleton’s discovery,and would not therefore have been presentin the block containing the remains thatwas removed to the laboratory. That theindividual may have been resting on lowfrequencies of sharp pieces of stone alsostands to reason, if one supposes that theindividual had been sleeping when theshelter wall or brow collapsed—you wouldexpect it to have cleared a comfortable areafor rest, which would presumably haveincluded removing sharp pieces of stone.Such behaviour does not require a modernhuman mind, or a culturally prescribed

60 . .

notion of cleanliness—finding and maintain-ing a comfortable resting place is a commonenough behaviour in the animal world.Therefore, the absence of lithics beneath theSaint-Césaire Neandertal might insteadbe construed as evidence supporting thesuggestion that the individual had beenresting in that place when it died, andwas subsequently buried by collapsingsediments.

Considerable additional support for theexistence of a destructive (and potentiallylethal) depositional environment in the areaof the skeleton comes from Backer’s (1993)examination of site structure at Saint-Césaire. In the vicinity of the Neandertalremains (Figure 4, top left) one finds thehighest frequencies of so-called naturallymodified lithic debris—i.e., more than 400pieces in the 1#1 m square containing theskeleton (Backer, 1993:117, reproducedhere as Figure 4, bottom left). Among otherthings, this observation contrasts withVandermeersch’s account of a low frequencyof lithic artefacts within the block containingthe skeletal remains (Vandermeersch,1993:130). Backer (1993:121) also reportsthat the debitage that would have resultedfrom natural breakage, i.e., the ‘‘small sur-plus flakes,’’ do not co-occur with the‘‘naturally’’ modified lithics. This ‘‘lendscredence to the possibility that the Saint-Césaire Neandertal was purposefully bur-ied’’ (Backer, 1993:121). Yet this analysis isnot supported by the data. These smallflakes are ‘‘expected to result from stonemodification activities, because tiny retouchflakes and debris are likely to remain embed-ded in the substrate once they fall to theground’’ (Backer, 1993:116). Again, it mustbe assumed that this distribution would haveincluded the unretouched flakes thatresulted from ‘‘natural modification.’’ The‘‘small-flake surplus’’ distribution is shownin Figure 4, bottom right (Backer,1993:119, Figure 9.8). It clearly shows thatthe densest concentration occurs in a square

adjacent to the one in which the hominidremains lay. In addition, far from being‘‘absent’’ from the square containing theremains, small-flake surplus occurs inrelatively high numbers.

In addition, one can suppose that thecategory of ‘‘unretouched flakes less than2 cm’’ (Backer, 1993:109, Figure 9.3, repro-duced here as Figure 4, top right), includedat least some amount of the debris of naturalmodification. A glance at the distributionof ‘‘unretouched flakes less than 2 cm’’reveals one of the densest concentrationsin square E4. Thus, in just that placewhere one finds the crushed and fragmentedhominid skeleton, one also finds the highestconcentrations of ‘‘naturally modified’’lithics, a high concentration of small,unmodified flakes that could have resultedfrom natural modification, and a relativelyhigh density of what is acknowledged to bethe debitage from natural modification.

Thus, rather than being an importantindicator of purposeful disposal of thedead, the distribution of lithic debris givesadditional support to the thesis that destruc-tive deposition was occurring in the vicinityof the skeleton—enough, that is, to haveproduced four and five times as many‘‘naturally modified pieces’’ in the squarecontaining the Neandertal remains as in 22of the 31 squares for which such data werereported. Finally, in accord with the patternof destructive deposition in the vicinity ofthe remains, the highest concentrations of‘‘whole’’ (i.e., undamaged) flakes occurseveral metres away.

SummaryThe excavators and analysts of La Roche àPierrot explicitly argue that purposefulburial preserved the hominid specimen.However, there are perfectly plausible natu-ral explanations for its preservation. First,there are numerous limestone cobbles inEJOP

superior. This accumulation of bedrockpieces is repeatedly termed a rockfall by the

61

multidisciplinary team that excavated it(Backer, 1993; Lévêque & Vandermeersch,1980). By any measure, it was a destructivedepositional environment, with crushedand broken skeletal material and brokenstone artefacts in abundance. No burial pitwas recognized, and no new stratumobserved. Tenuous arguments are offered toaccount for the state of preservation, thenear-total absence of articulation, thevertical compression and the relativelyconfined horizontal space in which thespecimen was found. Added to this,interpretation of the lithics inferred to becontemporaneous with the skeleton intro-duces some confusion, and ultimatelyundermines much of the burial argument. Inshort, purposeful burial seems the leastlikely explanation for the Saint-Césaireskeleton’s preservation. The Neandertalindividual lay in an area of destructive bed-rock breakdown, which not only producedthe highest concentrations of ‘‘naturallymodified lithics,’’ ‘‘unmodified flakes lessthan 2 cm,’’ and ‘‘small-flake surplus,’’ butalso the highest concentrations of naturallymodified Neandertal bones in the site. Withso much clear evidence for what Ferrand(1985:26) termed ‘‘crushed debris,’’ thereseems little support for the suggestionthat Saint-Césaire 1 was purposefullyburied. The individual could have diednaturally, and been buried in the destructivedepositional environment under circum-stances of delayed decomposition. Takentogether, the disarticulated, incomplete,fragmentary condition of the remains,coupled with the destructive depositionalenvironment in the shelter lead to theconclusion that natural deposition cannotbe ruled out as the explanation forSaint-Césaire 1’s preservation.

Kebara Cave

The Kebara 2 Neandertal specimen wasdiscovered in 1983 (Arensburg et al., 1985;

Bar-Yosef et al., 1986, 1988, 1992), duringthe extensive modern excavation of the cavenear Mount Carmel, in Israel. Kebara 2 is apartially disarticulated, partial skeleton. Itwas recovered in the deepest trench of theexcavation, 7·85 m below datum, in unit XII(Figure 5). Its archaeological context hasbeen dated by thermoluminescence to59,500&3500 years BP (Valladas et al.,1987, 1989). In all, the remains of 23 sub-adult and adult hominids have been recov-ered from Kebara Cave’s MP contexts. Onlytwo of these, a partial adult, Kebara 2, anda partial infant, Kebara 1, are claimed aspurposeful burials. Kebara 1’s excavationtook place earlier this century (Schick &Stekelis, 1977), and although it, too, isclaimed to have been buried (Smith &Arensburg, 1977), I do not deal with ithere because of inadequate documentation.Suffice it to say that there is nothing in thedescription of its discovery to warrantthe belief that it was purposefully buried.The adult, on the other hand, was excavatedunder technically masterful conditions, andwith the utmost care in recording anddescription. For this reason alone theKebara 2 specimen has led the majority ofpalaeoanthropologists to accept claims thatthis Neandertal was purposefully buried,and on that basis to perpetuate the beliefthat such behaviour was common in the MPand earlier. Close reading, however, revealsa number of unwarranted assumptionswhich weaken the arguments.

The remains are described as follows(Figure 6):

The skeleton was lying on its back . . . Theright hand lay on the thoracic cavity at thelevel of the left scapula. The left hand lay alittle lower at the level of the lumbar ver-tebrae. The cranium was missing, with theexception of the right, third upper molar.The cervical vertebrae were in anatomicalsequence, with the atlas positioned betweenthe branches of the mandible. The latter wastilted toward the vertebral column, indicat-ing that the head of the skeleton originally

62 . .

leaned forward. The head lay at a slightlyhigher level than the rest of thebody against the steep northeastern side ofthe burial pit. In its primary position thehead was probably facing westward.

The right humerus was turned inwardwith its lateral side facing up. The rightinnominate was in place, and the typicalsideward collapse after the decay of the fleshhad not occurred. These observations meanthat the right side of the body was leaningagainst the northern wall of the burial pit,limiting the amount of bone movementwhich normally occurs with the decompo-sition of the soft tissue. The exact position ofthe bones therefore furnishes the neededinformation concerning the existence ofthe northern and northwestern edges of thepit. The western side of the pit remainsunknown, as the right lower limb is missing.The left side of the skeleton has beenaffected by diagenetic processes whichcaused the alteration of the elbow, the pelvis,and the proximal part of the left femur.(Bar-Yosef et al., 1992:527–528)

The excavators thus spend considerableeffort describing the disposition of the skel-etal remains, as part of their argument thatthis was a burial. However, their interpret-

ation is not the only plausible one, and aparsimonious natural explanation appears tobe in evidence.

Figure 5. Kebara Cave. Stratigraphic profile showing position of Kebara Hominid 2 (Bar-Yosef et al.,1988:23, Figure 1). Note that the skeleton does not appear to lie on the unconformity that the excavatorsdescribe as a burial pit. Note also the presence of an earlier, similar feature, near the bottom. Reproducedwith permission from L’Homme de Neandertal: Actes du Colloque International de Liege (4–7 Decembre 1986),5 La Pensee (1988) Liege: Eutudes et Recherches Archeologique de l’Universite de Liege. Copyrightretained by the University of Liege.

Protection of the corpseThe partial skeleton was lying strati-graphically superior to an erosionalunconformity—a basin-shaped depression—that truncates previous deposits. Thedepression itself is stratigraphically superiorto other similar features, which suggest thatit is merely the latest product of a time-transgressive phenomenon (most likelya karstic process). The depression is20–25 cm deep, and is clearly distinguish-able in the underlying profile (Figure 5).The remains are repeatedly described ashaving rested in a ‘‘burial’’ pit. There aretwo reasons to doubt that the ‘‘pit’’ was aburial feature. First, to securely infer that apit was created for the purposes of buryinganything, it must be possible to discern anew stratum created at the time of burialdistinct from those into which the grave isdug and from those accumulating above it.

63

Figure 6. Kebara Cave. Plan of Kebara Hominid 2 (after Arensburg et al., 1985:229). Note the following:absence of cranium; disarticulation of the right humerus, scapula, and clavicle; fragmented anddisarticulated right elbow; disarticulated left hand; absence of the right lower limb. Reproduced withpermission from Comptes Rendus des seances de l’Academie des Sciences, Paris (1985) Series 2, 300, 227–230.Copyright retained by l’Academie des Sciences.

64 . .

In the case of Kebara 2, the sedimentssurrounding it were indistiguishablefrom those occurring immediately above(Bar-Yosef et al., 1988:19).

Second, if the unconformity were theresult of a grave, one would expect theskeleton to be lying on the contact withthe lower sediments. However, it is not clearfrom the profile, or from the excavators’description, if the skeleton was lying at thebase of the depression, or was depositedsometime after the depression began to fillin. Such ambiguity renders the claim forpurposeful burial less robust.

Further evidence that the unconformitymay not have been a prepared grave comesfrom the uncertainty regarding its extent.Bar-Yosef et al. (1992) were unable to locateits limit on the right side of the skeleton.Finding no clear stratigraphic evidence forthe extent of the erosional unconformity, theexcavators argue that the right innominate’sposition indicates the confines of the ‘‘pit.’’The right innominate was more or less inanatomical position, suggesting to the exca-vators that it had remained this way due tothe inferred presence of the (unrecogniz-able) boundary of the depression. In con-trast, the left innominate was disarticulated,and more or less horizontal. Explaining thedifferential articulation of right and leftinnominate is a relatively simple matter ofmodelling allochthonous deposition in thecave. The right innominate was on the sidenearest the cave’s entrance. Thus, it wasnearer the source of wind-blown and col-luvial sediments that form the majority ofthese deposits. During gradual burial of askeletonizing corpse, sediments would haveaccumulated against the right innominateon the side opposite the articulation with thesacrum—in other words it would haveformed a ‘‘sediment dam’’ behind whichsediments accumulated to hold it in place.The left innominate was furthest away fromthe source of sediments, and the rate ofsedimentation on the side opposite the

articulation with the sacrum would havebeen less. Thus, there would have beennothing to stop the left element frombecoming disarticulated. Here, then, is aplausible alternative to the excavators’ claimthat the right innominate was held in placeby the inferred pit boundary. The position ofthe right innominate is at best equivocalevidence for the ‘‘northern and north-western edges’’ (Bar-Yosef et al., 1992:528)of the presumed burial pit.

The lack of clear stratigraphic evidencefor the contemporaneity of the erosionalunconformity and the skeleton requirefurther examination and explanation. If anerosional unconformity of indeterminateextent was being filled naturally before theindividual was deposited, one would expectthat the horizontal extent of the originalfeature might be some distance from theremains, and might therefore have eludedexcavators. A feature such as the erosionalunconformity in evidence beneath Kebara2 is the sort of place where sedimentsnaturally accumulate at a more rapid ratethan elsewhere in the cave. This wouldhelp explain how a skeletonizing corpsecould have been preserved with portionsintact.

Finally, there is evidence in the publishedprofile that several similar sized depressionsinferior to the skeleton had developed in thesame place in the cave, and been filled in.This raises the possibility that the morerecent depression, the one containing thehominid remains, was created naturally.Thus, while it is reasonable to entertain thepossibility that the depression was purpose-fully created as a burial pit, it is not the onlyinference that one can draw from the obser-vations, and the possibility that it wascreated naturally, and filled in gradually,cannot be ruled out.

Condition of the remainsKebara 2 is an incomplete skeleton, missingall of the cranium except the right upper

65

third molar, most of the left lower limb, andall of the right lower limb. If only some partof a vertebrate skeleton is preserved, andpurposeful burial is inferred, explanationmust be provided as to why the remainderperished. The excavators suggest the auxil-iary hypothesis that the cranium’s absencewas the result of a deliberate act of post-interment ritual. The left lower limb’sabsence inferior to mid-femur is argued tohave been dissolved by diagenesis (Weiner &Goldberg, 1990; Bar-Yosef et al., 1992;Weiner et al., 1993, 1995). That same pro-cess is cited to explain the left olecranon andleft innominate’s poor preservation. Thisprocess is well documented, and adequatelyaccounts for the degradation and absence ofparts of the skeleton’s left side. However,diagenesis cannot adequately explain theright lower limb’s absence, nor that ofthe cranium. The excavators do mentionthe missing right lower limb, yet offer noexplanation for its absence. Nor do theymake a case for why they think it couldhave no bearing on the question of pur-poseful burial. Diagenesis cannot explainthe right lower limb’s absence, unless theprocess affected everything up to, but notincluding the acetabulum of the rightinnominate, a fairly unlikely possibility.Some actor or process must have displacedboth the cranium and the right lower limbafter the soft tissue decomposed. It is poss-ible to use knowledge of decompositionand disarticulation to construct a parsimo-nious natural explanation for the conditionof the remains, one that supports an infer-ence that the skeleton underwent a processof gradual burial.

Decomposition and disarticulationBar-Yosef et al. (1992) argue that theskeleton’s articulated state stands as the bestevidence for purposeful burial.

Most of the anatomical connections werestill intact; for example, the extremely

mobile hyoid bone remained in placebetween branches of the mandible. Therewas no evidence for the collapse of thethoracic cavity after decomposition of thesoft tissue. The position of the componentsof the scapular girdle and the obliquity ofthe left clavicle indicate that the shoulderswere contracted slightly upwards. Theseobservations suggest that the body decom-posed in a filled grave and that the burialpit was somewhat deeper at the level of thethorax. Despite the abundance of evidencefor hyena-gnawed bones in the cave, nocarnivore marks were noticed on thesebones. The position of the upper limbs,especially the right hand reaching theleft shoulder, supports the hypothesis ofimmediate inhumation, perhaps preced-ing rigor mortis. The positions of themandible, the hyoid bone, and the rightupper third molar, which fell from itssocket next to the right lower third molar,exclude the hypothesis that the skull wasremoved by an animal. Furthermore, nocranial fragments were found. These obser-vations suggest that the skull was removedby humans following the complete decay ofthe atlanto-occipital ligaments. This is, inour view, the first clear-cut case recordedin a Mousterian context for later humanintervention in a primary burial. Theabsence of the skull precludes any fur-ther interpretations in terms of mortuarypractices.(Bar-Yosef et al., 1992:528–529)

In the preceding excerpt, the linksbetween observation and conclusion are notalways explicit. However, one is able to infertheir lines of reasoning from the way inwhich the evidence is presented and theconclusions they reach. Their argumentshinge on expectations of what should hap-pen when a hominid skeleton decomposes.By describing the remains in this way theexcavators are clearly implying that, if leftunburied (1) the hyoid should have beenmobilized, (2) the ribs should have comeapart from the backbone prior to theskeleton’s thorax collapsing, (3) the scapulaand clavicle should have been in anatomicalposition, (4) the bones should have beengnawed by hyenas, (5) the right hand could

66 . .

not have been where it was, (6) the man-dible, hyoid and upper third molar could nothave been preserved where they were ifan animal had removed the cranium. Sincethe authors depend on expectations ofwhat should have happened, one wouldhave expected some reference to actualisticstudies underpinning their arguments.Studies of natural decomposition, and dis-articulation, though rare, do provide back-ground knowledge that renders most of theirexpectations open to question. It is possibleto evaluate the excavator’s assertionswith reference to eight aspects of naturaldisarticulation.

First, quadruped forelimbs are the firstbody portion to become disarticulated asdecomposition proceeds (Hill, 1979;Haglund, 1997), because there is no liga-mentous attachment between the scapulaand the rest of the skeleton. In hominids,however, the clavicle would make the pro-cess of forelimb disarticulation more com-plex. Thus, in a gradually decomposingskeleton, the articulation between theacromion and clavicle can plausibly explainthe position of the scapular girdle and clav-icle, which was taken to indicate that ‘‘theshoulders were contracted slightly upwards’’(Bar-Yosef et al., 1992:528). As the scapulacomes free of the ribcage, it would form acontinuous skeletal element ‘‘chain’’ withthe clavicle and humerus, which, beinganchored at the sternum, would mean thatthe whole structure would have ‘‘slid’’ dor-sally and cephalically off the ribcage. Thus,rather than acting as evidence for immediateburial, the position of the left upper limband scapular girdle can plausibly stand asevidence for this individual’s having incom-pletely decomposed while being naturallyburied.

Second, Bar-Yosef et al. (1992) makemuch of the preserved hyoid and the absentcranium, asserting that the cranium hadbeen carefully removed in a postintermentritual. However, they are somewhat vague

when it comes to their reasons for conclud-ing that animal disturbance cannot beinvolved to explain the absent cranium, andthat ‘‘human’’ intervention was responsible.They base their conclusion on theposition of the mandible, hyoid and upperthird molar. Presumably the authors aresuggesting that an animal bent on removingthe cranium would not have been able toavoid disturbing the mandible, hyoid andupper third molar, with the clear implicationthat a Neandertal would have been able to.That seems unlikely. Whether it was a homi-nid with the intention of removing the cra-nium, an animal intent on scavenging thehead, or a less intentional act such as stub-bing a toe, the mandible, hyoid, and toothwould have been left behind unintentionally,because the soft tissue connecting them tothe cranium had previously decomposed.The mandible is freed of its association withthe cranium early in the sequence of dis-articulation (Hill, 1979). The hyoid de-velops in soft tissue inferior to the mandible.It should come as no surprise that the man-dible and the anatomically associated hyoidremained, while the cranium was displaced,without demonstrably affecting the positionof either the mandible or the hyoid. For thehyoid to have been discovered between the‘‘branches’’ of the mandible, the mandiblemust have first dropped forward on to thethorax. This is to be expected, given that theindividual was lying on its back.

Third, the atlas is also described as beingbetween the branches of the mandible, andit was in fact displaced slightly anteriorly(Tillier et al., 1991:92). The atlas wouldhave come free of the rest of the vertebralcolumn prior to separation from the cra-nium (Hill, 1979). Thus, the true anatomi-cal location of the atlas no more explains itsfinal physical relation to the other elementsof the preserved Kebara 2 skeleton thandoes that of the hyoid—neither are articu-lated with the mandible, but both are whereone might expect them to be after a gradual

67

process of natural disarticulation followedby displacement of the cranium. The cra-nium must have been removed once theatlanto-occipital ligament was decomposed,because the atlas did not go with it. Thus,the most parsimonious explanation for thehyoid’s position is that it was bound up inthe soft tissue associated with one of the firstelements to become disarticulated, in anatural sequence of decomposition in theopen air. In a depression that was collectingsediments, one would expect that the man-dible and hyoid, once having dropped on tothe thorax, would have been buried some-what earlier than the rather more prominentNeandertal cranium. The cranium wouldthus have remained exposed for muchlonger.

Fourth, the Kebara 2 right humerus isdescribed as having been ‘‘turned inwardwith its lateral side facing up.’’ This is to beexpected if natural disarticulation hadoccurred prior to natural burial, because thehumerus also becomes free early in thesequence of disarticulation (Haglund, 1997;Hill, 1979). Although the excavators men-tion that the right humerus was turnedmedially, they do not mention that it wasalso displaced at least 5 cm from the glenoidfossa of the scapula. This can be seen clearlyin the plan of the hominid remains in situ(Figure 6). As with the missing lower rightlimb, Kebara 2’s excavators offer no expla-nation for the disarticulated right humerus,nor do they explain why they think it isunimportant. However, it is, once again,potential evidence that parts of this skeletondecomposed and disarticulated prior toburial. Displacement of the right humerusoccurred only once decomposition hadadvanced to the point where it became freeof the scapula. Such displacement, I wouldargue, was unlikely to have occurred afterpurposeful burial. Whereas smaller ele-ments, such as phalanges, can be removedby bioturbation, displacement of a majorlong bone is much less likely once it

has been buried under any quantity ofsediments—it would probably have requiredsubstantial disturbance to have moved itseveral centimetres sideways. Any distur-bance process capable of such effects wouldprobably have disturbed neighbouring ele-ments. Displacement of the right humerusdoes appear to have been associated withother disturbance, notably that which frag-mented the individual’s right elbow region.Judging from the nature of that breakage itcould only have occurred once the boneswere dry, and it seems improbable that suchdamage could have been inflicted after pur-poseful burial. However, if a heavy visitortrod on this place while the specimen was inthe process of being naturally buried, suchdamage might be expectable. There is still aslight chance, however, that such damageoccurred after purposeful burial, but onlyif the individual had been buried in anextremely shallow grave.

Fifth, some of Kebara 2’s condition isperfectly expectable, whether it had beenpurposefully buried or not. The right handbones are more or less articulated, whilethose of the left are scattered about. Theright hand bones were supported on the leftribs, while the left hand decomposed overthe fleshy abdomen, resulting in the lefthand bones scattering as the soft tissuedecomposed.

Sixth, although some of Kebara 2’s dis-articulation would have been possible had itbeen purposefully buried, the cranium couldhave been displaced in only one of two ways.Either it was purposefully buried and thenexhumed, as the excavators contend, or itwas naturally displaced in the course ofdecomposition and natural burial. It ishighly unlikely that the cranium’s absence isdue to scavenging. By the time the atlanto-occipital connection is severed, decompo-sition is well advanced, and it is unlikely thatthe remains would have been attractive.Once the soft tissue decomposed, and theconnection between atlas and occipital were

68 . .

severed, the cranium was free to roll or bekicked away. Crania are among the mostmobile of hominid skeletal elements,especially those of the more recent,encephalized species (see e.g., Voorhies,1969). Exposure during decompositionwould parsimoniously explain the presenceof the third upper molar with the Kebara 2remains, even though the cranium isabsent. For this tooth to have come freeof the cranium, considerable time musthave elapsed before the cranium wasdisplaced—a minimum of a year in a dryenvironment, and probably longer if it didin fact mummify (McKeown & Bennett,1995). While purposeful disinterment of thecranium cannot be ruled out, discoveries ofisolated crania in many other MP contextssuggest that, at least at times, crania do getaway from the rest of their body-mates.Finally, unless one is prepared to view thelower right limb’s absence as evidence of asimilar ritual, nothing should persuade us tothink that the cranium’s absence was any-thing other than an expectable taphonomicprocess. The damage to Kebara 2’s rightelbow region, disarticulation of the righthumerus, and ‘‘removal’’ of the craniumcould easily have occurred in a single distur-bance episode, during natural burial, andneed not be the result of postinterment ritual.

Seventh, last in Hill’s sequence of dis-articulation are the ribs and vertebrae (cf.Haglund, 1997). The joints of the spine andribs are the last to decompose, because theyare tightly bound together with a densefabric of ligamentous soft tissue. Bar-Yosefet al. (1992) say this of the ribcage: ‘‘Therewas no evidence for the collapse of thethoracic cavity after decomposition ofthe soft tissue.’’ I take this to mean that theribcage must have collapsed prior to decom-position of the soft tissue binding it to thespine. It is by no means clear why this standsas evidence for purposeful burial, and notgradual natural burial. Because the individ-ual was lying on its back, the backbone and

rib articulations would have lasted intrinsi-cally longest and been buried earliest, withnatural burial prolonging soft tissue decom-position (cf. Toots, 1965), leaving theshaft of the ribs to collapse without theproximal portions being displaced from theiranatomical positions. By the time the softconnections were broken down, the ribs hadalready succumbed to breakage. That softtissue connections can survive buried for along time is exemplified by the naturalmummies of the American Southwest. Inextremely arid conditions mummifiedremains can survive with soft tissue adheringfor thousands of years. Preserved skin andhair of the Spirit Cave mummy fromNevada, for example, are approximately9415 radiocarbon years old (Tuohy &Dansie, 1996; Kirner et al., 1997). And,while Kebara Cave was by no means as aridas the North American Great Basin whenthe hominid died, neither did the depo-sitional circumstances require Kebara 2’ssoft tissue to preserve intact for 9000 yearsbefore it was naturally buried. Thus, toargue that the condition of the ribs is evi-dence for purposeful burial is to ignore goodpresent-day models for how temperatureand humidity can combine to retarddecomposition and disarticulation.

Lastly, Bar-Yosef et al. (1992) use theposition of the right hand to suggest thatburial occurred prior to decomposition. Pre-sumably they are arguing that bloating,which often accompanies decomposition,would have straightened the limb. But thisignores the variability inherent in the processof decomposition. Under cool conditionssuch bloating will not occur (Micozzi,1997). In addition, cool conditions retarddecomposition, allowing desiccation to pre-vail, which then ensures that the skeletonwill stay intact for longer.

Characteristics of the sedimentsThe sediments that accumulated around theKebara 2 specimen suggest that relatively

69

gradual sedimentation was occurring. Dur-ing the Mousterian occupation of Kebarathe predominant sediments are colluvial clayand aeolian sands and silts, with theaddition of materials clearly of anthro-pogenic origin (i.e., lithic debris, faunalbone, and combusted plant fuel, Bar-Yosefet al., 1992). Kebara’s stratification presentsa complex picture of allochthonous andautochthonous natural and anthropogenicsedimentation, postdepositional chemicalalteration, karstic processes including sink-hole development and subsequent slumping,and significant bioturbation. Given the com-plexity of the cave’s depositional history, andthe erosional unconformity near which theremains were found, it is not possible to saywith any certainty how rapidly the individualwould have been covered. On the evidenceof particle size and origin, relatively slowdeposition was occurring in the cave at thetime of Kebara 2’s deposition. However,because the remains came to rest wheredeposition was relatively more rapid—i.e., ina localized sediment trap—they would havebeen more rapidly encased in sediments thanif they had been exposed on a plane surface.

SummaryThe excavators of Kebara 2 claim that apurposeful burial took place nearly 60,000years ago in Kebara Cave (Bar-Yosef et al.,1992). They argue for this based on thedegree of articulation and the presence of astratigraphically inferior erosional uncon-formity. They document a partially com-plete, partially articulated skeleton, thecondition and stratigraphic context of whichis amenable to explanation by other thanpurposeful burial. When first reporting thediscovery, the excavators suggested that thecorpse had been ‘‘protected’’ by wood,branches or skins until a process of skel-etonization had been completed, at whichtime the cranium was removed (and pre-sumably the right leg), and the remainderburied (Bar-Yosef et al., 1988:20; cf. Villa,

1989:325, who also concludes that longexposure of Kebara 2 is indicated). It is notclear why Bar-Yosef et al. changed theirminds four years later, and concluded thattheir ‘‘observations suggest that the bodydecomposed in a filled grave’’ (Bar-Yosefet al., 1992:528), and that it was probably an‘‘immediate inhumation’’ (Bar-Yosef et al.,1992:529). One cannot, of course, rule outthe possibility of Kebara having been apalaeo charnel-house, and that the remainswere purposefully protected from somedisturbance processes prior to purposefulburial. Neither can one rule out naturalprocesses. The evidence is consistent withslow decomposition and natural burial.

Amud Cave

The original excavations of Amud Cave,Israel, recovered evidence for fourNeandertal individuals. I previously argued(Gargett, 1989a) that Amud 1 (Suzuki &Takai, 1970) was recovered in unstratifiedsediments, near the wall of the cave. There-fore nothing about its discovery suggeststhat purposeful burial occurred, other thanits degree of articulation, which I do notregard as evidence of anything other thanspecial depositional circumstances. I will nottherefore deal with the results of Suzuki andTakai’s excavations again. Amud has nowyielded three more hominids. One of thesewas Amud 7, a 10-month-old Neandertal(Rak et al., 1994), discovered during the1991–1992 field season. Amud 7 is a frag-mentary, partial skeleton, with some ele-ments still articulated. It is claimed to havebeen purposefully buried (Hovers et al.,1995).

Amud 7 is described as follows:

The specimen is an articulated skeleton of a10-month-old infant, which was found lyingon its right side directly on the bedrock in asmall niche in the cave’s north wall . . . Thecranium is collapsed, and only the occipital,parietal, and temporal bones are present. . . .

70 . .

the face is badly damaged and cannot bereconstructed. Several upper teeth wereretrieved when the sediment around theskeleton was sieved. The mandible, on theother hand, is relatively complete. Whereasthe vertebral column and ribs are fairly wellpreserved, the long bones, the pelvis, thescapula, and the sternum are incomplete.The phalanges of the hands and feetwere discovered in their proper anatomicalposition.(Rak et al., 1994:314)

The authors argue that purposeful burial isindicated.

In the highly dynamic environments of theLevantine caves during Mousterian times,hominid occupation commonly alternatedwith the activities of other animals, and theresidues of both were often subjected tosevere disturbance prior to further sedimentdeposition. Under such circumstances, thearticulation of Middle Paleolithic hominidskeletons is the major criterion for theirdesignation as intentional burials (Belfer-Cohen & Hovers, 1992). In the case ofAmud 7, this claim is enhanced by thediscovery of a red deer (Cervus elephus) max-illa leaning against the pelvis of the buriedhominid.(Rak et al., 1994:314)

The authors assume that the depositionalenvironment at Amud would have meantdisarticulation of an exposed carcass. Yetthere is no good way of assessing the priorprobability of disturbance, and it cannotsimply be assumed. In spite of the exca-vators’ assertions, the location and dispo-sition of the Amud 7 skeleton provides cluesas to why this individual was preservedbetter than some of the others recovered atthe same site.

Protection of the corpseThe excavators found no burial pit or newstratum created when the corpse wasdeposited. There is thus no stratigraphicevidence of purposeful burial.

Grave offeringsAlthough the excavators assert that the deermaxilla resting on the pelvis of Amud 7 is a‘‘burial gift’’ (Hovers et al., 1995:52), thereis no reason to accept the assertion. First,there is no clear evidence that this individualwas intentionally buried. There is no evi-dence that such a discovery is unique or rarein the excavated sample. There is nothing tosuggest that its preservation in an out-of-the-way place would be unexpected. There-fore, it is not logical to argue that the skeletalremains of one animal in proximity to those ofa hominid are contemporaneous, much lessa ‘‘burial gift’’ from a bereaved conspecific.

CompletenessA complete, articulated skeleton offers goodevidence of a corpse protected from distur-bance. However Amud 7 was not insulatedfrom disturbance. It is incomplete. It lacksthe frontal and facial bones. Moreover, thelong bones, pelvis, scapula and sternumhave portions missing. Given that the speci-men is incomplete, one has to ask whatbecame of the missing element and portionsof elements. The excavators are not forth-coming with models to explain the missingparts, although they are quick to point outthat carnivore disturbance is not indicated(Hovers et al., 1995:52).

Decomposition and disarticulationHovers et al. (1995) and Rak et al. (1994)assert that articulation is good evidence forpurposeful burial. However, parts of Amud7 appear not to be articulated. The pub-lished stereo photograph (Rak et al.,1994:316, Figure 2) clearly shows the stateof the cranial elements in situ. First, thecranium is not, strictly speaking, articulated.Instead, it appears to be lying on the occipi-tal condyles on the more or less horizontalbedrock. In other words, it has been rotatedthrough 90) in the coronal plane. If indeedthis individual had been placed on its rightside and then buried, as the excavators

71

assert, its head should not be where it wasfound. That is, of course, unless the headwas purposely twisted at right angles to thebody while being prepared for burial. Whileit is impossible to rule out such an arcaneburial scenario, based on the condition ofthe skeleton itself, the simpler explanation—that it was buried gradually in an out-of-the-way place—seems equally likely.

In addition to the evidence for extremeneck flexion, Amud 7’s unfused cranial ele-ments are not in any sense articulated. Theoccipital bone forms a bowl in which aredeposited the left calotte elements. Theleft parietal and temporal, having beenseparated from the rest of the cranium, areresting almost vertically near the midline ofthe basicranium and occipital. The righttemporal and parietal somehow escaped thelateral displacement experienced by the left,and instead, the fragments of the right tem-poral and parietal are lying near the rightmargin of the basicranial fragment, nestedlike fragments of a broken bowl. If thisindividual had been purposefully buried,one would have to postulate considerablemicroscale variability in the disturbance pro-cess, and differential distortion of the gravefill in the area of the cranium, to account forthe different ways in which the left and rightside calotte elements are positioned. Indeed,one would have to argue that two verydifferent, dynamic subsurface disturbanceprocesses had been at work within a scantfew centimetres of one another to explainhow the right cranial bones collapsed insideone another while the left ones remainedmore or less in their anatomical orientation,while being displaced several centimetresmedially.

It is not necessary to postulate any highlyunlikely permutations of postdepositionaldisturbance to explain the preservation andthe location of this individual’s skeletalelements—the cranial bones’ orientationand fragmentation suggests simply that thebreakage took place while the cranium was

exposed. This is supported by the orien-tation of the cranium and mandible inrelation to the rest of the skeleton. Theunarticulated state of the cranium arguesstrongly for exposure prior to burial. Itappears as if the head tipped on to itsbasicranium as the individual decomposed.There followed a gradual process of naturalburial, which would have exposed theprominent cranium and other elements todamage.

Depositional circumstancesAmud 7’s location in a natural niche in thecave wall would have given it a better thanaverage chance at preservation—if onlybecause trampling would have been virtuallyruled out. In addition, the specimen’s occur-rence in a shallow cave, in a relatively aridarea, means that its small body would haveundergone rapid desiccation. Carnivoreswould then have been uninterested (Gifford,1981). With respect to Amud 7’s location ina natural niche, it would be impossible torule out purposeful behaviour, regardless ofwhether or not it had been purposefullyburied. On the other hand, it would beequally difficult to rule out the possibilitythat this was a chance combination of cir-cumstances resulting in the specimen’sexcellent preservation.

SummaryAmud 7’s excavators found no burial pit ornew stratum created at the time of theskeleton’s deposition. Taking into accountthe skeleton’s location in a natural niche, theposition of the various cranial bones, and theposition of the skull and mandible, onecannot rule out the possibility that it wasnaturally buried.

Dederiyeh Cave

In 1993 a joint Syrian and Japanese exped-ition recovered the partially articulated, par-tial skeleton of a two-year-old Neandertal in

72 . .

Dederiyeh Cave, about 400 km north ofDamascus, northern Syria (Akazawa et al.,1995a,b). The infant’s remains were foundnear the back of the cave, about 1·5 m belowthe surface, ‘‘lying on its back with armsextended and legs flexed’’ (Akazawa et al.,1995b:79) in what is reported as a burial pit.Infant skeletons are known to be extremelysusceptible to destructive processes, andpurposeful burial would be one possibleexplanation for the presence of well-preserved immature remains. At presentthere is very little published on the discov-ery, which makes critical appraisal of thedepositional circumstances difficult. How-ever, scholarly perception of the rarity ofinfant remains, and the care with which theywere recovered would be enough to con-vince most palaeoanthropologists that theyhad been purposefully buried. However, inAkazawa et al.’s (1995b:79) account thereare several observations on the depositionalcircumstances that suggest an alternativeexplanation.

Characteristics of the sedimentsExamination of the cross-section ofDederiyeh Cave (Figure 7, top) reveals a10 m by 15 m chimney at the back of thecave, open to the plateau above, whichappears to be the apex of an enormous vaultdeveloping in the cave’s ceiling. Beneath thischimney is an equally enormous talus conethat clearly resulted from ceiling breakdown,and must at times have included materialfalling or being flushed down from theplateau above. Beneath this chimney iswhere the infant’s skeleton was found, insome sort of local low spot, depression,or pit. Whether one assumes purelyautochthonous sedimentation prevailingat the time the infant was deposited, orsome combination of allochthonous andautochthonous sedimentation, I wouldexpect relatively rapid burial of anythingcoming to rest on the cone. Thus, relativelygood preservation of even an infant’s

skeleton should be the expectation, ratherthan the exception.

Protection of the corpseAkazawa et al. (1995b:79) refer to a burialpit, and observe that in ‘‘the mostly sterilefill of the burial pit, there was a sub-rectangular limestone slab above the headand a small triangular piece of flint wherethe infant’s heart was located.’’ However,Akazawa (1995) states that the pit does notconform to expectations for what a burial pitwould look like, and is therefore unable tosay whether or not it was dug for the pur-pose of burying the infant. As I suggestedabove, there are a number of ways that lowspots can form in caves, and any low spotwill tend to fill more or less rapidly. Thus,any vertebrate that found its way intosuch a depositional environment will have abetter-than-average chance of survival.

There may be a clue to the depositionalcircumstances of the infant in Akazawaet al.’s (1995b:79) description of the pit fillas ‘‘mostly sterile.’’ This implies either thatthe pit was dug into mostly sterile sedi-ments, and that therefore its fill containedlittle evidence for hominid activity, or thatthe pit fill contrasts with the surroundingsediments in not containing traces of homi-nid behaviour. The latter seems more likely,given the often-noted difficulty of recogniz-ing such features in cave sediments—if thefill contrasted with the sediments in which alow spot formed it would have made it easierto discern during excavation. And if thesurrounding sediments included moreevidence of hominid behaviour, this mightsuggest that the pit had been filled withmostly natural sediments in a process ofnatural burial. However, final interpretationmust await publication of the individual’sdepositional context. At present, there isan incompletely warranted claim that theinfant’s remains were purposefully pro-tected, and the remains themselves, whichare more or less well-preserved.

73

Figure 7. Dederiyeh Cave. Top: cross-section through cave and talus slope. Bottom: plan (after Akazawaet al., 1995b:78–79, Figures 2 and 3). Hominid infant remains are shown on plan by a dashed circle. Notethat the remains were recovered from the back of the cave, at the chimney’s base, near the talus cone’sapex. Reproduced with permission from Paleorient (1995) 21/2, 77–86. Copyright retained by C.N.R.S.,Paris.

Condition of preservation, decomposition anddisarticulationA plan drawing of the partially articulated,partial Dederiyeh infant in situ clearly dem-onstrates that a number of fragile elementswere preserved, albeit fractured and frag-mented (Figure 8). In general, immaturevertebrate remains are quite fragile. Ifdensity-mediated destruction were theexplanation for this skeleton’s condition,one would expect the denser elements to

have survived while the less dense, thinner,and smaller elements succumbed (Brain,1967, 1969; Guthrie, 1967; Lyman, 1984).However, the catalogue of surviving ele-ments reveals no clear relationship todensity. For example, most of the spongy,fragile, innominate elements survived. Manyof the facial bones did not. While notspongy, facial bones tend to be thin andtherefore relatively fragile. They might beexpected to be as susceptible to destruction

75

as the innominate. The left clavicle, leftischium, the distal left femur, proximal anddistal left fibula, and most of the small bonesof the hands and feet did not survive, whiletheir opposites did. The distal left tibia wasdamaged. Not so the right. Such differentialpreservation cannot be explained by differ-ences in bone density—some very fragilebones survived, while some robust ones didnot. Moreover, paired elements should bedeleted simultaneously if affected by thesame process. One must ask, for example,why all three right innominate elementswere preserved, while only the left ilium andpubis survived. Indeed, none of the innomi-nate parts should have survived given that aportion of the much denser left femur suc-cumbed. The right clavicle survived. Whynot the left? There does appear to have beensome process differentially affecting the ele-ments of the left side: i.e., the shoulder, hip,and lower limb. But what are the processesresponsible?

Whether the Dederiyeh infant had beennaturally or purposefully buried, it wouldbe impossible to rule out bioturbationas an explanation for the skeletal partrepresentation—these are relatively smallskeletal elements, and most would besusceptible to displacement by burrowingrodents or root activity. Indeed, on theplan drawing it appears as if a number ofelements have been separated from theiranatomical neighbours, including severalrib fragments that have been displaced by40–50 cm. Alternatively, if rockfall wasresponsible for this individual’s death, it ispossible that burial was only partial, and thatsome of the individual’s left side remainedexposed, allowing differential decompo-sition. The exposed parts would have been

more vulnerable to postdepositionaldestruction and displacement. On balance,it appears as if diagenesis of the sort inevidence at Kebara Cave can be ruled out,because some elements are absent fromareas in which others have survived—themissing clavicle, for example. Thus, whilethe evidence of the skeleton’s condition pre-cludes density-mediated destruction anddiagenesis, neither does it provide evidencefor purposeful burial.

Figure 8. Dederiyeh Cave. Infant remains in situ. Plan of excavated skeletal remains and interpretation oforiginal position (inset) (after Akazawa et al., 1995b:80, Figure 5). Note the contortion of the right wrist,and buckling of the right leg. Note also the displacement of rib fragments in lower portion of this view.Reproduced with permission from Paleorient (1995) 21/2, 77–86. Copyright retained by C.N.R.S., Paris.

Position of the corpseAlthough it is described as ‘‘flexed’’ the poseof this infant is anything but flexed in theway most would associate with flexed burial.In the excavators’ interpretation of its orig-inal position (upper left of Figure 8) theright wrist is extremely flexed, while theright elbow is almost fully extended. The leftwrist is extremely flexed, while the elbow isslightly flexed, about 45) (note that the plandrawing does not illustrate the left hand,while the excavators’ interpretation does).The right femur appears to be abducted 90)to the spinal column, and the right tibiais parallel to the femur. The left femur isabducted about 45), and the right knee isflexed about 120) (note that in the plandrawing, this angle is closer to 150)). Thus,in both the plan drawing and in theinterpretation the skeleton’s position looksmore like that which one would expect if theinfant had died in a fall, or had been caughtin a rockfall, and not what one might expectto see in a purposeful burial. The flexion ofthe right wrist and leg more resembles theawkward positions of limbs that have beenbuckled as an individual collapsed, or werefolded underneath by a fall. Given theposition of its limbs at the time of burial, it is

76 . .

difficult to ignore this infant’s location nearthe top of a cone of talus developing under avault, with a chimney open to the plateauabove. By itself, the skeleton’s position is atbest weak evidence for purposeful burial. Onthe other hand, it is better evidence of somesort of catastrophic event.

On the evidence of its location and itsposition, one can propose at least two alter-native hypotheses to explain this infant’sdeath and preservation: (1) rockfall, and (2)a fall down the chimney, followed by rela-tively rapid burial in falling sediments. Arockfall would not be out of the question,considering the way the cave has formed,and its location in a tectonically activeregion. One line of evidence in support ofthe rockfall scenario is reported by Akazawaet al. (1995b), when they state that mostlysterile sediments encased the infant. Adearth of artefactual material in the encasingsediments would be expected if the baby haddied in a small rockfall. Moreover, whetherthe infant was killed instantly, by rockfall, ordied after a fall down the chimney and wasthen buried by gradual, more or less rapidbreakdown, some quantity of sedimentswithout artefactual material should beexpected.

SummaryOn present evidence, the absence of a newstratum created at the time of burial, theinfant’s position, its location directlybeneath an enormous vault and chimney,and the depositional regime that could beexpected in a cave in a tectonically activearea, all combine to make it impossible torule out natural processes to explain thisNeandertal’s preservation.

Discussion

In this paper I have re-examined the contextof a number of MP hominid specimens.Based on present knowledge of the range ofnatural processes known to form archaeo-

logical sites, I have offered alternative expla-nations for MP hominid traces claimed tohave been purposefully buried. There isnothing unusual, or even especially creativein what I have done—subjecting hominidbehavioural inferences to examination froma taphonomic perspective has become acommonplace in archaeological inquiry.There are a number of good examples:Brain’s (1981) critique of Dart’s inference ofan australopithecine osteodontokeratic tech-nology at Makapansgat (Dart, 1957a,b),Stiner’s (1991) and White & Toth’s (1991)alternative to inferences of ritual cannibal-ism at Grotta Guattari (Monte Circeo)(Blanc, 1958), Koby’s (1941, 1953) andKurtén’s (1976) demystification of cavebear cults at the Drachenloch cave (Bächler,1921). In all these cases, what wereonce widely accepted theories for theway pre-modern hominids behaved havebeen replaced by more parsimoniousexplanations that rely on an empiricallyderived specification of the range of pro-cesses capable of forming such archaeologi-cal traces.

In drawing their inferences, excavators ofthe claimed MP burials employed a similarkind of reasoning to the one I employ. Themajor difference in our approaches, andwhat underlies the differences in our expla-nations for the hominid remains, is the set ofassumptions that underpin our interpretivemodels. Meticulous recovery techniques,exemplified by those of Kebara 2’s exca-vators, and prompt publication of sitedescriptions have enabled me to undertake ataphonomic re-analysis, drawing on knowl-edge of a wider range of taphonomic pro-cesses than have heretofore been employedin interpreting these discoveries.

In assessing my claims, the reader shouldremember that the archaeological recordcontains inherent checks against un-warranted assumptions, whether mine orthose of other workers. I have suggested thatthe archaeological record makes it difficult

77

to rule out natural burial of MP hominids.If, for example, my suggested alternativeswere implausible, or just false, the exca-vators’ observations could be used to coun-ter my assertions. For example, if Kebara 2was as fragmented and crushed as the Saint-Césaire specimen, I would be hard pressedto explain how it could have happened in alow-energy depositional environment, whereany vertebrate would probably have beengradually buried in fine sediments. If,indeed, the Kebara skeleton were to havebeen more like that of Saint-Césaire’s—i.e.,crushed, fragmented, disarticulated, andvertically compressed—my suggestion ofgradual, natural burial of Kebara 2 mightlegitimately be seen to stretch the limitsof plausibility. Likewise, if Qafzeh 9were partly disarticulated, and was missingportions of the skeleton, and did not bearevidence of having been crushed at or nearthe time of death, my suggestion that thisindividual was entombed in a rockfall wouldbe seriously weakened. Thus, a nuancedappreciation of the processes contributingto preservation of vertebrate skeletonsprovides the basis for alternatives to theclaims of purposeful burial. In sum, thisre-examination allows me to draw twoconclusions.

The evidence for MP burialFirst, there is no direct evidence for pur-poseful burial in the MP. Neither the reportsfrom La Chapelle-aux-Saints, La Ferrassie,Le Moustier, La Grotte du Régourdou,Shanidar, Teshik-Tash, Roc de Marsal,Kiik-Koba, La Quina, Amud (1), Tabunand Skhul (Gargett, 1989a,b), nor those Ihave examined in the present paper containany direct evidence for purposeful burial.Direct evidence would take the form of anartificial stratum containing the remains,which was created at the time of interment,and which could be distinguished fromnaturally accumulating sediments by beingdistinct from those upon which or within

which it occurs, and from those accumulat-ing above it and the original surface (cf.Drucker, 1972:5; Harris, 1979:95). Inthe absence of direct evidence, purposefulburial has been inferred on the basis ofweakly warranted assumptions. Forexample, the presence of hominid skeletalmaterial in or near erosional unconfor-mities has been used to argue that suchdepressions were created for the purpose ofburying hominid corpses. However, I haveargued that erosional unconformities canbe created by many processes, most ofwhich do not rely on a belief in an after-life. In addition, proximity to erosionalunconformities and low spots would helpto preserve articulated hominid skeletalmaterial.

Second, it is clear from pubished accountsof claimed MP burials that inferences ofpurposeful burial are based on assumptionsabout the way the archaeological record musthave formed, often without recourse to therange of processes capable of forming sites.In particular, most of the arguments explic-itly or implicitly hinge on the degree ofarticulation of skeletons and portions ofskeletons. Archaeologists have assumed thatskeletons would naturally decompose andbecome disarticulated, and on that basisthey have felt justified in making claims ofpurposeful burial. However, there is noadequate warrant for the claim that articu-lated skeletons would not preserve naturally.Instead, with a more nuanced view of caveand rockshelter deposition, and the priorprobability of preservation, it becomes clearthat purposeful burial is not the only processcapable of preserving articulated hominidremains in MP archaeological contexts.Thus, when examined from a taphonomicviewpoint, the archaeological contexts of allthe claimed burials described in this paperand in Gargett (1989a) provide evidenceof processes that are capable of naturallypreserving articulated hominid skeletalmaterial.

Specimen

La Chapelle-aux-Saints

Com

plet

e

Art

icu

late

d

Roc

kfal

l ev

iden

t

Cru

shed

sed

imen

ts

Alm

ost

com

plet

e

Tec

ton

ism

Nat

ura

l n

ich

e

Par

tial

ly c

ompl

ete

Par

tial

ly d

isar

ticu

late

d

Low

spo

t

Dis

arti

cula

ted

Fra

gmen

ts

Car

niv

ore

mod

ific

atio

n

Bas

in d

epos

its

Cry

otu

rbat

ion

Amud 1

Qafzeh 10

Shanidar 4

Shanidar 7

Qafzeh 9

Shanidar 1

Shanidar 3

Qafzeh 15

Shanidar 5

Shanidar 2

Qafzeh 8

Tabun CI

Dederiyeh

Qafzeh 11

Amud 7

Shanidar 6

Shanidar 9

La Ferrassie 1

La Ferrassie 2

Saint-Césaire

Kebara 2

Shanidar 8

Roc de Marsal

Kebara 1

Kiik- Koba

La Ferrassie 4a

La Ferrassie 6

La Ferrassie 3

La Ferrassie 4b

Teshik-Tash

La Ferrassie 5

Figure 9. Seriation by correspondence analysis of depositional environment and state of preservation for32 claimed MP burials. From left to right the data are ordered from complete to fragmentary, and fromarticulated to disarticulated. From top to bottom, the data are ordered from more rapid to more gradualdepositional contexts.

79

What the evidence tells usWhen the data on MP burial are looked at asa whole, one can see patterns that providestrong support for my proposed alternativesto purposeful burial. Figure 9 presents pub-lished data from 32 claimed burials. I havenow dealt in detail with most of them (seealso Gargett, 1989a). In general, I acceptthe observations of the excavators. Howeverin some cases I have made inferences thatare at variance with them, as follows. First,contrary to the excavators’ descriptions, Iobserve that Amud 7 and Kebara 2 are onlypartially articulated. Second, I think thatthere is good evidence for the occurrence ofrockfall at Qafzeh and Dederiyeh. Note thatcrushed sediments occur both where theexcavators have observed rockfall (i.e., atShanidar and Saint-Césaire), and at Qafzehand Dederiyeh where there is no explicitrecognition of it. Farrand (1985), remem-ber, considers crushed sediments goodproxy evidence for rockfall in the absence ofother sedimentary clues. In that vein, atQafzeh there are abundant crushed sedi-ments, in the form of the several hominidremains, while at Dederiyeh, rockfall wasinevitable beneath the chimney. The onlyother area in which I disagree with theobservations of the excavators is in the cat-egory of erosional unconformities, which Ihave termed ‘‘low spots.’’ Most are claimedto have been burial pits. La Ferrassie 1 and 2were discovered in a low spot created by thewall of the shelter and the sloping surface onwhich they rested. Thus, with the fewexceptions noted, the data presented inFigure 9 are consistent with the claims of theexcavators.

Using MV-Arch (Wright, 1992), the datawere subjected to a seriation by correspon-dence analysis, in which a parsimonioussolution is achieved by grouping recordswith similar sets of attributes. A seriation bycorrespondence analysis provides a graphi-cal representation of patterns in the data.It does not imply causal relationships, nor

does it attach statistical significance to theoutcome—it merely sorts like with like.Viewed along either axis of Figure 9,MV-Arch sorted the data along a continuumfrom complete to fragmentary, from articu-lated to disarticulated, and from (what Iwould consider) more rapid to more gradualdeposition. The specimens fall roughly intotwo groups: (1) those that occur along withevidence for rockfall or tectonism, or both;(2) those that occur in low spots as I havedefined them. Straddling this division arethose specimens found in natural niches orsimilar protected locations. Half of thosefound in protected contexts occur in siteswith evidence of rockfall; half where moregradual deposition was taking place.

It is immediately clear from Figure 9 thathalf of all the claimed burials are found insedimentary contexts that, on the evidence,are consistent with rockfall (i.e., crushedsediments and skeletal material, as well asevidence of bedrock breakdown). Theremainder are to be found in niches andlow spots.

A striking correspondence exists betweenthe better-preserved specimens and depo-sitional environments that I and others haveinferred to include rockfalls. From thesecome all three of the complete specimens,as well as a majority of almost complete(63%) and partial skeletons (58%). As onemight expect, given knowledge that suddenbedrock breakdown will tend to preserveskeletal material intact, only 27% of thefragmentary specimens come from contextswith evidence of rockfall.

There is also correspondence betweenrockfall contexts and the degree of articu-lation. A majority of the articulated speci-mens (69%) are found in depositionalenvironments bearing evidence of suddenbedrock breakdown, while 40% of the partlydisarticulated and only 27% of the disarticu-lated specimens are to be found in suchenvironments. Thus, in line with expec-tations, contexts containing evidence of

80 . .

rapid deposition are more likely to be thosewhere hominid skeletons are preserved morecompletely, and articulated.

The next clear pattern in these data is thatthe majority of the less well preserved speci-mens occur in natural niches or low spots.Natural niches can be expected to protectskeletal material somewhat from mechanicaldisturbance, and low spots act like sedimenttraps, which then experience a more rapiddepositional rate than open areas. All of thenatural niches in which MP hominids havebeen found occur in regions with activetectonism, and most co-occur with evidencefor rockfall. Thus, the evidence for rapiddeposition can explain the better preservedspecimens. However some specimens inniches fared less well, most probablybecause they were more gradually buried(e.g., Amud 7).

What I have termed low spots areespecially important to the discussion ofpurposeful burial, because at least some ofthese features are claimed to have beenburial pits. However, one sees remarkablyfew well-preserved specimens in these con-texts, a somewhat counterintuitive finding, ifone assumes that purposeful burial shouldpreserve hominid skeletons well. Figure 9illustrates that in low spots one finds none ofthe complete specimens, and only a minorityof almost complete (38%) and partiallycomplete (17%) specimens. Looked at inanother way, in low spots one finds themajority (67%) of fragmentary specimensand 75% of the less than fully articulatedspecimens. These data suggest that, in largemeasure, depositional rates are determiningthe state of preservation.

There are several clear inferences to bemade from the data in Figure 9. The first ofthese I have said before—all of the so-calledMP burials are found in special depositionalcircumstances that are capable of naturallypreserving hominid skeletal material. Not-withstanding the problematic claims fromTaramsa Hill, Egypt (Vermeersch et al.,

1998), there is not a single open-air contextwhere purposeful burial is indicated, asthere is from the early Upper Palaeolithicin Eastern Europe and from some of theearliest documented open-air sites inAustralia.

The second inference to be drawn fromFigure 9 is that there appears to be a nega-tive correspondence between the length oftime a corpse is exposed and the degreeto which it is preserved complete andarticulated. Rockfall, for example, isresponsible for rapid deposition in caves androckshelters. It is likely that even a smallrockfall would be capable of killing andinstantly burying a hominid, thus improvingits chances of preserving as an articulatedspecimen. On the other hand, deposition ina natural niche, or in a low spot, would bemore rapid than that occurring on moresalient surfaces, but would nevertheless beslower than rockfall. Thus, low spots andnatural niches would expose the remains todestructive processes for longer. Underthose circumstances skeletal material inthem would be less likely to preserve com-plete and articulated. Figure 9 bears witnessthat where more rapid deposition occurs,hominid skeletal material is much morelikely to be complete and articulated,whereas in circumstances where naturalburial would have been slower, the skeletalmaterial tends to be more fragmentary anddisarticulated.

The third major inference I draw fromFigure 9 does not refer to the data presentedthere. Instead I call attention to the skeletalevidence for hundreds of MP hominidsthat are too fragmentary to be considered‘‘burials.’’ I estimate that the number offragmentary MP hominids exceeds that ofclaimed burials by at least 10 to 1. Further-more, I estimate that the MP hominidsthat lived and died without leaving a tracewould outnumber the so-called burials byseveral orders of magnitude more than10 to 1. Thus, the vast number whose

81

remains did not survive were unluckyenough to die in places and at times whendepositional circumstances did not con-tribute to their preservation, i.e., eitherin the areas of caves and rockshelters thatwere subjected to the greatest destructiveprocesses, or out in the open, away fromthe protected environments of caves androckshelters.

The role of meticulous excavationThis re-examination was motivated by thenearly universal belief that rigorous exca-vation techniques provide secure inferencesof purposeful burial from Qafzeh, Kebara,La Roche à Pierrot, Amud and Dederiyeh.However, while rigorous excavation tech-niques are necessary to extract the maxi-mum information from an archaeologicalsite, they do not interpret the evidence, norcan they provide the models on whichrobust inferences depend. Rigorous exca-vation techniques can provide archaeologistswith the sedimentary clues with which toconstruct a depositional history, but theycannot provide an assessment of the priorprobability of preservation in a given depo-sitional environment. Rigorous excavationtechniques can recover plenty of pene-contemporaneous evidence; however, theycannot provide excavators with knowledgeof inherently variable processes with whichto infer a taphonomic history. In the end,rigorous excavation techniques cannotrescue inferences of purposeful MP burialfrom unwarranted assumptions, nor savearchaeologists from the consequences ofchoosing to ignore or downplay certainkinds of evidence, and seeing no value inothers. I hope I have demonstrated thatrigorous excavation techniques do nothingto help identify the kinds of evidence thatcan and do underpin satisfying alternativeexplanations. Finally, rigorous excavationtechniques cannot adopt the practice ofruling out natural causes before invoking

hominid behaviour to explain archaeologicaloccurrences.

Fortunately, for the purposes of thisre-examination, rigorous excavation tech-niques and their detailed publication doallow assessments of the kind presentedhere, illuminated by current understandingof site formation processes and taphonomy.The modern excavators have documentedmany more useful observations than wasthe case for La Chapelle-aux-Saints, LaFerrassie, Le Moustier, La Grotte duRégourdou, Shanidar and Teshik-Tash(Gargett, 1989a,b). The evidence recoveredfrom Qafzeh, Kebara, La Roche à Pierrot,Amud and Dederiyeh is therefore muchmore useful than was the case for thoseearlier discoveries. However, it is stillequivocal evidence for purposeful burial,and does not allow one to rule out naturalprocesses.

Implications for the archaeology of modernhuman originsAfter re-examining the evidence for pur-poseful burial at Qafzeh, Amud, Dederiyeh,Saint-Césaire and Kebara, I conclude thatclaims for purposeful burial in the MP relytoo heavily on weakly warranted assump-tions about how vertebrates are preserved incaves and rockshelters. My work fits in alarger framework of archaeological inquiry,fuelled by the suspicion that in the MP onemay be dealing with hominids that werecognitively different from modern humans.In recent years, other claims for modernhuman behaviours among MP hominidshave been called into question. Mousteriantechnology is not nearly so varied as wasonce thought (see e.g., Dibble, 1984;Barton, 1990; Rolland & Dibble, 1990).Levallois core and flake technology can nolonger be assumed to provide evidenceof a modern human mind (e.g., Noble &Davidson, 1996: Chapter 7). Notwithstand-ing strident claims to the contrary, so-calledevidence for MP representation is rarely

82 . .

unequivocal, and is often the result of natu-ral processes that have simply been misinter-preted (see e.g., Chase & Dibble, 1987;Davidson, 1990; D’Errico & Villa, 1997;D’Errico et al., 1998). Neandertals hunted,but not in a way that modern humans do,until perhaps 55,000 years ago (Stiner,1994), by which time there is some evidencethat modern humans were already on thescene (Davidson & Noble, 1992; Brookset al., 1995; Yellen, 1996). Spatial pattern-ing of MP archaeological sites cannot beinterpreted in the same way as that of mod-ern human sites, because many animals,including cave bears, wolves, and hyenas arecapable of creating behaviourally meaning-ful patterning. Thus, before one can arguethat MP spatial patterning stands as evi-dence for culturally-mediated (and thusmodern human) behaviour, archaeologistsmust first determine what kinds of spatialpatterning are unequivocal evidence ofmodern human mental abilities (Gargett,1996). The foregoing re-examination ofclaims for purposeful burial is therefore buta strand of a larger fabric of inquiry, one thatcalls into question a long-standing belief inthe abilities of MP hominids.

I have not presented the foregoing argu-ments with the intention of ‘‘proving’’ thatMP hominids were not burying their dead.Because even if one accepts my thesis thatnatural processes cannot be ruled out, therewould still be no basis on which to argue thatthe Neandertals and their contemporarieswere incapable of burying their dead. Norwould it be possible to say that they did nothave thoughts similar to our own with re-spect to death. One cannot, after all, arguethat absence of evidence is evidence ofabsence. MP hominids may have performedelaborate open-air rituals upon the death of agroup member, and have normally left thecorpse in a tree, or on a promontory. Thereis, for example, plenty of ethnographic evi-dence demonstrating that humans who leavea sparse and expedient archaeological record,

like that inferred for the Middle Palaeolithic,nevertheless take part in a complex spirituallife, with artistic expression almost entirelyrestricted to body painting, and painting onother perishable media. Neandertals, on thatmodel, may simply have chosen to make rep-resentational imagery on bark, skin, or in thesand, and not to paint or engrave rock out-crops or deep caves. However, archaeologistscannot be interested in what might have hap-pened. What really matters in the debatesabout modern human origins is not a matterof speculation, it is a matter of dealing withthe evidence as it exists. Thus, if one acceptsthe thesis that there is little unequivocal evi-dence for artistic representation or burial inthe MP archaeological record, it providesthe basis for scepticism that modern humanbehaviour began prior to about 40 to 60 kaago. However it does not give one licence toinfer that such behaviour did not exist priorto that time.

If, on the other hand, one chose to ignorethe alternative explanations presented in thispaper and continued to employ claims forburial and symbolic representation in argu-ments for an MP dawning of modern humanbehaviour, one would have to do so on shakygrounds. Moreover, one is left, under thosecircumstances, with a conundrum. In placesand at times over the last 40 to 60 ka,humans have left traces that do not look sodifferent from that of the MP. However,looked at globally, the archaeological recordof the recent period looks different from thatof the MP, and overall demonstrates greatvariability and flexibility. One can argue thatpaint wears off, and that caves painted200 ka ago might today retain no traces ofthat activity. If that is the case, and evidencefor MP engraving is so rare and contentious,one is left drawing conclusions about thecognitive abilities of MP hominid from otheraspects of behaviour. Thus, with tenuousevidence for burial and symbolic represen-tation in the MP, and with well-documentedevidence for different MP procurement

83

strategies, and questions about the mentalabilities required to produce MP stone tools,it remains an open question just why thearchaeological record of the MP is so differ-ent in character from that of the UpperPalaeolithic and later, and why it stayedmore or less static across vast distances,and through (according to conservativeestimates) 200 ka.

Thus, one is left with purposeful burial asreally the only behaviour that might demon-strate modern human thought processes(ignoring the leap required to get from pur-poseful burial to spirituality). In this paper Ihave argued that even the evidence for pur-poseful burial is equivocal. I would argue,therefore, that until archaeological under-standing improves, and our observations ofthe depositional circumstances of hominidskeletal material is such that we can ruleout natural processes, the existence of pur-poseful burial in the MP needs to be seen forwhat it is—an argument and not a fact. Anduntil there is much better evidence thatarticulated hominid skeletal material couldnot preserve naturally in caves androckshelters, MP burial will remain an argu-ment based on insufficiently warrantedassumptions.

Conclusion

Over the last 25 years, archaeologists haveincreasingly applied rigorous excavationtechniques in early Upper Pleistocene sitescontaining evidence of human ancestry.This contrasts with a great many discoveriesin the first half of the twentieth century.Recent publications include detailed sedi-mentological observations, adequate, if rep-resentative, illustrations of stratification, andmore attention to the taphonomic history ofthese sites. This has resulted in a muchricher set of data on the depositional cir-cumstances of MP hominids. Given theseimprovements in our understanding ofthe MP archaeological record, one might

have expected, contra Gargett (1989a), anavalanche of observations, construed asevidence, in arguments supporting claimsof purposeful MP burial. Yet despite anincrease in the quality and quantity of exca-vated material, arguments for purposefulMP burial have changed little since theearlier part of the century. For this reasonthey remain unsatisfactory, because theyrely on the same premises as the earlierclaims.

The arguments remain unconvincing fortwo reasons. First, they do not take intoaccount current understanding of the rangeof processes that could naturally preservearticulated hominid remains. Second, evenwhen archaeologists acknowledge thepotential for natural processes to affectpreservation, they do not acknowledge thevariability inherent in those processes.Therefore, expectations of what went on inthe past are limited to a narrow range ofpossibilities.

Good evidence may exist for purposefulMP burial. However, to date none has beenrecovered archaeologically. More often thannot, claims rest on the unwarranted assump-tion that skeletal articulation would onlyoccur if a corpse had been purposefullyburied. Occasionally, these claims are sup-ported by additional evidence, such as thepresence of an inferred burial pit. But in thispaper I have argued that even the obser-vation of an erosional unconformity is insuf-ficient to prop up the claim that purposefulburial had occurred.

I have proposed alternative explanationsfor the preservation of what are claimed tobe purposeful MP burials. My alternativesare based on empirically grounded knowl-edge of inherently variable taphonomicand geomorphic processes that could easilyhave preserved hominid skeletal material.In the section on the taphonomy of homi-nid burial I described the special, naturaldepositional circumstances in which articu-lated skeletons might be expected to occur,

84 . .

and provided evidence to support my con-tention that such depositional circum-stances are in evidence in all cases ofclaimed MP burial.

In summary,- The extraordinary occurrence of several

individuals in the small space of CoucheXVII at Qafzeh Cave can be explainedwithout having to invoke hominidbehaviour—the combination of friablebedrock and active tectonism would haveled to episodic deposition of breakdownrubble in large and small quantities, andthus would have contributed to thehominid skeletons’ preservation.

- The Saint-Césaire hominid’s highly frag-mented and fragmentary remains suggestprolonged exposure in a destructivedepositional environment. Indeed, atLa Roche à Pierrot there is abundantevidence of rockfall.

- Kebara 2’s incompleteness and less thanperfect articulation argues for naturalincorporation in the sediments, despiteclaims that there was a ‘‘burial pit’’ toprotect the corpse. Moreover, the ‘‘pit’’may even have begun to fill in before theskeleton was deposited, which argues forgradual burial in a naturally fillingdepression.

- The natural niche in which the Amudinfant came to rest would have meant abetter than average chance of preserva-tion, and its condition bears evidencethat it had been buried gradually.

- The Dederiyeh infant was discovered in adepression of unknown origin, at the topof an enormous talus cone beneath achimney. Such a depositional environ-ment can easily explain both the torturedpose for the little skeleton, and preserva-tion of the infant’s delicate bones. Onpresent evidence it is as difficult toargue in favour of burial as it is to argueagainst it.

Together with the results of a previousstudy (Gargett, 1989a) the present analysis

brings the matter of MP burial up todate. However, together, these two paperscannot settle the question of whether ornot MP hominids buried their dead.Rather, they underscore the need forarchaeologists to acknowledge the ambi-guity inherent in the traces they recover,and to refine their ability to distinguishbetween naturally and purposefully buriedhominid remains.

Acknowledgements

This research has had a long gestation, andbecause of it, I am indebted to many peoplefor its culmination. I am personallyacquainted with many of the people whoseinferences I criticize in this paper. And,while I clearly disagree with their inter-pretations, they already know this. I bearthem no ill will, and my remarks are madewithout prejudice. I am especially grateful toOfer Bar-Yosef for inviting me to take partin the excavations at Kebara Cave in 1989,and to Paul Goldberg for recommendingme. My conversations with the équipe atKebara were always stimulating, especiallythose with Bernard Vandermeersch andAnna Belfer-Cohen. Anna Backer graciouslyentertained my queries about Saint-Césaire1’s depositional circumstances. WilliamKimbel and Yoel Rak have entertained mewith good-humoured jibes about Neandertalburial. I gratefully acknowledge KnutFladmark for teaching me that archaeolo-gists must always attempt to rule out naturalprocesses before inferring human or homi-nid behaviour to explain the material tracesthey find, while to Brian Hayden go thanksfor teaching me that archaeologists can anddo ignore geomorphological and tapho-nomic processes in favour of fanciful inter-pretations of human agency. Funds fortravel to Israel were provided through anL. S. B. Leakey Trust grant and a RobertH. Lowie Scholarship. Thanks especiallyto Wendy Beck, Iain Davidson, Alice

85

Gorman, Rosalind James, StephanieLambretti, and Catherine Mitchell, whoread and commented on drafts, to PeterGrave for producing the seriation by corre-spondence analysis and to Douglas Hobbsfor producing the illustrations. Moreover, Iam grateful to JHE’s editor, Terry Harrison,five anonymous reviewers, and two not-so-anonymous reviewers—Paul Goldbergand Ofer Bar-Yosef—for asking challengingquestions, making pertinent suggestions andthrowing the brickbats that helped me totighten and strengthen my arguments.Belatedly, I wish to express my gratitude toGlen Guthrie, who was a ready soundingboard during an earlier phase of thisresearch. Finally, I must thank FrançoisLévêque for useful comments, and TakeruAkazawa for providing me with a translationof pertinent material. All of these people areresponsible for helping to make this papermore amenable to interpretation; however,they cannot be held accountable for itsarguments.

References

Akazawa, T. (1995). The Dederiyeh Neanderthalresuscitation project. In (T. Akazawa, Ed.) Dederiyeh:Neanderthals and Modern Humans, pp. 4–5. Tokyo:The University Museum, University of Tokyo.

Akazawa, T., Dodo, Y., Huhesen, S., Abdul-Salam, A.,Abe, Y., Kondo, O. & Mizoguchi, Y. (1993). TheNeanderthal remains from Dederiyeh Cave, Syria:interim report. Anthrop. Sci. 101, 361–387.

Akazawa, T., Muhesen, S., Dodo, Y., Kondo, O. &Mizoguchi, U. (1995a). Neanderthal infant burial.Nature 377, 585–586.

Akazawa, T., Muhesen, S., Dodo, Y., Kondo, O.,Mizoguchi, Y., Abe, Y., Nishiaki, Y., Ohta, S.,Oguchi, T. & Haydal, J. (1995b). Neanderthal infantburial from the Dederiyeh Cave in Syria. Paléorient21/2, 77–86.

ApSimon, A. M. (1980). The last Neanderthal inFrance? Nature 287, 271–272.

Arensburg, B., Bar-Yosef, O., Chech, M., Goldberg,P., Laville, H., Meignen, L., Rak, Y., Tchernov, E.,Tiller, A. M. & Vandermeersch, B. (1985). Unesépulture néandertalienne dans la grotte de Kébara(Israel). C.r. Acad. Sci. Paris 300, 227–230.

Bächler, E. (1921). Das Drachenloch bei Vättisim Tamintal. Jahr. St. Gallischen Naturwiss. Ges.57(1).

Backer, A. M. (1993). Spatial distributions at La Rocheà Pierrot, Saint-Césaire: changing uses of a rockshel-ter. In (F. Leveque, A. M. Backer & M. Guilbaud,Eds) Context of a Late Neandertal: Implications ofMultidisciplinary Research for the Transition to UpperPaleolithic Adaptations at Saint-Césaire, Charente-Maritime, France, pp. 105–127. Madison, WI:Prehistory Press.

Bahn, P. G. (1996). Treasure of the Sierra Atapuerca:a mountain range in northern Spain yields a cornu-copia of human remains, including the world’s oldestknown burials. Archaeology 49, 45.

Bar-Yosef, O. (1988). Evidence for Middle Palaeolithicsymbolic behaviour: a cautionary note. In (M. Otte,Ed.) L’Homme de Néandertal: Actes du ColloqueInternational de Liège (4–7 Décembre 1986), 5 (LaPensée), pp. 11–16. Liège: Etudes et RecherchesArchéologiques de l’Univérsité de Liège.

Bar-Yosef, O., Laville, H., Meignen, L., Tillier, A. M.,Vandermeersch, B., Arensburg, B., Belfer-Cohen,A., Goldberg, P., Rak, Y. & Tchernov, E. (1988). Lasépulture néanderthalienne de Kébara (unité XII). In(M. Otte, Ed.) L’Homme de Néandertal: Actes duColloque International de Liège (4–7 Décembre 1986), 5(La Pensée), pp. 17–24. Liège: Etudes et RecherchesArchéologiques de l’Univérsité de Liège.

Bar-Yosef, O. & Meignen, L. (1992). Insights intoLevantine Middle Paleolithic cultural variability. In(H. Dibble & P. Mellars, Eds) The Middle Paleolithic:Adaptation, Behavior, and Variability, pp. 163–182.Philadelphia: University Museum, University ofPennsylvania.

Bar-Yosef, O., Vandermeersch, B., Arensburg, B.,Belfer-Cohen, A., Goldberg, P., Laville, H.,Meignen, L., Rak, Y., Speth, J. D., Tchernov, E.,Tillier, A.-M. & Weiner, S. (1992). The excavationsin Kebara Cave, Mt Carmel. Curr. Anthrop. 33,497–550.

Bar-Yosef, O., Vandermeersch, B., Arensburg, B.,Goldberg, P., Laville, H., Meignen, L., Rak, Y.,Tchernov, E. & Tillier, A.-M. (1986). New data onthe origin of modern man in the Levant. Curr.Anthrop. 27, 63–64.

Barton, C. M. (1990). Beyond style and function: aview from the Middle Paleolithic. Am. Anthrop. 92,57–72.

Belfer-Cohen, A. & Hovers, E. (1992). In the eye of thebeholder: Mousterian and Natufian burials in theLevant. Curr. Anthrop. 33, 463–471.

Bergounioux, F. M. (1958). Spiritualité de l’hommede Néandertal. In (G. H. R. von Koenigswald,Ed.) Hundert Jahre Neanderthaler (NeanderthalCentenary), pp. 151–166. New York: Wenner-GrenFoundation.

Binford, L. R. (1981). Bones: Ancient Men and ModernMyths. New York: Academic Press.

Binford, L. R. (1985). Human ancestors: changingviews of their behavior. J. Anthrop. Archaeol. 4,292–327.

Binford, L. R. & Ho, C. K. (1985). Taphonomy at adistance: Zhoukoudian, ‘‘the cave home of Beijingman’’? Curr. Anthrop. 26, 413–442.

86 . .

Binford, S. R. (1968). A structural comparison ofdisposal of the dead in the Mousterian and UpperPaleolithic. Southwest. J. Anthrop. 24, 139–151.

Blanc, A. C. (1958). Torre in Pietra, Saccopastore,Monte Circeo: on the position of the Mousterianin the Pleistocene sequence of the Rome area. In(G. H. R. von Koenigswald, Ed.) Hundert JahreNeanderthaler (Neanderthal Centenary), pp. 167–174.New York: Wenner-Gren Foundation.

Bonch-Osmolovskij, G. A. (1940). Grot Kiik-Koba (inRussian). Moscow: Paleolit Kryma.

Bonifay, E. (1962). Un ensemble rituel moustérien à lagrotte de Régourdou (Montignac, Dordogne). InProceedings of the 6th International Congress of Prehis-toric and Protohistoric Sciences, pp. 136–140. Rome:ICPPS.

Bonifay, E. (1964). La grotte du Régourdou (Montignac,Dordogne): stratigraphie et industrie lithiquemoustérienne. Anthropologie, Paris 68, 49–64.

Bonifay, E. & Vandermeersch, B. (1962). Dépôts ritu-els d’ossements d’ours dans le gisement Moustériendu Régourdou (Montignac, Dordogne). C.r. Acad.Sci. Paris D:255, 1635–1636.

Bordes, F. & Lafille, J. (1962). Découverte d’unsquelette d’enfant moustérien dans le gisement duRoc de Marsal, commune de Campagne-du-Bugue(Dordogne). C.r. Séanc. Acad. Sci. Paris D:254,714–715.

Boule, M. (1909). L’homme fossile de la Chapelle-aux-Saints (Corrèze). Anthropologie, Paris 20,257–271.

Bouyssonie, A., Bouyssonie, J. & Bardon, L. (1908).Découverte d’un squelette humain moustérien à laBouffia de la Chapelle-aux-Saints (Corrèze). Anthro-pologie, Paris 19, 513–518.

Bouyssonie, A., Bouyssonie, J. & Bardon, L. (1913).La station moustérienne de la ‘‘Bouffia’’ Bonneval,à la Chapelle-aux-Saints. Anthropologie, Paris 24,609–634.

Bouyssonie, J. (1954). Les sépultures moustériennes.Quaternaria 1, 107–115.

Brain, C. K. (1967). Hottentot food remains and theirbearing on the interpretation of fossil bone assem-blages. Scientific Papers of the Namib Desert ResearchStation 32, 1–7.

Brain, C. K. (1969). The contribution of Namib DesertHottentots to an understanding of australopithecinebone accumulations. Scientific Papers of the NamibDesert Research Station 39, 13–22.

Brain, C. K. (1981). The Hunters or the Hunted? AnIntroduction to African Cave Taphonomy. Chicago:University of Chicago Press.

Bricker, H. (1989). Comment on ‘‘Grave shortcom-ings: the evidence for Neandertal burial,’’ by R. H.Gargett. Curr. Anthrop. 30, 177–178.

Brooks, A., Helgren, D. M., Cramer, J. S., Franklin,A., Hornyak, W., Keating, J. M., Klein, R. G., Rink,W. J., Schwarcz, H., Smith, J. N. L., Stewart, K.,Todd, N., Verniers, J. & Yellen, J. E. (1995). Datingand context of three middle Stone Age sites withbone points in the upper Semliki Valley, Zaire.Science 268, 548–553.

Capitan, L. & Peyrony, D. (1909). Deux squeletteshumains au milieu de foyers de l’époque moustéri-enne. Revue de l’Ecole d’Anthropologie 19, 402–409.

Capitan, L. & Peyrony, D. (1910). Deux squeletteshumains au milieu de foyers de l’epoque moustéri-enne. Bull. Soc. Anthrop. Paris 1, 48–53.

Capitan, L. & Peyrony, D. (1911). Un nouveausquelette humain fossile. Rev. Préhist. 6, 129–132.

Capitan, L. & Peyrony, D. (1912a). Station préhis-torique de la Ferrassie. Rev. Anthrop. 22, 29–50,76–99.

Capitan, L. & Peyrony, D. (1912b). Trois nouveauxsquelettes humains fossiles. Rev. Anthrop. 22, 439–442.

Capitan, L. & Peyrony, D. (1921). Découverte d’unsixième squelette moustérien à La Ferrassie(Dordogne). Rev. Anthrop. 31, 382–388.

Chase, P. G. & Dibble, H. (1987). Middle Paleolithicsymbolism. J. Anthrop. Archaeol. 6, 263–296.

Courty, M. A., Goldberg, P. & Macphail, R. (1989).Soils and Micromorphology in Archaeology. Sydney:Cambridge University Press.

Dart, R. (1957a). The Makapansgat AustralopithecineOsteodontokeratic Culture. Livingston: Chatto &Windus.

Dart, R. (1957b. The Osteodontokeratic Culture ofAustralopithecus promethus. Transvaal MuseumMemoirs, No. 10. Pretoria: Transvaal Museum.

Davidson, I. (1990). Bilzingsleben and early marking.Rock Art Res. 7, 52–56.

Davidson, I. & Noble, W. (1989). The archaeology ofperception: traces of depiction and language. Curr.Anthrop. 30, 125–155.

Davidson, I. & Noble, W. (1992). Why the first col-onisation of the Australian region is the earliestevidence of modern human behaviour. Archaeol.Ocean. 27, 113–119.

Defleur, A. (1993). Les Sépultures Moustériennes. Paris:C.N.R.S.

D’Errico, F. & Villa, P. (1997). Holes and grooves: thecontribution of microscopy and taphonomy to theproblem of art origins. J. hum. Evol. 33, 1–31.

D’Errico, F., Villa, P., Pintollona, A. C. & Idarraga,R. R. (1998). A Middle Palaeolithic origin ofmusic—using cave-bear bone accumulations toassess the Divje Babe 1 bone flute (Neandertal).Antiquity 72, 65–79.

Dettwyler, K. A. (1991). Can Paleopathology provideevidence for ‘‘compassion’’? Am. J. phys. Anthrop.84, 375–384.

Dibble, H. L. (1984). Interpreting typological variationof Middle Paleolithic scrapers: function, style, orsequence of reduction? J. field Archaeol. 11, 431–436.

Dreybrodt, W. (1988). Processes in Karst Systems:Physics, Chemistry, and Geology. New York: Springer-Verlag.

Drucker, P. (1972). Stratigraphy in Archaeology: AnIntroduction. New York: Addison-Wesley.

Duff, A. I., Clark, G. A. & Chadderdon, T. J. (1992).Symbolism in the Early Palaeolithic: a conceptualodyssey. Cambridge Archaeological Journal 2, 211–229.

87

Evans, G. L. (1961). The Friesenhahn Cave. Bull. Tex.Mem. Mus. 2, 1–22.

Farizy, C. & Masset, C. (1989). Comment on ‘‘Graveshortcomings: the evidence for Neandertal burial’’ byR. H. Gargett. Curr. Anthrop. 30, 179–180.

Farrand, W. R. (1985). Rockshelter and cave sedi-ments. In (J. K. Stein & W. R. Farrand, Eds)Archaeological Sediments in Context, pp. 21–39.Orono, ME: Center for the Study of Early Man.

Ford, T. D. (1976). The geology of caves. In (T. D.Ford & C. H. D. Cullingford, Eds) The Science ofSpeleology, pp. 11–60. New York: Academic Press.

Frayer, D. & Montet-White, A. (1989). Comment on‘‘Grave shortcomings: the evidence for Neandertalburial’’ by R. H. Gargett. Curr. Anthrop. 30, 180–181.

Galloway, A. (1989). Decay rates of human remains inan arid environment. J. forens. Sci. 34, 607–616.

Gamble, C. (1989). Comment on ‘‘Grave shortcom-ings: the evidence for Neandertal burial’’ by R. H.Gargett. Curr. Anthrop. 30, 181–182.

Gamble, C. (1993). Timewalkers: Prehistory of GlobalColonization. Stroud: Alan Sutton.

Gargett, R. H. (1989a). Grave shortcomings: theevidence for Neandertal burial. Curr. Anthrop. 30,157–190.

Gargett, R. H. (1989b). Reply to comments on ‘‘Graveshortcomings: the evidence for Neandertal burial’’.Curr. Anthrop. 30, 326–329.

Gargett, R. H. (1996). Cave Bears and Modern HumanOrigins: The Spatial Taphonomy of Pod Hradem Cave,Czech Republic. Lanham, MD: University Press ofAmerica.

Garrard, A. (1992). Comment on ‘‘The excavations inKebara Cave, Mt. Carmel,’’ by O. Bar-Yosef et al.Curr. Anthrop. 33, 535.

Garrod, D. A. E. & Bate, D. M. A. (1937). The StoneAge of Mount Carmel, Vol. I: Excavations at the WadyEl-Mughara. Oxford: Clarendon Press.

Gifford, D. P. (1981). Taphonomy and paleoecology: acritical review of archaeology’s sister disciplines. In(M. B. Schiffer, Ed.) Advances in ArchaeologicalMethod and Theory, Volume 4, pp. 365–438. NewYork: Academic Press.

Gifford-Gonzalez, D. (1991). Bones are not enough:analogues, knowledge, and interpretive strategies inzooarchaeology. J. anthrop. Archaeol. 10, 215–254.

Gifford-Gonzalez, D. P. (1989). Modern analogues:developing an interpretive framework. In (R.Bonnichson & M. Sorg, Eds) Bone Modification, pp.43–52. Orono, ME: University of Maine, Center forthe Study of the First Americans.

Gilman, A. (1989). Comment on ‘‘Grave shortcom-ings: the evidence for Neandertal burial,’’ by R. H.Gargett. Curr. Anthrop. 30, 182.

Gore, R. (1996). Neandertals. Nat. Geog. 189, 2–35.Gould, R. A. (1977). Puntutjarpa Rockshelter and the

Australian Desert Culture. New York: AmericanMuseum of Natural History.

Guthrie, R. D. (1967). Differential preservation andrecovery of Pleistocene large mammal remains inAlaska. J. Paleont. 41, 243–246.

Haglund, W. D. (1997). Dogs and coyotes: Postmor-tem involvement with human remains. In (W. D.Haglund & M. H. Sorg, Eds) Forensic Taphonomy:The Postmortem Fate of Human Remains, pp. 367–381.Boca Raton, FL: CRC Press.

Harris, E. C. (1979). Principles of ArchaeologicalStratigraphy. London: Academic Press.

Harrold, F. B. (1980). A comparative analysis ofEurasian Paleolithic burials. World Archaeol. 12, 195–211.

Hayden, B. (1979). Palaeolithic Reflections: Lithic Tech-nology and Ethnographic Excavation among AustralianAborigines. Canberra: Australian Institute ofAboriginal Studies.

Hayden, B. (1993). The cultural capacities of Neander-tals: a review and re-evaluation. J. hum. Evol. 24,113–146.

Haynes, G. (1982). Utilization and skeletal distur-bances of North American prey carcasses. Arctic 35,266–281.

Haynes, G. (1990). Taphonomy: science and folk-lore. In (S. Solomon, I. Davidson & D. Watson,Eds) Problem Solving in Taphonomy, pp. 7–17. StLucia: University of Queensland, AnthropologyMuseum.

Heim, J.-L. (1968). Les restes nèandertaliens de LaFerrassie: nouvelles donnèes sur la stratigraphie etinventaire des squelettes. C.r. Acad. Sci. Paris D:266,576–578.

Hill, A. (1979). Butchery and natural disarticu-lation: an investigatory technique. Am. Antiq. 44,739–744.

Hill, A. & Behrensmeyer, A. K. (1984). Disarticulationpatterns of some modern East African mammals.Paleobiology 10, 366–376.

Hovers, E., Rak, Y., Lavi, R. & Kimbel, W. H. (1995).Hominid remains from Amud Cave in the context ofthe Levantine Middle Paleolithic. Paléorient 21/2,47–61.

Jennings, J. N. (1985). Karst Geomorphology. Oxford:Blackwell.

Kerbis-Peterhans, J. C. (1990). The roles of porcu-pines, leopards and hyenas in ungulate carcass dis-persal: implications for paleoanthropology. Ph.D.Dissertation, University of Chicago.

Kirner, D. L., Burky, R. & Southon, J. R. (1997).Dating the Spirit Cave mummy: the value of re-examination. Nevada Hist. Soc. Quart. 40, 54.

Klein, R. G. (1996). The Mousterian of EuropeanRussia. Ph.D. Dissertation, University of Chicago.

Klein, R. G. (1994). The problem of modern humanorigins. In (M. H. Nitecki & D. V. Netecki, Eds)Origins of Anatomically Modern Humans, pp. 3–17.New York: Plenum Press.

Koby, F. E. (1941). Le ‘‘Charriage à sec’’ desossements dans les cavernes. Ed. geol. Helv. 34,319–320.

Koby, F. E. (1953). Les paléolithiques ont-ils chassél’ours des cavernes? Act. Soc. jurass. Émul. 57, 157–204.

Kurtén, B. (1976). The Cave Bear Story. New York:Columbia University Press.

88 . .

Lam, Y. M. (1992). Variability in the behaviour ofspotted hyaenas as taphonomic agents. J. archaeol.Sci. 19, 389–406.

Leroi-Gourhan, A. (1975). The flowers found withShanidar IV, a Neandertal burial in Iraq. Science 190,562–564.

Lévêque, F., Backer, A. M. & Guilbaud, M. (Eds)(1993). Context of a Late Neandertal: Implications ofMultidisciplinary Research for the Transition to UpperPaleolithic Adaptations at Saint-Césaire, Charente-Maritime, France. Madison, WI: Prehistory Press.

Lévêque, F. & Vandermeersch, B. (1980). Découvertede restes humains dans un niveau Castelperronien aSaint-Césaire, Charente-Maritime. C.r. Acad. Sci.Paris D:291, 187–189.

Lieberman, D. E. & Shea, J. J. (1994). Behavioraldifferences between Archaic and modern humans inthe Levantine Mousterian. Am. Anthrop. 96, 300–332.

Lindly, J. M. & Clark, G. A. (1990). Symbolism andmodern human origins. Curr. Anthrop. 31, 233–261.

Lyman, R. L. (1984). Bone density and differentialsurvivorship of fossil classes. J. anthrop. Archaeol. 3,259–299.

Marean, C. W. & Ehrhardt, C. L. (1995). Paleoanthro-pological and paleoecological implications of thetaphonomy of a sabertooth’s den. J. hum. Evol. 29,515–547.

Marshack, A. (1989). Evolution of the human capacity:the symbolic evidence. Yearb. phys. Anthrop. 32,1–34.

Martin, H. (1923). L’homme Fossile de La Quina. Paris:Octave Doin.

McBrearty, S. (1990). The origin of modern humans.Man (N.S.) 25, 129–143.

McKeown, A. M. & Bennett, J. L. (1995). A prelimi-nary investigation of postmortem tooth loss. J. forens.Sci. 40, 755–757.

Mellars, P. (1989). Major issues in the emergence ofmodern humans. Curr. Anthrop. 30, 349–385.

Mellars, P. (1995). The Neanderthal Legacy: AnArchaeological Perspective from Western Europe.Princeton, NJ: Princeton University Press.

Mercier, N., Valladas, H., Joron, J. L., Reyss, J. L.,Leveque, F. & Vandermeersch, B. (1991). Thermo-luminescence dating of the Late Neanderthalremains from Saint-Césaire. Nature 351, 737–739.

Micozzi, M. S. (1986). Experimental study of postmor-tem change under field conditions: effects of freezing,thawing, and mechanical injury. J. forens. Sci. 31,953–961.

Micozzi, M. S. (1997). Frozen environments and softtissue preservation. In (W. D. Haglund & M. H.Sorg, Eds) Forensic Taphonomy: The Postmortem Fateof Human Remains, pp. 171–180. Boca Raton, FL:CRC Press.

Miskovsky, J.-C. & Lévêque, F. (1993). The sedimentsand stratigraphy of Saint-Césaire: contributions tothe paleoclimatology of the site. In (F. Lévêque,A. M. Backer & M. Guilbaud, Eds) Context of a LateNeandertal: Implications of Multidisciplinary Researchfor the Transition to Upper Paleolithic Adaptations at

Saint-Césaire, Charente-Maritime, France, pp. 9–14.Madison, WI: Prehistory Press.

Mithen, S. (1995). Palaeolithic archaeology and theevolution of mind. J. archaeol. Res. 3, 305–332.

Mithen, S. (1996). The Prehistory of the Mind: A Searchfor the Origins of Art, Religion and Science. London:Thames & Hudson.

Movius, H. L. (1953). Mousterian cave of Teshik-Tash, southeastern Uzbekistan, Central Asia. Bull.Am. Sch. prehist. Res. 17, 11–71.

Newell, R. (1984). The archaeological, human biologi-cal, and comparative contexts of a catastrophically-terminated Kataligaaq house at Utqiagvik, Alaska(BAR-2). Arctic Anthrop. 21, 5–51.

Noble, W. & Davidson, I. (1989). Reply to commentson ‘‘The archaeology of perception: traces of depic-tion and language,’’ by I. Davidson and W. Noble.Curr. Anthrop. 30, 337–340.

Noble, W. & Davidson, I. (1991). The evolutionaryemergence of modern human behaviour: languageand its archaeology. Man (N.S.) 26, 602–632.

Noble, W. & Davidson, I. (1993). Tracing the emer-gence of modern human behavior: methodologicalpitfalls and a theoretical path. J. anthrop. Archaeol.12, 121–149.

Noble, W. & Davidson, I. (1996). Human Evolution,Language and Mind: A Psychological and Archaeologi-cal Inquiry. Melbourne: Cambridge University Press.

Okladnikov, A. (1949). Issledovanie musterskoi stoiankii pogrebeniia neandertaltsa va grotet Teshik-Tash,iuzhnyi Uzbekistan (Sredniaia Aziia) [Investigationof the Mousterian site and burial of a Neanderthalin the cave of Teshik-Tash, Southern Uzbekistan(Central Asia). In Teshik-Tash: Paleoliiticheskiichelovek [Teshik-Tash: Palaeolithic Man], pp. 7–85.Moscow: Trudy Nauchno-Issledovatelskogo InstitutaAntropologii.

Peyrony, D. (1921). Les moustériens inhumaient-ilsleur morts? Bull. Soc. Hist. Archaeol. Périgord, 132–139.

Peyrony, D. (1930). Le Moustier: ses gisements, sesindustries, ses couches géologiques. Rev. anthrop. 40,48–76, 155–176.

Peyrony, D. (1934). La Ferrassie: Moustérien, Périgor-dien, Aurignacien. Préhistoire 3, 1–92.

Rak, Y., Kimbel, W. H. & Hovers, E. (1994). ANeandertal infant from Amud Cave, Israel. J. hum.Evol. 26, 313–324.

Roe, D. (1992). Comment on ‘‘The excavations inKebara Cave, Mt. Carmel,’’ by O. Bar-Yosef et al.Curr. Anthrop. 33, 538.

Rolland, N. & Dibble, H. L. (1990). A new synthesis ofMiddle Paleolithic variability. Am. Antiq. 55, 480–499.

Rosenberg, K. R. (1992). Comment on ‘‘The exca-vations in Kebara Cave, Mt. Carmel,’’ by O. Bar-Yosef et al. Curr. Anthrop. 33, 538–539.

Schepartz, L. A. (1992). Comment on ‘‘The exca-vations in Kebara Cave, Mt. Carmel,’’ by O. Bar-Yosef et al. Curr. Anthrop. 33, 539.

Schepartz, L. A. (1993). Language and modern humanorigins. Yearb. phys. Anthrop. 36, 91–126.

89

Schick, T. & Stekelis, M. (1977). Mousterian assem-blages in Kebara Cave, Mount Carmel. Eretz Israel13, 97–149.

Schiegl, S., Goldberg, P., Bar-Yosef, O. & Weiner, S.(1996). Ash deposits in Hayonim and Kebara Cave,Israel: macroscopic, microscopic and mineralogicalobservations, and their archaeological implications.J. archaeol. Sci. 23, 763–781.

Shackley, M. (1980). Neanderthal Man. London:Duckworth.

Smirnov, Y. (1989). Intentional human burial: MiddlePaleolithic (Last Glaciation) beginnings. J. WorldPrehist. 3, 199–233.

Smith, F. H. (1992). Comment on ‘‘The excavations inKebara Cave, Mt. Carmel,’’ by O. Bar-Yosef et al.Curr. Anthrop. 33, 540–541.

Smith, P. & Arensburg, B. (1977). A Mousterianskeleton from Kebara Cave. Eretz Israel 13,164–176.

Solecki, R. (1955). Shanidar Cave, a Paleolithic site innorthern Iraq. Smithsonian Institution Annual Reportsfor 1954, 389–425.

Solecki, R. (1960). Three adult Neanderthal skeletonsfrom Shanidar Cave, northern Iraq. SmithsonianInstitution Annual Report for 1959, 603–625.

Solecki, R. (1961). New anthropological discoveries atShanidar, northern Iraq. Trans. N.Y. Acad. Sci. 23,690–699.

Solecki, R. (1963). Prehistory in the Shanidar Valley,N. Iraq. Science 139, 179–193.

Solecki, R. (1971). Shanidar: The First Flower People.New York: Knopf.

Solecki, R. & Leroi-Gourhan, A. (1961). Paleoclima-tology and archaeology in the Near East. Ann. N.Y.Acad. Sci. 95, 729–739.

Stiner, M. C. (1991). The faunal remains from GrottaGuattari: a taphonomic perspective. Curr. Anthrop.32, 103–117.

Stiner, M. C. (1994). Honor Among Thieves: A Zoo-archaeological Study of Neandertal Ecology. Princeton,NJ: Princeton University Press.

Stiner, M. C. & Kuhn, S. L. (1992). Subsistence,technology and adaptive variation in Middle Paleo-lithic Italy. Am. Anthrop. 94, 306–339.

Straus, L. G. (1982). Carnivores and cave sites inCantabrian Spain. J. anthrop. Res. 38, 75–96.

Stringer, C. & Gamble, C. (1993). In Search of theNeanderthals: Solving the Puzzle of Human Origins.London: Thames & Hudson.

Sutcliffe, A. J., Bramwell, D., King, A. & Walker, M.(1976). Cave paleontology and archaeology. In(T. D. Ford & C. H. D. Cullingford, Eds) TheScience of Speleology, pp. 495–520. New York:Academic Press.

Suzuki, H. & Takai, F. (Eds) (1970). The Amud Manand His Cave Site. Tokyo: University of Tokyo Press.

The Oxford Dictionary of Quotations. (1979). TheOxford Dictionary of Quotations. Oxford: OxfordUniversity Press.

Theunissen, R., Balme, J. & Beck, W. (1998). Head-room and human trampling: cave ceiling-heightdetermines the spatial patterning of stone artefacts at

Petzkes Cave, northern New South Wales. Antiquity72, 80–89.

Tillier, A.-M. (1990). Une controverse dépassée:l’existence de pratiques funéraires au Paléolithiquemoyen. Les Nouvelles de l’archéologie 40, 22–24.

Tillier, A.-M., Arensburg, B., Vandermeersch, B. &Rak, Y. (1991). L’Apport de Kébara la palethnologiefunèraire des néanderthaliens du Proche-Orient. In(O. Bar-Yosef & B. Vandermeersch, Eds) Le SqueletteMoustérien de Kébara 2, pp. 89–95. Paris: C.N.R.S.

Toots, H. (1965). Sequence of disarticulation in mam-malian skeletons. University of Wyoming Contributionsto Geology 4, 37–39.

Trinkaus, E. (1989). Comment on ‘‘Grave shortcom-ings: the evidence for Neandertal burial’’ by R. H.Gargett. Curr. Anthrop. 30, 183–185.

Trinkaus, E. & Shipman, P. (1993). The Neandertals:Changing the Image of Mankind. New York: Knopf.

Tuohy, D. R. & Dansie, A. (1996). Oldest mummy inNorth America is at Nevada State Museum. NevadaSt. Mus. Newsl. 24, 4–5.

Valladas, H., Joron, J. L. & Valladas, G. (1989).Application de la thermoluminescence à la datationdes niveaux moustériens de la Grotte de Kébara(Israel): ages Préliminaires des unités XII, XI et VI.In (O. Bar-Yosef & B. Vandermeersch, Eds) Investi-gations in South Levantine Prehistory, pp. 97–100.Oxford: British Archaeological Reports.

Valladas, H., Joron, J. L., Valladas, G., Arensburg, B.,Bar-Yosef, O., Belfer-Cohen, A., Goldberg, P.,Laville, H., Meignen, L., Rak, Y., Tillier, A. M.,Tchernov, E. & Vandermeersch, B. (1987). Thermo-luminescence dates for the Neanderthal burial site atKebara in Israel. Nature 330, 159–160.

Valladas, H., Reyss, J. L., Joron, J. L., Valladas, G.,Bar-Yosef, O. & Vandermeersch, B. (1988). Ther-moluminescence dating of Mousterian ‘‘Proto-Cro-Magnon’’ remains from Israel and the origin ofmodern man. Nature 331, 614–615.

Vandermeersch, B. (1996). Nouvelles découvertes derestes humains dans les couches Levalloiso-moustériennes du gisement de Qafzeh (Israël). C.r.Acad. Sci. Paris D:262, 1434–1436.

Vandermeersch, B. (1969). Les nouveaux squelettesmoustériens decouverts à Qafzeh (Israël) et leursignification. C.r. Acad. Sci. Paris D:268, 2562–2565.

Vandermeersch, B. (1970). Une sépulture moustéri-enne avec offrandes découverte dans la grotte deQafzeh. C.r. Acad. Sci. Paris D:270, 298–301.

Vandermeersch, B. (1976). Les sépultures néandertali-ennes. In (H. de Lumley, Ed.) La PréhistoireFrançaise, pp. 725–727. Paris: C.N.R.S.

Vandermeersch, B. (1981). Les Hommes Fossiles deQafzeh (Israel). Paris: C.N.R.S.

Vandermeersch, B. (1990). De Neandertalers en deeerste mensen. In 5 miljoen jaar menselijk avontuur,pp. 68–86. Brussels: King’s printer.

Vandermeersch, B. (1993). Was the Saint-Césaire dis-covery a burial? In (F. Leveque, A. M. Backer & M.Guilbaud, Eds) Context of a Late Neandertal: Impli-cations of Multidisciplinary Research for the Transition

90 . .

to Upper Paleolithic Adaptations at Saint-Césaire,Charente-Maritime, France, pp. 129–131. Madison,WI: Prehistory Press.

Vermeersch, P. M., Paulissen, E., Stokes, S., Charlier,C., Van Peer, P., Stringer, C. & Lindsay, W. (1998).A Middle Palaeolithic burial of a modern human atTaramsa Hill, Egypt. Antiquity 72, 475–484.

Villa, P. (1989). Comment on ‘‘Grave shortcomings:the evidence for Neandertal Burial,’’ by R. H.Gargett. Curr. Anthrop. 30, 325–326.

Voorhies, M. R. (1969). Taphonomy and PopulationDynamics of an Early Pliocene Vertebrate Fauna,Knox County, Nebraska. Laramie: University ofWyoming.

Weigelt, J. (1989). Recent Vertebrate Carcasses and theirPaleobiological Implications. Chicago: University ofChicago Press.

Weiner, S. & Goldberg, P. (1990). On-site Fouriertransform-infrared spectrometry at an archaeologicalexcavation. Spectroscopy 5, 46–50.

Weiner, S., Goldberg, P. & Bar-Yosef, O. (1993). Bonepreservation in Kebara Cave, Israel using on-siteFourier-transform infrared spectrometry. J. archaeol.Sci. 20, 613–627.

Weiner, S., Schiegl, S., Goldberg, P. & Bar-Yosef, O.(1995). Mineral assemblages in Kebara andHayonim, Israel: excavation strategies, bone preser-

vation and wood ash remnants. Israel J. Chem. 35,143–154.

White, T. D. & Toth, N. (1991). The question of ritualcannibalism at Grotta Guattari. Curr. Anthrop. 32,118–124.

White, W. B. (1988). Geomorphology and Hydrologyof Karst Terrains. New York: Oxford University Press.

Willermet, C. & Clark, G. A. (1995). Paradigm crisis inmodern human origins research. J. hum. Evol. 29,487–490.

Wolpoff, M. H., Spuhler, J. N., Smith, F. H.,Radovcic, J., Pope, G., Frayer, D., Eckhardt, R. &Clark, G. (1988). Modern human origins. Science241, 772–773.

Worsaae, J. J. A. (1843). Danmarks Oldtid. Copenhagen.Wright, R. V. S. (1992). MV-Arch: A Multivariate

Statistical Package for Archaeology. Sydney: R. V. S.Wright.

Wylie, A. (1989). Archaeological cables and tacking:the implications of practice for Bernstein’s ‘‘Optionsbeyond objectivism and relativism’’. Phil. soc. Sci. 19,1–8.

Wynn, T. (1993). Two developments in the mind ofearly Homo. J. anthrop. Archaeol. 12, 299–322.

Yellen, J. E. (1996). Behavioural and taphonomic pat-terning at Katanda 9: a Middle Stone Age site, KivuProvince, Zaire. J. archaeol. Sci. 23, 915–932.

Documents

Robert H. Gargett Middle Palaeolithic burial is not a dead ... · human origins, taphonomy, site formation Middle Palaeolithic burial is not a dead issue: the view from Qafzeh, Saint-Césaire,