2007. 'Totems', 'Ancestors', and 'Animism'. The Archaeology of Ritual, Shrines, and Sacrifice amongst the Tallensi of Northern Ghana. (In), Barrowclough, D., and Malone, C. (eds.),

Chapter 43

.TOTEMS

"'ANCESTORS "

AND'ANIMISM' .THEARCHAEOLOGY OF RITUAL, SHRINESAND SACRIFICE AMONGST THE TALLENSI

OF NORTHERN GHANA

Timothy InsoII

Introduction'Totems', 'ancestors', and 'animism' are all seeminglyfamiliar terms, applicable across wide geographical andtemporal contexts to describe religious phenomena. Yet inreality they are not as straightforward as they might initiallyappear either as regards definition or archaeologicalapplication. This paper seeks to briefly explore the utilityof these terms in relation to the results of an archaeologicalproject begun in 2004 and focussed upon the Tallensipeople, subsistence agriculturalists who speak a Gurlanguage of the Oti-Volta group (Naden 1998,12), and wholive in the Upper East Region of northern Ghana in WestAfrica (Figure 43.I).In so doing attention will be primarilygiven to the archaeology of ritual, shrines, and sacrifice inthe Tongo Hills, the area inhabited by one of the two maingroups of clans forming the Tallensi, the Hill Talis. This isbecause in concentrating upon these aspects what isimmediately clear is that the conceptualisation of religiousphenomena such as 'totems', 'ancestors' or 'animism' inthe singular is flawed and instead interpretive plurality isrequired. A premise which it will be suggested is probablyequally valid for archaeological contexts elsewhere.

The Tallensi of Northern Ghana are well-known via theseminal anthropological studies of Fortes (e.g. 1945; 1949;1987). Fortes provides a wealth of detail on, for instance,kinship and aspects of Tallensi religion and ritual practicein relation to ancestor worship and to a lesser extent shrines.However, his functionalist approach means that Tallensimaterial culture is comparatively neglected as is thehistorical dimension of their existence, though the latterhas been redressed recently (Allman and Parker 2005).Thus until the start of the current research project in 2004(Insoll, Kankpeyeng, and Maclean 2004;2005; in press a)no archaeological research had been completed in the TongoHills, the epicentre of Tallensi settlement, beyond an

ethnoarchaeological study of Tallensi house compounds(Gabrilopoulos 1995), and an inventory of cultural heritagecompleted by one of the project co-directors (Kankpeyeng

2001).Potentially, Tallensi derived material provides a way of

beginning to examine religious phenomena and definitions,and their material signatures and archaeologicalimplications for the majority of the Tallensi remainfollowers of traditional religion. This, however, is complexand Tallensi religion has as central elements both ancestraland earth cults. Ancestral worship functions on variouslevels. Each segment of a Talis composite clan has a lineageshrine to distinguish it from other segments (Fortes 1949,6), to which they sacrifice individually (Fortes 1950, 253).Contrasting with this, groups of maximal lineagesbelonging to different clans (but not necessarily united byties of clanship) collaborate, 'in the cult of their collectiveancestors' (Fortes 1949,6) via joint sacrifice. Materially,the ancestral cult is represented by the ancestral shrine orba'a (Allman and Parker 2005,43) which can be manifestin various ways, as a household shrine, or in an externalshrine such as a sacred grgve or cave (Kankpeyeng 2001),but involving ritual practices which have no specialistpriests (Fortes 1987, 150).

The latter, however, are specifically involved in the earthcult, for this is linked with a class of specialist priest, theTbndaana, the 'Custodian of the Earth' (Fortes 1987, 43),and is materially manifest through sacred places, ten, withinwhich are located the earth shrines, tengbana, which thepriests serve. Secondly, it also differs in the beliefs it reflectsbeing, obviously, connected with the sacred aspect of theearth, which is personified, and which is envisaged,according to Fortes (1987, 135-6), as 'a living force...complementary to the collective ancestorso. To the ancestraland earth cults must also be added totemism, as another

329

T]URKINA FNSO

l. Lawra2. Bolgatanga3. Tongo4. Bawku

2 su*"y Area

Figure 43.1. Map of Northern Ghana.

element of Tallensi religion. All these components,however, form the ultimate whole which is Tallensi religion,and the existence of a metaphysical upper 'tier' to Tallensireligion must also be acknowledged (see Fortes 1987).

The archaeology of ritual and shrinesMaterially, Tallensi religion in all its varied forms ismanifest via shrines which exist in a bewildering range ofconfigurations. Moreover that these are not static as astrictly typological approach might suggest (Insoll inpreparation a) is clearly indicated via archaeology, i.e. thatboth shrines and the ritual practices they sustain andgenerate can instead be adaptive, dynamic, and flexible.Part of the problem here in relation to shrines lies indefining exactly what constitutes a religious place, a focusof devotion, for this is not always best served by the Englishterminology which is often used within the context ofdescribing relevant material associated with eithercontemporary traditional religions or in prehistoricarchaeological contexts, i.e. 'shrine'. The primary reasonfor this being that 'shrine' is ultimately derived from theLatin scrinium - meaning 'box' or 'receptacle', as in

'containers of sacred meaning and power'(Courtright 1987,299). This being a term, which it can be suggested, thatsingularly fails to describe the range of structures includedwithin its boundaries (Insoll 2004,105). For simply stated,the shrine can contain shrines, as with sacred groves -shrines - that contain standing stones therein - shrines -which are in turn associated with sacred pots - shrines - ascan be seen at Nyoo. Hence in the attempt to 'typologise'shrines complexity could easily be lost and an unwarranteddegree of uniformity instituted instead.

Innumerable shrines exist in the Tongo Hills, but threeshrines dominate, Tonna'ab/Yaane, Bonaab, and Nyoo. Ofthese Nyoo is perhaps of primary interest to thearchaeologist being a large 'sacred grove' functioning asan earth shrine and of further significance in being usedannually as a dancing ground at some point from lateFebruary to early April during the pre-agricultural Gologoor Golib festival (Fortes 1987,34; Kankpeyeng 2001,26).That Nyoo was not a natural sacred grove was immediatelyindicated by archaeological survey. For far-spreadingarchaeological vestiges visible on the surface clearly showedthat this was in fact an extensive enshrined archaeologicalsite divided into a series of different zones. These included

330 TIMOTHY INSOLL

a 'field' of standing stones, an area of stone arrangementswith an associated spread of poftery covering some 300 meast to west, and an active sacrificial area lacking suchovert archaeological features (Figure 43.2). Furthermore,test excavations completed in 2005 confirmed that this wasnot a 'natural' sacred grove with, for example, the potteryspread seen to be formed of both complete pottery vesselsand sherds which had been forced into the ground to amaximum depth of 30 cm, and thus seemingly rep-resentative of an act of structured 'ritual' deposition (Insoll,Kankpeyeng, and Maclean 2005). Hence Nyoo was madethe focus of excavation in July 2006 during which twounits were opened.

NYOO 06 (A)The first of these units was assigned the code NYOO 06(A) and measured 8 m x 4 m. Only shallow archaeologicallayers were encountered with a maximum depth of c.I5-20cm before sterile deposits were reached. This noted, thematrix that was removed was densely filled witharchaeological material, predominantly sherds, many fromcomplete vessels apparently broken in-situ, but alsocontaining an assemblage of 35 lithic objects comprisinglargely stone grinder/pounder/rubbers, both fragmentary,and complete, but also lumps of quartz (Insoll, Kankpeyeng,and Maclean in press b).

Interspersed amongst the pot filled deposits were sevenstone arrangements (Figure 43.3). Originally, it wasthought that these might represent cairns, but this idea wasdiscarded on the basis that the stone arrangements werealmost entirely composed of a single layer of stones.Similarly, there was apparently no significance in thenumbers of stones composing the arrangements whichranged between 16 and 51. However, care had definitelybeen taken in the uurangements of the stones, possibly withsome concern evident as to the colour patterning of theaplite (fine-grained granite), Bongo granite, ild schistpresent. The colours red, pink, black, and grey were noted,with white represented by smaller fragments of quartzfrequently found, as well as by the banding in some of thegranite. In so-doing perhaps providing a further manifesta-tion of the oft-noted red-white-black colour symbolismevident in sub-Saharan Africa (Turner 1985; Jacobsen-Widding 1979). Removal of the stone arrangementsproduced very little material, and it would seem that theFigure 43.2. Plan of Nyoo.

Figure 43.3. IWOO 06 (A) indicating the stone arrangements and pot fiIled deposits in-situ (all photos T. Insoll).

33r

pots represented by the pot filled layers were depositedafter the stones had been arranged (Insoll, Kankpeyeng,and Maclean in press b).

Dates from this and the other excavation completed arein progress, but it is not unreasonable to suggest thatperhaps one deposition event involving a lot of people andpottery is represented, rather than a build-up of materialover some length of time involving multiple depositionevents. This suggestion can be made based upon pre-liminary indications with regard to the pottery assemblagewhich indicate little variability present, but rather astandard range of vessel types found (Ashley 2006). Theabsence of any contextual difference also supports this ideaof a single, or at least a rapid, deposition event. Though thegradation in density of pottery present indicates that asimplistic uniform infilling around the stone arrangementsdid not take place.

What the deposition of the pot and other materials mightrepresent, or be associated with, is unclear. Varioussuggestions can be made, with perhaps the most compellingbeing that it might broadly function within the frameworkof commemorating or supplicating the ancestors ordeceased. Though the absence of funerary remainsprecludes a direct link with the dead (similarly absent inthe other unit as well). This interpretation would seemplausible based upon broad parallels elsewhere, less so,perhaps, with Tallensi practices today, but certainlyreminiscent of, for instance, the Akan Asensie, or 'place ofpots' (Bellis 1982). A point qualified with the proviso thatdirectAkan connections are notbeing proposed, just generic

parallels suggested. The deliberate destruction of some ofthe pots, with holes forced, bored, or chipped in their basesfor example, might support this association with theancestors or deceased. An interpretation lent further weightby the results of the excavations in NYOO 06 (B) discussedbelow.

Although it is unwise at this preliminary stage toadvance too far in interpretation, it can again tentatively besuggested that based on contemporary parallels what mightalso be represented by the stone arrangements at Nyoo isthe residue of perfonnance or movement, perhaps dance.Again, during the ritual activities observed at Nyoo Biil,dance around and between the stone arrangements wasseen to be a key part of the Golib festival (Figure 43.4).That such activities might also have occurred in the area ofstone arrangements excavated in Nyoo would not seeminconceivable considering the contemporary parallels withTallensi practices which exist.

NYOO 06 (B)The second unit excavated reinforced this image of Nyooas a complex archaeological site rather than a 'natural'sacred grove. This excavation was completed within thearea of Nyoo most densely filled with standing stones. Intotal 12 clusters of either paired or single standing stoneswere recorded (Figure 43.5). Almost immediately uponexcavating groups of iron bracelets and other artifacts suchas points which had been placed adjacent to the standingstones were recorded. Ultimately, eight complete iron

Figure 43.4. Dance in progress in Nyoo Biil.

332 TIMOTHY INSOLL

Figure 43.5. NYOO 06 (B) indicating the standing stones and associated ritual pots in-situ-

bracelets, seven iron bracelet fragments, one ifon finger

ring, five iron points, and two fragments of iron strip were

recovered. The bracelets were almost uniformly of simple

design and their presence was described as representing

the interring of 'personal gods' associated with the dead,

i.e. intimate personal possessions, possibly following the

instructions of diviners to carry out such actions (R. Yin

pers. comm.).This would broadly concur with Fortes (1987, 267)

description of the notion of sii and its links with concepts

of the person - personhood - and personal possessions.

Specifically, to quote (Fortes 1987,267),'sli, therefore, in

one of its aspects, is the focus, one might almost say the

medium, of personal identity which is objectively rep-

resented in possessions characteristic of a person's sex and

status'. Such possessions as described by Fortes (1987)

included the individuals clothing, and for a woman her

personal ornaments such as brass bracelets and beadwork,

and for a man, his tools such as the hoe and axe, or bow

and arrow. Iron bracelets would and could certainly

constitute such a category of intimate personal possessions,

of precise ontological status and association, even if we

cannot go so far as to interpret which gender and/or age or

initiation status mighthave been linked with these artefacts.

Considering the evidence from both units it would

appear that distinct differences in ritual practice are

manifest in Nyoo, both in the excavated materials recovered

and the surface features recorded. This notion of differential

ritual practices across Nyoo is lent further support by

thirteen complete, and one partially complete, pots which

were recovered below and slightty adjacent to the standing

stones. The ritual nature of the deposits in this area of Nyoo

was thus firmly indicated by the discovery of these pots and

by two pear-shaped clay objects, broken, but still conjoining,

which were uncovered in association with one of these

complete pots from NYOO 06 (B). These clay objects (or

object, as they had been joined) had a hole in each of the

two pear-shaped segments (Figure 43.6). Again, providing

a precise interpretation as to what this object was used for

is impossible, and opinions sought from community

members varied though generally consensus existed in that

it was:

a) A ritual object.b) Probably offered libation and/or sacrifice (hence the

holes).c) Functioned as a 'Personal God'.

That this object might have functioned within the

context, perhaps, of fertility concerns or rituals would be

entirely plausible, if unproven. Their actual shape, 'pear-

.TOTEMS"'ANCESTORS" AND'ANIMISM' aaa

JJJ

Figure 43.6. The clay ritual object.

shaped' being a somewhat neutral description, isreminiscent of a pair of testicles, and if correct, a fertilityassociation would thus not seem unwarranted. Directparallels have not been found, though preston Blier oggi..48) refers to small earthern balls which are produced by theBatammaliba midwife following the birth of a child. Theseare made around the infant's birth sack 'as a symbol of thecreat ive process' (Preston Bl ier l9g7, l l9) . No suchmeaning is proposed for the Nyoo object, and the latter isnot even a ball shape, but it does indicate the ritual use ofclay ro produce generically similar ritual objects by ulinguistically related group from Togo.

As to the meaning of the standing stones, bracelets,pots, clay objects or stone arrangements and their associatedrituals in totality, precise information from the Tallensiwas not forthcoming - it was seemingly lost. Rather, theirgeneral meanings were understood, namely that thismaterial was linked into negotiating destiny via the agencyof the aforementioned 'personal gods' and hence functionedwithin the framework of ancestral worship. Moreover,general consensus existed in that Nyoo should be consideredas the great shrine for all the Tallensi, where worship startedand spread from (J. B. zuure pers. comm.). It could thus ineffect be called a reservoir or nucleus of ritual practice, andis broadly analogous to the notion of 'symbolic reservoirs'as discussed by Sterner (1992, l7 l_I7Z; and seeMacEachern 1994), but 'symbol' would here be translatedinto 'ritual'.

The archaeology of sacrifice and the sensesThe practice of sacrifice could be indicated archaeologicallyby many of the categories of material recovered, the brokenpottery or the iron bracelets already described, for throughtheir deposition it is possible that they too underwent'sacrifice'. The shedding of blood, however, is more usuallyequated with sacrifice in Africa (see De Heusch 19g5) andthe almost total absence of faunal remains in either of theexcavations is of interest in this respect for two primaryreasons. Firstly, for totemic observances in relation to theconsumption of certain species are a key element of rallensibelief and social structure today (see Fortes r987, 136 and249). And secondly, for the simple reason that bloodsacrifice of domestic animals such as chickens, dogs,donkeys, sheep, goats, and cattle is one of the primary actsof ritual practice by the Tallensi, either undertakenindividually or collectively (Insoll 2006). How then do weaccount for the absence of sacrificial residues in thearchaeological record - perhaps as an absence of bloodsacrifice in earlier, but as yet undated ritual practice asmanifest in Nyoo? This is very unlikely, rather it is possibleto look to the ethnographic present again to provide possibleexplanations. Specifically, that such remains disappear forsacrifice is only the act and the residues are not wasted.Instead they are carefully divided according to strict criteriawhich reflect clan association and social structure.Moreover, what is left; blood, or stomach contents forinstance, is soon consumed by scavengers such as dogs,vultures, and dung beetles.

Hence the sacrifice remains are distributed throughoutthe landscape rather than clustered in one place to enter thearchaeological record as at Nyoo. But this is significant forthe absence of faunal remains is again reflecting ritualpractice, but here seemingly in accord with what occursamong the Tallensi today, rather than at variance with it asis represented by much of the evidence from Nyoo. It is alsonotable that where faunal remains are distributed in thelandscape they can also take on dimensions inaccessible tothe archaeologist. This is especially so with the powerfulobjects formed of blood, fat, skin, bones, and fur which aresometimes attached to the exterior of the house shrine orritual structure, the zong, which is associated with ancestralworship (Gabrilopoulos 1995, 70).

The example in Figure 43.7 is slowly rotting but isplaced adjacent to the main entrance of a Tallensi chief'shouse in the Tongo Hills. These objects give off a powerfulsmell and are still accorded a highly prominent position inthe house of a prominent individual. They provide a truesensory experience, of smell, often lacking in archaeo-logical contexts but which would have been represented atplaces of sacrifice by the all-pervading smell of blood, as isimmediately apparent in sacrificial areas today. It is notimprobable, if unproven, that the excavated areas in Nyoocould also have been used for sacrifice. Hence besides the

r

334 TIMOTHY INSOLL

Figure 43.7. Tallensi powerful objects formed of fat, skin, bone etc. on Zong.

recognition of sound as a sense potentially of interpretiverelevance in relation to dance and performance as alreadydescribed, smell might also have been a significant, butmissing element of our elusive sensory archaeology as well.

ConclusionsIt is thus evident that both 'shrines' and 'sacrifice' areevident archaeologically but indications of 'totemism','ancestors', and 'animism' are not. Or is such a simplestatement warranted? It is not, for all these aspects andnone are represented if such a contradiction is possible.This is because what is structuring all the elements presentis complexity; i.e. defining shrines is problematic, sacrificemight be evident in terms of pottery vessels but not faunalremains, whereas 'ancestors' are potentially placated bythe deposition practices evident, whilst 'animism' or atleast 'animistic' concepts may be ascribed to materials suchas rock (Insoll 2006), and totemic affiliation could berepresented, in the absence of any other discerniblepatterning, via the use of the stone arrangements bydifferent clans. All of this is of course supposition, but thisis the key point. For attempting to generate singularinterpretations of religious phenomena based uponarchaeological material be it in relation to 'totemism','ancestor cul ts ' , 'animism' or the current vogue,

'shamanism' (Pearson2002; Insoll 2004) is nearly alwaysimplausible for these are usually based upon supposed' clear-cut' modern definitions which bare little resemblanceto the complex entities which exist in reality (Levi-StraussI99l; Bowie 2000; Insoll in press a).

Instead what is apparent is that the one aspect so farmissing from this concluding discussion but included inthe title, 'ritual', is actually the most clearly discerniblearchaeologically. But notritual as some sort of vague catch-all category for the otherwise not understood, rather ritualas repeat action of a religious nature in relation to materialculture (Bell 1992; Insoll -2004). Here we perhaps have awindow into the dynamic nature of Tallensi or 'proto'-Tallensi ritual via archaeology. The specifics of these ritualsmight be bluned, even with the existence of supportinganalogies drawn from extant practices available, but whatis recurrent is the importance both of performance and thesensory dimension. The latter potentially manifest inrelation to dance, and by association sound. Anotherassociated sensory feature which has been briefly exploredis smell, whilst vision could be similarly considered andboth are significant in Tallensi ritual today.

This research project, and hence its interpretations, arestill at an early stage. Nonetheless, it is hoped that as

research proceeds further insights will be gained into thearchaeology of ritual and religion in the Tongo Hills.

335

However beyond this West African context, the potential

interpretive relevance of this material is already clear be itfor the Neolithic of the British Isles (Insoll 2006) or theBronze Age of southern England (Insoll in preparation b).For what is immediately apparent is that complexity inreligious definition and configuration must, seemingly, beinvoked and singular phenomena discounted as more akinto contemporary western conceptualisation of what'traditional' religions are rather than, potentially, a moreaccurate configuration of the complexity of past religions.

ReferencesAllman, J. and Parker, J.2005. Tongnaab. Bloomington, Indiana

University Press.Ashley, C. 2006. Unpublished Ceramics Field Notes from

Tengzug, 2006. Department of Archaeology, University ofManchester.

Bell, C. 1992. Ritual Theory, Ritual Practice. Oxford, OxfordUniversity Press.

Bellis, l. O. 1982. The 'Place of the Pots' in Akan FuneraryCustom. Unpublished Report. African Studies Programme,Indiana University.

Bowie, F. 2000. The Anthropology of Religion. Oxford,Blackwells.

Courtright, P. 1987. Shrines. In M. Eliade (ed.). The Encyclopediaof Religion London, Macmillan, 299-302.

de Heusch, Luc. 1985. Sacrif ice in Afr ica. Manchester,Manchester University Press.

Fortes, M. 1945 [1969]. The Dynamics of Clanship among theTallensi. Oosterhout, Anthropological Publications.

Fortes, M. 1949 11967l. The Web of Kingship among the Thllensi.Oosterhout, Anthropological Publications.

Fortes, M. 1950. The Political System of the Tallensi of theNorthern Territories of the Gold Coast. In M. Fortes and E. E.Evans-Pritchard (eds). African Political Systems. London,Oxford University Press, 238-7L.

Fortes, M. 1987. Religion, Morality and the Person. Essays onTallensi Religion. Cambridge, Cambridge University Press.

Gabrilopoulos, N. 1995. Ethnoarchaeology of the TallensiCompound Unpublished MA Thesis, University of Calgary.

Insoll, T.2004. Archaeology, Ritual, Religion. London, Routledge.Insoll, T.2006. Shrine Franchising and the Neolithic in the British

Isles: Some Observations based upon the Tallensi, NorthernGhana. Cambridge Archaeological Journal 16, 223-38.

Insoll, T. In Press a. Archaeology. The Conceptual Challenge.London, Duckworth.

Insoll, T. In Preparation a. Natural or Human Spaces? TallensiSacred Groves and Shrines and their Implications forPhenomenological Interpretations of the Past (submitted toNorw e gian Archaeolo gical Review).

Insoll, T. In Preparation b. Negotiating the Archaeology of Destiny.Observations from the Tallensi of the Tongo Hills, NorthernGhana.

Insol l , T., Kankpeyeng, B., and Maclean, R. 2004. AnArchaeological Reconnaissance in the Tong Hills, and GaruArea, Upper East Region, and Nakpanduri, Northern Region,Ghana. Nyame Akuma 62,25-33.

Insoll, T., Kankpeyeng, B., and Maclean, R. 2005. Excavationsand Surveys in the Tongo Hills, Upper East Region, Ghana.July 2005. Fieldwork Report. Nyame Akuma 64, 16-23.

Insol l , T., Kankpeyeng, B., and Maclean, R. In Press a.Excavations and Surveys in the Tongo Hills, Upper East Region,Ghana. July 2006. A Preliminary Fieldwork Report. NyameAkuma.

Insoll, T., Kankpeyeng, 8., and Maclean, R. In Press b. TheArchaeology of Shrines among the Tallensi of Northern Ghana:Materiality and Interpretative Relevance. In A. Dawson and P.Shinnie (eds). Shrines in African Societies. Calgary, TheUniversity of Calgary Press.

Jacobson-Widding, A. 1979. Red-White-Black as a Mode ofThought. Uppsala, Almqvist and Wiksell.

Kankpeyeng, B. 2001. The Cultural Landscape of Tongo-Tengzuk.Traditional Conservation Practices. In Anon. TraditionalConservation Practices in Africa. Nantes, CRATerre-EAG, 23-3r.

L6vi-Strauss, C. 199I. Totemism. London, Merlin Press.MacEachern, S. 1,994. 'Symbolic reservoirs' and Inter-group

Relations: West African Examples. African ArchaeologicalReview 12.205-224.

Naden, T. 1988. The Gur Languages. In M. E. Kropp Dakubu(ed.). The Innguages of Ghana. London, Kegan Paul, 1249.

Pearson, J. L. 2002. Shamanism and the Ancient Mind. WalnutCreek, Altamira.

Preston Blier, S. 1987. The Anatomy of Architecture. Ontologyand Metaphor in Batammaliba Architectural Expression.Cambridge, Cambridge University Press.

Sterner, J. 1992. Sacred Pots and 'Symbolic Reservoirs' in theMandara Highlands of Northern Cameroon. In J. Sterner andN. David (eds.). An African Commitment. Calgary, Universityof Calgary Press, 17l-79.

Turner, V. 1985. Colour Classification in Ndembu Ritual. In M.Banton (ed.). Anthropological Approaches to the Study ofReligion. London, Tavistock, 41 -84.