P07.11.207.032-P1 útg. 1 (21.01.2008 12:40) - Prentað 25.02.2008 14:44

DAVIDE ZORI

NAILS, RIVETS, AND CLENCH BOLTS: ACASE FOR TYPOLOGICAL CLARITY

This paper reevaluates the current terms and typology used for small functionaliron arlifacts from medieval nonhweslcm Europe and discusses the advantages ofa more uniform terminology for ulldcrslanding Ihe morphological and functionaldifferences between nails, rivets and cknch bolls. Examination of data collectedfrom the N<ltional Museum of Iceland illuSlrillCS some of the potcntialtypologi­cal difficulties that compromise c1assific31ion of iron artifacts. Examples fromEnglish- and Danish-language publications show th31lhcsc typological difficultiesarc a common problem across northwestern Europe. Focusing on Ihe inclusion ofclench bolts in Viking Age graves, the final section exemplifies the interpretivepotential of nuanced iron artifact type identification for yielding insighlS inlO awidespread mortuary prdetiee. Finally, the paper presents a revised terminology fornails. rivels. and clench bolls in English. Icelandic. and Danish.

Dln'ide 7AJri, Colsen IlI.willlle ojArchoeology.Uni>'ersily ojCaliiomia. Los Angeles, USA. Email: tf:ori@lIc1a.edll

Keywords: Middle Age.l: fce/and: iron artifact,': /)1iology: buriols

the importancemedieval society,

I. Introduction:Functional Iron ArtifactsNails, rivets, and clench bolts are notglamorous, They are rarely displayed asspecial items in museums, and mani­fest no great differences in form acrossregions and time. Often overlooked if notperceived as particularly important fora specific research question (Bill 1992:55), scholars have treated these function­al iron objects in a marginal fashion. Anexample of this marginalization is evidentin the comprehensive and seminal work,Viking Artifilcts: A Select Catalogue, inwhich James Graham-Campbell (1980)omits any individual treatment of nails,rivets, or clench bolts.

Consideringthese artifacts had in

ARCIIAEOLOGIA ISLANDlCA 6 (2007) 32-47

the scant mention of functional ironworkin the archaeological literature appearsunjustified. Nails, rivets, and clenchbolts, some of the simplest items pro+duced by the medieval ironsmith (Tyle­cote 1987: 262), were used in most con+struction contexts, They played a vitalrole in the fabrication of ships, houses,sheds, doors, roofs, carts, sledges,shields, boxes, coffins, and other objectsand structures made of wood, and are,as a result, frequently the most com­mon finds from medieval archaeologicalsites. In most excavation contexts, thesefunctional iron artifacts are the onlyremaining record of wooden objects orstructures, and therefore often have apivotal role in archaeological arguments(e.g. Skaaning 1992). The recording

P07.11.207.033-P1 útg. 1 (21.01.2008 12:45) - Prentað 25.02.2008 14:44

NAILS, RIVETS. ANI) CLENCII BOLTS: A CASE FOR TYPOLOGICAL CLARITY

and interpretation of their existence andrelative position is thus of considerableimportance for the reconstruction ofmedievallifeways.

With notable exceptions, suchas the work of Patrick Ottaway (1992),the limited attention devoted to function­al iron artifacts has resulted in the lackof a uniform terminology, thus obscur­ing the identity and interpretive valueof these artifact types. In an attempt toaddress this issue, [ begin by advancingspecific definitions of nails, rivets andclench bolts. Subsequently, my analy­sis employs a sample of artifacts fromthe National Museum of Iceland to testand elaborate on the criteria distinguish­ing the various artifact types. This datahighlights the interpretive complica­tions resulting from typological descrip­tive variability. Examination of severalEnglish- and Danish-language exam­ples, demonstrates that this typologicalvariability is a general phenomenon inmuseums and the international corpusof scholarly publications. [ then employthe example of clench bolts in mortuarycontexts from medieval northwesternEurope to illustrate the ways in whicha consistent terminology and a refinedunderstanding of artifact types can lendinsights into the past.

ItA Brief Introduction 10 DefinilionsBasic descriptive definitions of nails,rivets, and clench bolts are necessary inorder to frame the following discussionof functional iron artifacts. The form ofeach of these artifact types is constrainedprimarily by their function, a factor which

accounts for the similarities observedacross time and space; thus, these arti­fact definitions are widely applicable.Understood in the modern sense, a nailis "a small metal spike with a broadenedflat head, driven typically into wood witha hammer to join things together or toserve as a peg or hook" (Oxford Ameri­can Dictionary). A rivet is "a metal boltor pin having a head on one end, insertedthrough aligned holes in the pieces to bejoined and then hammered on the plainend so as to form a second head" (Ameri­can Heritage Dictionary). A clench boltis probably the most unfamiliar of thethree artifacts and is the only one madeup of two separate components: a nail anda rove. As Patrick Ottaway (1992: 615)explains, "[a] clench bolt was used forjoining overlapped timbers and consistsof a nail which, once passed through thetimbers to be joined, had a small piercedplate, the rove, set over its tip. The tipwas then burred or hammered over (Le.clenched) to hold the bolt in position."Characteristically, clench bolts are usedto join pieces of wood that either overlap,or are cut diagonally and then fit together(Figure I).

III. Artifacl Identification andthe Need for Typological Consistency:A Case Study from the NationalMuseum of IcelandThe artifact analysis in this section isbased on data that [ collected during thesummer of2003 at the National Museumof Iceland (l>j66minjasafn islands)1 whiledoing comparative research for findsfrom the Mosfell Archaeological Project

, This research was made possible by lhe generous a..iSlanee nf Halld"111 Asgeirsdl>nir, cnnservalOr al the Nalinna! Museumof Iceland and the help and guidance of Gu<'Jmundur 6lafsson. Head of lhe Archaeological Dcpanmem allhe Naliona!Museum.

33

P07.11.207.034-P1 útg. 1 (21.01.2008 12:38) - Prentað 25.02.2008 14:44

DAVIDE ZORI

I

l~. II~

IV

~Figure /, Clench bolts employed in 111"0 mel­hods for joining planks (ret/rawn by JennieDi/fonjrom Ouaway /992: 6/7)

(MAP).2 This work led me to revisit theterminology currently in use for nails,rivets and clench bolts, and the impor­tance of typological consistency forresearchers attempting to understand theoriginal function and use of such arti­facts.

The creation of archaeologicaldata is an on-going process that beginswith labeling find bags in the field, andcontinues with cataloging at the projectlevel and subsequent storage at regionalor national museums. Particularly at theselevels, a collection-wide database facili­tates access to data. Such a database isa living tool, the creation of which is,by nature, a cooperative task involving

many individuals and projects. Scholarlyuse of a database can be an opportunityfor reflection and adjustment in orderto make the database more amenable tofuture research. In this paper, I reflecton possible classification adjustments toarchived archaeological data concerningsmall functional iron artifacts.

Working with all catalog entriesin the National Museum pertaining tofunctional ironwork, I examined the arti­fact types ofnails, rivets, and clench boltspresent in each catalog category. TheNational Museum has created an exten­sive electronic database that includes fullcatalog entries for all the finds in thearchives. This user~friendly database iselectronically searchable across severalfields such as find material or identity,and easily generates complete lists offinds cataloged under those descriptions.From each of the find entries, the actualfind was easily found in the storage facili­ties of the museum. This database and theefficient access to the finds greatly facili­tate archival artifact research. The issuefor this paper was that each of the threeiron artifacts types (nails, rivets, clenchbolts) forming the basis ofthis study, werefound to be cataloged under variable cate­gory names that were not mutually exclu­sive. Because of this unintentional typo­logical variability, an electronic search ofthe database for a specific artifact typedid not yield a comprehensive list of theappropriate artifacts in the collections.This limited the usability of the data-sys­tem in searching for specific iron artifacttypes and meant that identification of theartifact types often required reference to

'MAP is an in1emalional archaeological project conducting interdisciplinary research in the Mosfell region in oouth­weS1 Iceland. MAr. directed by Jesse Byock and Phillip Walker. "mh in rollaoom1ion with I>j6dminjasafn Islands andMosfcllsba:r.

34

P07.11.207.035-P1 útg. 2 (21.01.2008 15:54) - Prentað 25.02.2008 14:44

NAILS. RIVETS. ANI) CLENCII BOLTS: A CASE 1'011. TYPOLOGICAL CLARITY

the more lengthy descriptions of artifactmorphology in the catalog or detailedstudy of the actual artifact.

Specific details will illustrate theissue ofdescriptive consistency. The Icelan­dic tenns used to classify nails, rivets, andclench bolts arc 'bolao5alllllll,.' (boat nails),'1II1oollagli' (riveting-nail), '/IIlOOSalllllllr'(riveting-nail), 'nagli' (nail), and '/Y>nagli'(nut- or rove- nail). Although ostcnsiblydefined and differentiated by their formalmeaning, thcse terms appear in the Ice­landic archives as overlapping categories.One ramification of this is the catalogingofclench bolts In all of the aforementionedcategories, 'batusaulIlur,' '!IIl()()nagli,''ImorJsuuml/r,' '"ugli,' and 'ronagli.'

In Archaeological Typology andPractical Realily: a Dialectical Approach10 Artij(lcf Classijicaiioll and Sorting(1991), William and Ernest Adams stressthat a uscfultypology must consist of dis­crete artifact types, identifiable by diag­nostic features. There should bc no pos­sibility that an artifact belongs to morethan one type.) Cataloging an artifact thatcannot be grouped into any of the typeswithin a typology should result in thecreation of a new typc that strcsscs theunique nature of the artifact. Adams andAdams (1991: Ch. 4) assert that a consist­ent system of classification provides thefoundation for artifact analysis, allowingfor the quantification and subscquent sta­tistical analysis of artifact types. The ter­minology uscd must consistently respectthe morphological differences exhibitedby the various artifact types and distin­guish the unique functions that can beinferred from the morphology. In an iron

collection with such similar objects asnails, rivcts, and clench bolts, a strict ter­minological protocol would help to limitobserver bias. The widespread lack ofattention to distinguishing between nails,rivets, and clench bolts has not led to theadoption of such a protocol, and there­fore, it can be suspected that collectionsof small functional ironwork, in general,have not been organized in such a waythat would meet the prerequisites for auseful typology. The typological descrip­tive variability observed in the Icelandiccollection support this hypothesis. A con­sistent typology, therefore, must underliefurther scientific study of these iron arti­facts.

In order to evaluate the artifacttypes that were included in each category,a total of 44 record entries in the NationalMuseum of Iceland were examined: I)all artifacts labeled 'ronagfi' 2) all arti­facts labeled 'hnoJllagli' 3) all artifactslabeled 'bcitaS(llllJlllr' 4) a selection of theartifacts labeled 'llagli.' Some entries, asin the boat burials, contained as many as500 individual artifacts. If the actual arti­facts were available, they were examinedto dcterminc their idcntity. If the artifactswere unavailable, I attempted to identifythem as nails, rivets, or clench bolts bytheir physical description and find con­text, as recorded in the catalog.

Criteria Usedfor the ldemijicalionofArtifacl 7j'Pe.\·The 44 artifact entries from the NationalMuseum were evaluated following thecriteria for identification of nails, rivetsand clench bolts outlined here.

j Ollen it may be impossible to S<."<:urely determine the arlifacl type bec.use the anifaet may be broken or bec.use of ironcorrosion obscuring the morphology Oflhc find. In such a case, x-raying the anifact may be able to re\"eal more "eeumtclythe original form of the find.

35

P07.11.207.036-P1 útg. 1 (21.01.2008 12:42) - Prentað 25.02.2008 14:44

DAVIDE ZORI

Figure 1. 1i1'o iron nails. The black bar is ollecelllimeler long. (pllOlOgraph taken by Iheallfhor at fhe National Museum ofIceland)

Artifacts having a pointed shankwith a single circular or square head wereidentified as nails (Figure 2). Broken riv·ets and clench bolts are easily confusedwith nails, and therefore it is often moredifficult to positively identify a nail thanit is to identify either of the two other arti­fact types.

Rivets have two heads connec­ted by a shank that does not extendpast either of the heads (Figure 3). Theheads of a rivet are usually rounded, butmay also have an irregular shape due totheir deformation from flallening witha hammer. [n working with the wriuencatalog entries of the National Museum,I classified an artifact as a rivet if it wasdescribed as having two heads withoutthe mention ofa 1'0 (rove). The find con­texts are particularly important for distin­guishing rivets from clench bolts, since,in the Middle Ages, rivets were used in avariety of objects, often holding togethertwo sheets of iron, whereas clench boltswere generally used to hold together twoor more overlapping pieces of wood.

Clench bolts characteristicallyhave a round head on one end and a dia­mond- or square-shaped head, or rove, onthe other end. The artifact commonly has

36

Figure 3. Rirer. No/e Iha//he shank does /lorprolmde from eilher head. The black bar isone centimeter long. (photograph taken by theamhor at the National Museum ofIceland)

a piece of the shank extending throughthe hole in the diamond- or square-shapedrove, sometimes protruding several mil­limeters (Figure 4). When direct exami­nation was nOI possible, I focused on thedescription of the heads or roves of theartifacts in order to identify clench bolts.[n Icelandic, 1'0 (plural rrer) means 'rove'or 'nut.' When ,.0 is used to refer to oneend of the shank instead of hailS ('head,'plural hal/sal') then the artifacts are mostlikely clench bolts, since neither nails norrivets have nuts or roves. This method ofclassification is not completely satisfac­tory because the term 'ronagli' is usedoccasionally to describe a rivet. It is pos­sible, therefore, that the conflation of thetwo types of artifacts (rivets and clenchbolts) could have been carried over tothe smaller constituent parts of rivets andclench bolts, plausibly resulting in heads(hausar) being referred to as roves (rrer).Because of this possibility, entries wereconsidered inconclusive unless the 1'6 wasdescribed as Jerkantaoar or Jerhymda('square,' literally 'four-sided' and 'four­cornered') or better still, tigulmynduJ('diamond-shaped').

P07.11.207.037-P1 útg. 1 (21.01.2008 12:44) - Prentað 25.02.2008 14:44

NAILS, RIVETS. ANI> CLENCII BOLTS: A CASE fOR TYPOLOGICAL CLARITY

-Figure 4. Clench bolfS alld a /'01·e. Note that the ends 0/ fhe shanks profrude from the roves 0/fhe{\I'O complete clench bolts. The black bar is olle centimeter long. (photograph takell by the ali/hoI'at fhe National Museum 0/Iceland)

ResullS ofArtifacildemificationThe following section presents and dis­cusses the types of artifacts identifiedunder each of the following current data­base categories: 'nagli,' 'hnoonagli,' and'r6nagli.' I identified the actual artifacttypes in each of the entries, noting thepresence/absence of nails, rivets, and/orclench bolts. Since an individual catalogentry often contains numerous artifactsand frequently more than one artifacttype, the number of identified artifacttypes typically exceeds the number ofactual database entries. The results aredepicted in the pie graphs (Figure 5),which calculate the percentage that eachartifact type (nail, rivet, clench bolt) rep­resents out of the total number of typeswithin each of the National Museum'sdatabase categories ('nagli,' hnoonagli,'and '"onagli') .

A sample of twelve 'nagli'('nail') entries was investigated to deter­mine whether they also included rivets orclench bolts. Nine 'nagli' entries werecomposed solely of nails. Two entriescontained artifacts that could not conclu­sively be assigned to a particular type.One entry (National Museum catalog # B:1939-88) included a single whole clenchbolt and eight broken artifacts that mayhave been either nails or clench bolts (seeFigure 5). The 'nagli' entry that includeda clench bolt consisted of the iron remainsfrom a coffin. The description of the ironfrom this coffin burial begins, "nails withwood remains on them, eight in number,however only one whole."4 The entry goeson to address the single whole nagl": "theonly whole one is 3.5 cm long and has around head at one end but a square rove atthe other."s The description contained in

• -[Nlaglar me.... vioorleifum a. 8 talsins.1><\ a<kins einn heill"l "Sa eini. scm heill cr. CT 3.5 em a<} I. og hefur kringl6nan haus a oorum enda. en ferhyrnda r6 a hinum"

37

P07.11.207.038-P1 útg. 1 (21.01.2008 12:39) - Prentað 25.02.2008 14:44

rivets0<>/0

clench bolts8 0

/0

DAVIDE ZORI

'Nagli'unknown

170/0

'Hno6nagli'

clench bolts60%

unknown0 0

/0

rivets300

/0

clench bolts61 0

/0

'R6nagli'unknown

190/0

nails8'/0

• nails

D rivets

• clench bolts

D unknown

Figure 5. Pie charts showing Ihe percentages of artifact types (nail, ril'et, and clench bolt) Ihatappear in Ihe categories of'rollagli: 'nagli: and 'Imoonagli' ill National Museum ofIceland.

38

P07.11.207.039-P1 útg. 1 (21.01.2008 12:42) - Prentað 25.02.2008 14:44

NAILS, RIVETS. ANI) CLENCII BOLTS: A CASE FOR TYPOLOGICAL CLARITY

this entry suggests that all of the supposednaglar (nails) in this context may haveoriginally been clench bolts. The entrywriter also notes as a concluding remark,"it seems that riveting did take place tosome degree in the making of the coffin."/> The hnoonegling (,riveting') in thisinstance refers to the functional action ofa clench bolt rather than a rivet. Thisexample demonstrates that even though'nagli 'may be one of the simplest artifactcategories in the classification system,referential inconsistency still occurs with­in the category.

The other t\vo categories, 'hnoo­nagli' and 'ronagli,' yielded even lessconsistency than the 'nagli' category(see Figure 5). There are seven entriesin the National Museum's databaselabeled 'hnoonagli' ('rivet'). Of the seven'hnoonagli' entries, three contain rivets,six contain clench bolts, and one has anail. Out of the 24 entries labeled 'rona­gli' ('rove-nail') in the catalog, 16 containclench bolts, three have rivets, two includenails, and five entries contain artifactsthat could not be securely identified.

The Effect ojTypologicallnconsislencyThe analysis above illustrates that the cate­gorization of iron artifacts in the databasedoes not consistently follow a typologyderived from the functional and morpho­logical characteristics of the individual arti­facts. The two particularly pervasive tenni­nological problems are the result of diamet­ric typological practices. The first problemis the use ofa single tenn to apply to severalartifact types ('lumping'). The second prob­lem is the use of several different tenns torefer to the same artifact type ('splitting').Ultimately, both problems complicate com-

parative and statistical analyses that rely ontenninological unifonnity.

The pie charts (Figure 5) docu­ment the extent of the typological prob­lems in the database. Each of the individ­ual pie charts reveals that a single termincludes two or more artifact types thathave been lumped together. For instance,the term 'hnoonagli' is used for the nail,rivet and clench bolt artifact types. Takentogether, the three pie charts show thatseveral different terms are used to identifyeach individual artifact type. For exam­ple, the three tenns 'nagli,' 'hnoonagli,'and 'ronagli' are all applied to a singleartifact type, the clench bolt.

The finds from boat burials illus·trate that the two typological problemsoutlined above also appear within a singlekind of archaeological context. In general,boat burials typically contain both nailsand clench bolts, which partially explainswhy there is little consistency in the cat­egorization of iron from boat burials. Theiron hardware from one boat burial in theNational Museum is cataloged as 'Imoo­nagli' (B: 1939: 71), whereas the iron arti­facts from two others are cataloged under'ronagli' (B: 1937-73 and B: 1964-110­1). Hardware from a fourth boat burial iscataloged under'balasaufIlur' (biltur=boat,

Sal/fIlllr= nail; B: 1946-53). Catalogingiron from a boat burial as 'boat nails' isunderstandable as an attempt to avoidmisidentifying some of the many artifactsin the entry with tcnns such as 'dmagli'or 'hnoOnagli,' but in practice the use of'biltasaulllur' relies on the find context forcataloging rather than artifact morphologyand thus only increases the typological con­fusion. Each individual catalog entry fromthe boat burials lumps multiple artifact

• "Viroisl hnooncslins !lVi a" cinhvclju Icyli hafa all seT Sla" i likkislusmi"inni"

39

P07.11.207.040-P1 útg. 1 (21.01.2008 12:42) - Prentað 25.02.2008 14:44

DAVIDE ZORI

types (nails and clench bolts) under a sin­gle term, an example of the first typologi­cal problem. As a group, the boat burialsinclude Iwo artifact types (nails and clenchbolts) that are each described by three dif­ferenl terms ('hnoonagli', 'ronagli' and'b/t((lsaumur'), thus exemplifying the sec­ond typological problem.

Typological inconsistencies inthe catalog seriously affect archaeologicalresearch by obscuring both the morpho­logical differences between the artifacttypes and the functional consistency with­in each of the types. This results in dif­ficulty and potential inaccuracy in the sta­tistical analysis of cultural materials. Theproblem would be solved by the applica­tion of a protocol using strictly consistentterminology that recognizes the distinc­tions between nails, rivets, and clenchbolts. Based on the preceding section, it isrecommended that different artifact typesfrom the same find context be catalogedseparately. Furthermore, the term 'nagli'should be reserved just for nails, 'llnoona­gli' for rivets, and 'ronagli' solely forclench bolts. Finally, 'bittasaumllr' oughtto be eliminated as a typological categorybecause it contlates objects of differentfunctions-nails and clench bolts-there­by violating one of the essential aspectsof a useful typology (Adams and Adams1991) and adding to the overall state ofclassificatory confusion.

A more uniform typology willassist in the statistical analysis of nails,rivets, and clench bolts and the compara­tive study of their find contexts, ultimate­ly contributing to a more nuanced under­standing of daily life in early Iceland.The National Museum of Iceland is in anideal position to serve as a pioneer among

1Danish,j<>m"'iron, nuglc"'nails1ri,ets/c1cnch bolls

40

museums in catalog consistency becauseof the manageable size of the nationalcollection and quality of the electronicdatabase that has already been compiled.

IV. Widespread TypologicalInconsistenciesTypological inconsistencies in the categori­zation of small functional ironwork are notan isolated phenomenon, but a widespreadterminological problem that exists in manylanguages and appears in other museumcollections, as well as numerous archaeo­logical publications and site reports. Thissection provides examples of typologicalinconsistencies from museum collections,site reports, and well-known publications inEnglish and Danish in order to demonstratethe relevance of Ihe argument in the widergeographic sphere of the North Atlantic.The identical two terminological problemsfound in the catalog at the National Muse­um are also pervasive in the sources ana­lyzed in this section. To reiterate, the firstterminological problem obscures differ­ences in artifact morphology by lumpingmore than one distinct artifact type undera single term, whereas the second problemcompromises the integrity of typologicaltenns by applying more than one tenn to asingle artifact type.

Archaeological publications fre­quently exhibit the first terminologicalproblem by using a single term such as'nail' in English, 'nagli' in Icelandic, or'nagle' in Danish 10 refer to all three typesof artifacts-nails, rivets and clenchbolts-without distinguishing betweenthem. In a Danish example from the pub­lication Trelleborg, Poul Norlund (1948)used the term 'jemnagler'7 to refer to agroup ofartifacts that he illustrated with a

P07.11.207.041-P1 útg. 1 (21.01.2008 12:38) - Prentað 25.02.2008 14:44

NAILS. RIVETS. ANI) CLENCII BOLTS: A CASE 1'011. TYPOLOGICAL CLARITY

photograph that shows only nails. A photo­graph published in 1991 shows that theartifacts that Norlund designated as 'jem­nagler' actually include roves and clenchbolts (Nielsen 1991: 132). The NationalMuseum of Denmark in Copenhagen con­tains an unpublished catalog of artifactsfrom Trelleborg which also uses the inclu­sive term 'llagler,' thereby perpetuatingthe terminological simplification. In anEnglish language example from the publi­cation Birka IV, Anne-Sophie Griislund(1980: 8) separates nails from rivets andclench bolts, but employs the term 'rivet'to refer to both rivets and clench bolts.This first type of typological inconsisten­cy, in which several artifact types aregrouped into the same category, obscuresthe distinct morphologies of the diITerentartifact types and renders it impossible toascertain the actual identity of the artifactswithout direct observation.

The second terminological prob­lem, in which several terms arc used todescribe a single artifact type, appears inthe publication Fyrkat II, in which ElseRoesdahl (1977) uses the Danish term'som' for nails, but applies two terms,'nagler' and occasionally 'klinknagler," inreference to clench bolts (e.g. Roesdahl1977: 85). In Danish, the term 'llagle' isusually used to refer to a rivet, but becauseRoesdahl employs this term 10 refer toclench bolts, no distinct term remains fora rivet. The practice of applying severalterms to the same artifact type leads to theimplication that identical artifacts are dif­ferent, as well as the identification ofarti­fact types with certain terms that arc moreappropriate for describing other artifacts.

As a whole, English-languagearchaeological publications exhibit the

second terminological problem by usingthe terms 'rivet,' 'clench nail,' and 'clenchbolt' to refer to the single clench boltartifact type. In general, when writing inEnglish, Scandinavian scholars tend torefer to clench bolts as rivets (e.g. Bill1994; Christensen 2002; Griislund 1980;Birkedahl & Johansen 1994). Using theterm 'rivet' to refer to a clench bolt isproblematic because it does not distin­guish between the presence and absenceof a rove and thereby fails 10 provideexact terms for different artifacts. Brit­ish scholars, however, commonly referto clench bolts as 'clench nails' (e.g.McGrail 2004; Richards 1991: lIS;Carver 1995). A clench nail is a nail thathas been bent to hold in place. The tech­nique of clenching a nail to secure tim­bers without the use of a rove was alsocommonly employed in the Middle Ages(McGrail 2004; Christensen 2002) and assuch is more appropriately considered asanother functional usc of the nail artifacttype. The third term referring to the nailand rove combination is 'clench bolt,'favored by the York Archaeological Trustand associated researchers such as PatrickOttaway (1992). The term 'clench bolt'has the advantage of indicating the bolt­ing function and two part composition ofthe artifact type, as well as stressing theunique nature of the artifact type as sepa­rate from nails, rivets or bent nails.

The terms 'rivet' and 'clench'have been the source of much terminolog­ical confusion and inconsistency, whichultimately derives from the facl that thetwo words entered modern English aswords of similar meaning from the OldFrench river and Middle English cleynch(McGrail 2004: ISO). The tenninologi-

• Klinkn"g/t' is a Danish I",m lha1 ref",.,; 10 clench bolls employed in the clinker slyle eon>1ruetion of ",.."lapping "'oodenplanks ch"r~Clcristic of Norse ships

41

P07.11.207.042-P1 útg. 1 (21.01.2008 12:37) - Prentað 25.02.2008 14:44

DAVIUE ZORI

cal debate is still lively, as evidenced bythe recently published exchange betweenSean McGrail (2004) and Arne-EmilChristensen (2002), in which they respec­tively provide reasons why 'rivet' and'clench nail' are inappropriate terms. Inlight of the problems inherent in 'rivet'and 'clench nail,' it is my guess that, whenthe dust settles, the term clench bolt willbe found to be the most morphologicallyspecific and typologically useful.

V. The Interpretive Value of a NuancedTypology: An Example from VikingAge Mortuary ContextsThe analytical benefits and interpretivepotential of employing a tenninologi.cally accurate and consistent typologythat recognizes the ditTerences betweennails, rivets, and clench bolts justifiesthe requisite etTort needed to establishsuch a typology. The foundation for theinterpretive value lies in understandingthe three artifact types as distinct fromeach other in morphology and function.Focusing on clench bolts in Viking Agegraves, this section argues that preciseartifact identification and careful analysisof the unique functions of these artifactscan yield new insights into a widespreadmortuary practice. Examples from cem­eteries across northwestern Europe showthat the phenomenon of clench bolts inburials has significance beyond the func­tional level and suggest ritual continuitywith the pagan practice of boat burial.

Burials with clench bolts, oftenarrayed in rows, appear in both paganand Christian cemeteries from the Mid­dle Ages across the wider North Searegion, including Fyrkat and Sebbersundin Denmark, Hrfsbru and HOfstaoir inIceland, Birka in Sweden, and Caistor-onSea, Ingleby, Thorpe-by-Norwich, York

42

Minster, and Barton-on-Humber in east­ern England (Birkedahl and Johansen1995: 162-163; Byock et a1. 2005;Carver 1992: 110; Gestsdottir 2006: 12;Griislund 1980: 24; Richards 1991: 115;Roesdahl 1977: 84, 113;). Clench boltsin burials are unlikely to derive from cof­fins, which are more easily built withnails. In contrast, the construction of abox with clench bolts is cumbersome andinefficient, necessitating more iron andgreater labor investment. Clench boltsjoin overlapping planks, but cannot fas­ten together perpendicular planks withoutthe use of angled mounts for the cornerjoints (Figure 6). No angled mounts havebeen found in any of the graves contain­ing clench bolts. The clench bolts mustbe the remnants of an object interred inthese graves other than a traditional box­shaped coffin. This argument has beensupported through accurate identificationand careful recording of the placementof rows of clench bolts in one grave atthe 9th_lOth century cemetery at Birkaon Bjorko, Sweden and two graves at the101h century cemetery at Fyrkat in Jut­land, Denmark. The analysis of the hard­ware from these three graves allowed forthe reconstruction of the original shapeof a clinker built wooden object and theresultant identification of a type of burialin which the deceased was interred in awooden cart (Gdislund 1980: 24; Roes­dahl 1977: 84, 113).

A number of burials with a dis­tribution of clench bolts different fromthat observed in the wooden cart burialsindicate the inclusion of another type ofclinker built object in graves. Roesdahl(1977: Ill) observed that a burial atFyrkat containing three jumbled rows ofclench bolts was different from the cartburials. The clench bolts had remnants of

P07.11.207.043-P1 útg. 1 (21.01.2008 12:36) - Prentað 25.02.2008 14:44

NAILS, RIVETS. ANI> CLENCII BOLTS: A CASE fOR TYPOLOGICAL CLARITY

= -

IIII

I~ -

1) simple box constructedwith clench bolts

2) simple box constructedwith nails

Figure 6. Schematic drawings ofbox construction /Ising clench bolts and nails

wood around them and clenched a lengthof 12-26 mm between the head and therove. In Iceland, at the Hrisbni cemeterydating from the late IOth_ I Ith centu­ries, a similar burial with three clearlydefined lines of clench bolts and twoadditional burials with numerous jum­bled clench bolts were uncovered in 2002and 2004 (Byock et at. 2005a; Byock etaL 2005b). The majority of these clenchbolts, clenching a length between 10 and20 mm, were similar in length to thoseuncovered at Fyrkat. It seems likely thatthe clench bolts in these graves representpieces of boats. The lengths of the clenchbolts at Fyrkat and Hrisbni are shorterthan would be expected for the planks ofa clinker built ship, but fit more closely toboards from a small boat.

The interpretation of boats orpieces of boats interred with the deceasedis supported by evidence from the 8lh_9th

century cemetery of Caister-on-Sea in

eastern England, where 12 out of the 150inhumation burials contained clench bolts(Carver 1995: 116-117). Martin Carver(1995: 117) observes that none of theseburials contain a whole boat, but that the"riveted timbers were more likely to havederived from pieces of clinker-built boatsthan any other source." Also, at the exca­vation at Sebbersund, an 8th_12th centurycoastal trading site in northern Jutlandin Denmark, Peter Birkedahl and ErikJohansen (1995) argue that the clenchbolts found in 41 out of 468 excavatedgraves are remnants of broken boat partsused by poor fishermen.

The important question iswhether the deposition of boat pieces ingraves served a purely functional role,such as Birkedahl and Johansen suggest(1995: 162-163), or whether the practicehad wider symbolic meaning for the indi­viduals burying their dead. Considering,for example, the postulated poverty of the

4)

P07.11.207.044-P1 útg. 1 (21.01.2008 12:43) - Prentað 25.02.2008 14:44

DAVIUE ZORI

fishermen at Sebbersund, it seems oddthat people did not extract the iron clenchbolts from the planks for resmelting, ashas been observed elsewhere (Birkedahland Johansen 1995; Fridriksson and Her­manns-Audard6ttir (992). Rather, theinvestment of energy into these burialsrepresented by the deposition of clenchbolts suggests a purposeful symbolicstatement (see e.g. Sinclair 1995: 55).

Furthermore, in all but one ofthe burials at Caistor·on-Sea, the clenchbolts were found on top of the skeletons,just as in the three burials at Hrisbru, Ice­land. The reused boat planks thereforedid not function as biers or cotTins, butrather as overlying covers or even as non­functional objects placed into the grave.Although coffin lids serve a practicalpurpose, Julian Richards (1991: (15)asserts that "this is really too mundane anexplanation. Given the Scandinavian tra·dition of ship burial it seems reasonablethat the symbolism of the boats' timberswas intentional."

In the North Atlantic culturalarea, the boat had great symbolic sig­nificance, beginning in the early [ron Ageand continuing throughout the VikingAge (Crumlin-Pedersen and Thye 1992).In this larger temporal framework, OleCrumlin-Pedersen (1992) views buri­als with parts of boats as belonging tothc same tradition as larger ship burials(Oseberg, Ladby, Sutton Hoo) and stoneship settings (Linholm H0je, Jelling, AlesStones) that are collectively a religiousreference to the Norse and Germanic godFrey's ship, Skioblaonir (Christensen et al.1992; Carver 1992; Green 1968; Krogh1983; Ramskou 1976; S0rensen (997). [agree with Richards (1991) that the inclu­sion of boat timbers with clench bolts inburial contexts was meant to convey a

44

symbolic message related to the ship-bur­ial tradition, but would push the interpre­tation further and stress that the boat partswere meant to convey the same messageas the whole ship, only at a different scaleand more affordable cost. As noted by Ric­hards (1991), this interprctation has widerimplications for ritual continuation of theship burial tradition into Christian cem­eteries, but also for the wider theoreticalpossibility that the presence of the parI isequivalent to the whole in ritual contexts.

The identification of the uniquecorpus of graves with clench bolts andtheir subsequent interpretation as sym­bolic boat burials was only possiblebecause of careful documentation anddifferentiation of the clench bolt artifacttype. There is little doubt that misidentifi­cation of iron artifacts has obscured otherexamples of clench bolts in burials andthat this practice of including portionsof boats in Viking Age graves is morewidespread than currently appreciated.In order for extant or future examples tobe incorporated into our understanding ofmedieval ritual and symbolic practices, aunified terminology differentiating theclench bolt artifact typc is a necessity.

VI. ConclusionI argue throughout this paper that nails,rivets, and clench bolts are significantartifact categories for archaeologicalanalysis. Currently, widespread typologi­cal inconsistencies exist in scholarly workconcerning these three artifact types,hindering type identification, statisticalanalysis, and comparative research. Nails,rivets, and clench bolts have unique mor­phologies and functions, which dictatetheir use in specific situations. Examin­ing the presence of the particular ironartifact types in various archaeological

P07.11.207.045-P1 útg. 1 (21.01.2008 12:42) - Prentað 25.02.2008 14:44

NAILS, RIVETS. ANI) CLENCII BOLTS: A CASE FOR TYPOLOGICAL CLARITY

contexts has great potential for illuminat­ing medieval use-patterns and enrichingour understanding of widespread culturaland ritual practices. The study of clenchbolts in burials illustrated this potential bysupporting an interpretation of the clenchbolts as an expression of the Scandina­vian boat burial practice. The interpretivepotential of nails, rivets and clench boltswill only fully be realized with a cleartypology and a consistently applied ter­minology that recognizes the differencesbetween the three artifact types. I there­fore suggest that the following terms beused in English, Icelandic, and Danish:

I) 'nail' (English) = 'nagli' (Icelandic) ='som' (Danish)

2) 'rivet' (English) = 'hlloonagli' (Ice­landic) = 'nagle' (Danish)

3) 'clench bolt' (English) = 'r6nagli'(Icelandic) = 'klinknagle' (Danish)

Acknowledgments[ am grateful for the support from the USNational Science Foundation, the town ofMosfellsbrer, the Mosfell ArchaeologicalProject (MAP), and the UCLA Friends ofArchaeology, which made my researchpossible. The staffat the National Museumof Iceland, particularly Guomundur 6laf­sson and Halld6ra Asgeirsd6ttir enabledme to analyze the Icelandic artifacts andgenerously offered their help, advice, andfruitful discussions. I am grateful to Jen­nie Dillon for redrawing Figure I. I par­ticularly thank Jesse Byock for his men­toring and guidance and Colleen Donleyfor her patience and editorial skills.

ReferencesAdams, William & Ernest Adams. (1991)

Archaeological "Typology and Practi­cal Reality: A Dialectical Approach

45

10 Artifact Classification and Sort­ing. Cambridge: Cambridge Univer­sity Press.

Appadurai, Arjun. (1986) "Introduction:Commodities and the Politics ofValue". In Appadurai, A. (ed.) TheSocial Life of Things: Commoditiesin ClIlfllral Perspec/ive. Cambridge:Cambridge University Press.

Bill, Jan. (1994) "Iron Nails in Iron Ageand Medieval Shipbuilding". In West­erdahl, C. (ed.) Crossroads in AncientShipbuilding: Proceedings of theSixth International Symposium onBoat and ship Archaeology, Roskilde1991. Oxford: Oxbow Books.

Birkedahl, Peter & Erik Johansen. (1994)"The Sebbersund Boat-graves". InCrumlin-Pedersen, O. & Thye B.(eds.) The Ship as Symbol in Pre­historic and Medieval Scandinavia.Copenhagen: Publications for theNational Museum of Denmark.

Byock, Jesse, Phillip Walker, Jon Erland­son, Per Holck, Davide Zori, MagnusGu6mundsson, & Mark Tveskov.(2005a) "A Viking Age Valley inIceland: The Mosfell ArchaeologicalProject". Medieval Archaeology 49:196-220.

Byock, Jesse, Phillip Walker, MariannaBelli, Jon Erlandson, Anne Enges­veen, Magnus Gu6mundsson, Mar­grel Hallsd6ttir, Magnus Hellqvist,Per Holck, Steinar Kristensen, Dag­fin Skre, & Davide Zori. (2005b)The Mos.fell Archaeological Pmjecl.2004: Excavation at Hrisbru. Mos­fellssveit. Unpublished report sub­mitted to Fornleifavernd Rikisins.

Carver, Martin. (1995) "Boat-burial inBritain: Ancient Custom or PoliticalSignal". In Crumlin-Pedersen, O. andThye, B. (eds.) The Ship as Symbol

P07.11.207.046-P1 útg. 1 (21.01.2008 12:39) - Prentað 25.02.2008 14:44

DAVIUE ZORI

i/l Prehistoric and Medieval Scandi­navia. Copenhagen: Publications forthe National Museum of Denmark.

Carver, Martin. (1992) "The Anglo­Saxon Cemetery at Sulton Hoo: AnInterim Report". In Carver, M. (ed.)The Age ofSulton Hoo: The SeventhCeil/lilY ill North-Westem Europe.Woodbridge.

Christensen, Arne Emil. (2002) "Reviewof Bout~· of the World" by McGrail,Sean.lntematiollul Journal ofNuuti­cal Archaeology 31/2: 300-3 I I .

Christensen, Arne Emil, Anne Stine Ing­stad, & Bjorn Myhre (eds.). (1992)Oseberg-DIVI/flillgens Grav: V(irArkeologiske Nwjollalskalf i NyttLy-~. Oslo: Schibsted.

Crumlin-Pedersen, Ole. (1995) "80at­burials at Slusegaard and the Inter­pretation of the Boat-grave Custom".In Crumlin-Pedersen, O. & Thye, B.(eds.) The Ship as Symbol ill Pre­historic alld Medieval Scalldinavia.Copenhagen: Publications for theNational Museum of Denmark.

Fridriksson, Thorbjorn & Margrct Her­manns-Audard6ttir. (1992) "lron­making in Iceland". Espelund, A.(ed.) BlOOIlIeI)1 lrollmakillg During2000 Years. Seminar in Budalen, Sor­Trondelag.

Gestsd6ttir, Hildur (ed.). (2006) Ho.ifiatJir2004: Illterim Report. Unpublishedreports of the Institute of Archaeol­ogy, Iceland: FS31 [-9101 [2.

Graham-Campbell, James. (1980) VikingArtifacts: A Select Catalogue. Lon­don: British Museum PublicationsLtd.

Green, Charles. (1968) Sulton Hoo: TheExcavatioll of a Royal Ship-Burial.London: Merlin Press.

Griislund, Anne-Sofie. (1980) Birka

46

tv. the Burial Customs: A s/lldy ofthe graves 011 Bj6rk6. Stockholm:Almqvist & Wiksell International.

Krogh, Knud. (1983) The Royal Viking­Age Monumets at Jelling in the Lightof Recent Archaeological Excava­tions. A Preliminary Report. ActaAl'chaeologica 53: 183-216.

Maddin, Robert. (1985) The Early Black­smith. In Scon, B. G. & Cleere, H(eds.) The Crafts of the Blacksmith.Essays presentcd to R. E Ty[ecoteat 1984 Symposium of the UISPPComite pour la Siderurgie Anciennein Belfast.

Mainman, Ailsa J. & Nicola S.H. Rogers.(2000) Craft, Indusoy. ulld EvelydayLife: Finds from Anglo-Scandina­vian York. York: Council for BritishArchaeology.

McGrail, Sean. (2004) "To Clench orto Rivet: That is the Question". TheImerna/ional JOl/rnal of NalllicalArclweology 33/1: 149-153.

Nielsen, Leif Christian. (1991) "Trelle­borg". Arborg for Nordisk Old­kyndighed og Historic: Copenhagen.

Norlund, Poul. (1948) Trelleborg. Copen­hagen: Oct Kongeligc Nordiskc O[d­skriflselskab.

Olsen, Olaf & Holger Schmidt. (1977)Fyrkat en Jysk Vikingeborg: I. Bor­gen og Bebyggelsen. Copenhagen:Oct Kongcligc Nordiske Oldskrift­selskab.

Onaway, Patrick. (1992) Anglo-Scandi­navian Ironwork from Coppergate.York: Council for British Archaeo­logy.

Ramskou, Thorkild. ([ 976) LindholmHojc Gravpladsen. Kobenhavn: OctKonge[ige Nordiske O[dskriftselkab.

Richards, Julian. (1991) Viking Age Eng­land. London: B.T. Batsford Ltd.

P07.11.207.047-P1 útg. 1 (21.01.2008 12:44) - Prentað 25.02.2008 14:44

NAILS. RIVETS. ANI) CLENCII BOLTS: A CASE 1'011. TYPOLOGICAL CLARITY

Roesdahl, Else. (1977) Fyrkat En JyskVikingeborg: II. Oldsagerne og Grav·pladsen. Copenhagen: Det KongeligeNordiske Oldskriftselskab.

Schmidt, Holger. (1999) VikingetidensByggeskik i Danmark. Moesgard:Jysk Arkrelogisk Sclskab.

Sinclair, Anthony. (1995) "The Techniqueas a Symbol in Late Glacial Europe".World Archaeology, Vol. 27, No. I:50-62.

Skaaning, Poul. (1991) Harald BlatandsSidste Kamp: Trelleborgene og Bryd­ningerne i Dansk Vikingetid 979-983.Lyngby: Dansk Historisk Handbogs­forlag.

S0rensen, Anne Christina. (1997) Ladby:Ship, Cemetery and Settlement. In

47

Nielsen K. & Jensen, C. (eds.) Bur­ial and Society: The Chronologicaland Social Analysis of Archaeologi­cal Burial Data. Aarhus UniversityPress.

Tylecote, Ronald F. (1987) The Early His­tory of Metallurgy in Europe. Long­man Archaeology Series: Londonand New York.

Thalin-Bergman, Lena. (1979) "Black­smithing in Prehistoric Sweden".Iron and Man in Prehistoric Sweden.Stockholm: LTs Forlag.

Zori, Davide. (2004) The InterpretiveValue of Nails, Rivets and ClenchBolts for Viking Age Economy andRitual. Unpublished MA Thesis, Uni·versity of California, Los Angeles.