Document Type : Original Article
Graphical Abstract
Highlights
This paper examines a small but coherent group of bronze bracelets (open rings) discovered at the entrance of the temple area of the fortress of Anzaf Superiore (Yukarı Anzaf) and subsequently inscribed with short Urartian dedications in the name of Išpuini, Minua, and Inušpua to the god Ḫaldi (Dinçol & Dinçol, 1995; CTU B 4-1A–D; B 4-2; B 4-3A–C). The objects, originally non-Urartian, feature a primary herringbone ornament at the terminals; the inscription is secondary, added to already finished artifacts. This practice lies at the crossroads of Urartian royal cult history, reuse of booty, and the circulation of metal ornaments in the southern Caucasus and the Caspian region.
The aim of this paper is twofold. On the one hand, to clarify what these objects are and how they functioned within the Urartian context, defining them not as epigraphic rings produced ad hoc, but as pre-existing bracelets re-functionalized through incision and votive deposition in the temple space. On the other hand, to assess what they reveal about political and territorial dynamics, the dedicatory formula mentions Amuša, a toponym which, in convergence with the late Urartian stele of Argišti, son of Rusa, from Tanahat in the Syunik region, points toward a location in southern Armenia (Dan, 2020: 128). Comparisons with materials from Gilan (Marlik) and Gegharkunik (Nerkin Getashen) show that the form of the open bracelet with geometric terminal ornaments fits into a Caucasian-Caspian tradition between the Final Bronze Age and the Early Iron Age, which provided the material reservoir upon which Urartian epigraphy intervened.
Methodologically, we adopt the term bracelets (open rings) for this material class. The analysis combines: (i) metric recording and tabulation of measurements (diameter, band width, thickness); (ii) technical observations on the incised grooves (depth, tool re-engagements, compressions at the junction) to demonstrate the secondary nature of the writing; (iii) palaeographic reading and comparison among specimens; (iv) find context and trans-regional parallels.
We anticipate three main outcomes: (1) demonstration of the operational chain raiding → inscription → temple deposition; (2) placement of the Anzaf bracelets within a Caucasian-Caspian network of forms and motifs; (3) an argument supporting the probable location of Amuša in southern Armenia, consistent with the Urartian expansion routes toward the upper Araxes River.
The fortress of Upper Anzaf (Yukarı Anzaf) stands on a prominent relief in the Anzaf valley, approximately 12 km northeast of Van, in a strategic position along the communication routes crossing the northern sector of the lake (Fig. 1) (see Belli, 1992a; 1992b; 1993a; 1993b; 1994; 1995; 1996; 1998a; 1998b; 1999a: 16-28; 2000; 2001; 2002; 2003; 2004; 2005; 2006; 2007; 2008; Belli & Ceylan, 2004; Dan, forthcoming). The elevation on which it is built reaches approximately 1,995 m above sea level, and from it, extensive portions of the surrounding territory are controlled, including the natural passages to the east and northeast (Belli, 1999a). The layout, irregular in shape but carefully adapted to the terrain morphology, occupies a total area of about 60,000 m², articulated into an upper citadel and a fortified lower city. The fortifications are built with double-facing walls of local stone blocks, filled internally with rubble and earth, according to a technique well attested at other Urartian sites in the Van region.
At the summit of the relief stood the susi, the tower-temple dedicated to the god Ḫaldi, known from two foundation inscriptions of Minua (CTU A 5-42; A 5-43), which record the construction of the sanctuary and the fortress walls. The sacred building, constructed of squared tufa and basalt stone blocks, exhibits a single-cell plan with a vestibule and lateral service rooms, set within a large artificial terrace surrounded by courtyards and secondary structures intended for ceremonial functions.
To the north of the tower-temple extends the palatial sector, characterized by a sequence of storage rooms, columned halls, and long distribution corridors. In this area, numerous epigraphic materials on stone, bronze, and clay have been discovered, documenting the intensive administrative and cultual activity carried out in the complex. Among these, notable finds include inscribed column bases (CTU A 5-100), pithoi with capacity marks and abbreviated inscriptions (CTU B CP An-1–11), a cuneiform tablet (CTU B CT An-1), and a clay bulla (CTU B CB An-1), as well as numerous metal objects attributed to various rulers up to certainly Sarduri II and perhaps even to later rulers (Belli & Salvini, 2006; Belli et al., 2004; 2009). These finds, associated with a rich architectural complex, allow us to outline a continuity of use of the fortress until the full reign of Sarduri II, and perhaps beyond down to Rusa, son of Argišti, with no evidence of post-Urartian renovations.
Figure 1. Map of the Upper Anzaf area. (Belli, 1999a: fig. 7).
In the western sector of the temple area, the inscribed votive rings CTU B 4-1A–D, B 4-2, and B 4-3a–b were discovered, constituting one of the most interesting epigraphic groups from the site (Dinçol & Dinçol, 1995: 30-36). The specimens, made of bronze and with a diameter ranging between 11 and 12.5 cm, bear nearly identical inscriptions written in the name of Išpuini, Minua, and Inušpua and dedicated to the god Ḫaldi, in which the conquest of the land of Amuša is mentioned. Rings B 4-1A–D were found near the entrance to the temple area, west of the tower-temple; B 4-2 comes from the same context; while the two fragments B 4-3a–b, with partial traces of cuneiform signs, were found in the same spot, delineating a coherent votive deposit (Belli, 1992b: 19-21, figs. 9-19; Belli, 1993a: 14-17, figs. 8-11).
The entirety of these epigraphic finds, distributed between the temple area and the palatial sector, shows a clear functional bipolarity between cult and administration. Furthermore, the presence of inscribed metal objects, sometimes of non-Urartian origin, perhaps obtained through raiding or as war booty, subsequently reused as supports for dedications, highlights practices of symbolic appropriation and integration of external materials into the Urartian cultural system (Belli et al., 2009: 102-109). This phenomenon helps to define Upper Anzaf not only as a fortified center of strategic importance but also as a place of ritual and ideological experimentation in the formation of the royal identity of Bia.
The inscribed bracelets from Anzaf, discovered at the western entrance of the temple area (Belli, 1992b: 19-21, figs. 9-19; Belli, 1993a: 14-17, figs. 8-11), form a coherent group in terms of typology and text but vary in preservation (Fig. 2 & Fig. 3). The discovery of votive bracelets with inscriptions in the temple built during the reign of Minua does not constitute evidence for the existence of an earlier sanctuary in the same location, as has been argued in the past (Dinçol & Dinçol, 1995: 36). It is instead more plausible that such cult objects, similarly to the rings of Išpuini (CTU B 2–7), were transferred to Anzaf Superiore at a later time, on the initiative of Minua himself. Their transfer may be connected to a process of cultic reorganization or reconsecration of the new tower-temple (susi), where they remained until the destruction of the building by fire.
All are open bracelets with overlapping terminals, a sub-circular cross-section, and an original herringbone (chevron) decoration near the overlap; on these already finished supports, the votive formula in the name of Išpuini, Minua, and Inušpua to Ḫaldi (with mention of the toponym Amuša) was subsequently incised. The herringbone pattern is conceived as an ornamental motif of the bracelet; occasionally, the beginning of the text is placed next to this motif purely for spatial convenience. The secondary writing shows, on all specimens, typical traits of reuse: burin depth not perfectly uniform, minimal re-engraving overlaps, micro-chipping along the grooves where the patina has been incised.
Figure 2. Drawing of the five bracelets discussed in the text (Salvini, 2012: 27).
Figure 3. Bracelet designs with copy of inscriptions (Dinçol & Dinçol, 1995: plates 6-10, 15-16).
CTU B 4-1A:
Morphology: nearly intact ring; short and well-finished overlap; homogeneous outer band.
Decoration: complete hatched band at the beginning of the text (6–8 notches); original ornament, respected by the text's incipit.
Epigraphy: clear and continuous ductus; no lacunae; minimal compression near the junction.
Measurements: external diameter 11.3 cm; band width 1.05 cm; section thickness 0.40 cm; overlap length 1.8 cm; average sign height 0.30–0.35 cm; usable epigraphic length ≈ 33.0–33.5 cm.
State of preservation: very good; compact patina; slight marginal abrasions.
CTU B 4-1B:
· Morphology: intact; slightly more extended overlap; slight thickening at the junction.
· Decoration: marginal hatching shorter than in 4-1A; functions as a "marker" of the incipit.
· Epigraphy: signs incised slightly more deeply; minimal alignment oscillations where the edge is worn.
· Measurements: external diameter 11.2 cm; band width 1.00 cm; thickness 0.38 cm; overlap 2.1 cm; signs 0.30–0.35 cm; epigraphic length ≈ 32.5–33.0 cm.
· State of preservation: good; small abrasions at the closure point.
CTU B 4-1C:
· Morphology: partial edge abrasion; thinned terminal; regular band on the preserved portion.
· Decoration: chevron/hatching barely visible; nevertheless, it retains its text-delimiting function.
· Epigraphy: local interruptions due to metal loss; sequence reconstructable by comparison.
· Measurements: external diameter 11.6 cm; band width 1.00 cm; thickness 0.38 cm; overlap ≈ 1.7 cm (partial); signs 0.28–0.33 cm; preserved useful length ≈ 70%.
· State of preservation: medium; widespread abrasions.
CTU B 4-1D:
· Morphology: lacunary ring; missing portions at the edge and in the overlap zone; slightly thinner band.
· Decoration: remnants of hatching; remnants of the original ornament; strokes.
· Epigraphy: estimated lacuna of ≈ 4 signs; wider spacing to compensate for edge irregularity.
· Measurements: external diameter 11.5 cm.
· State of preservation: mediocre, but sufficient legibility of the intact portion.
CTU B 4-2:
· Morphology: specimen with the largest diameter; particularly regular band; accurate junction.
· Decoration: faint but continuous marginal band; clearly frames the incipit.
· Epigraphy: relaxed and well-modulated ductus; clear and "wider" signs; optimal spatial management.
· Measurements: external diameter 12.5 cm.
· State of preservation: very good; homogeneous patina; micro-abrasions not affecting legibility.
CTU B 4-3a–c (fragments):
· Morphology: three fragments of a curved sheet of an additional ring; curvature and thickness consistent with the group.
· Decoration: remnants of hatching/chevron on the preserved margin, still legible as a "frame" for the epigraphic field.
· Epigraphy: few cuneiform traces, but compatible with the same votive formula (palaeographic comparison).
· Measurements: local radius derivable from fragments ≈ 5.6–6.2 cm, compatible with an original diameter of ≈ 11.2–12.4 cm; precise diameter not determinable.
· State of preservation: fragmentary; worn surfaces but legible in raking light.
In all specimens, the herringbone decoration, original to the bracelet, is respected by the engraver and remains autonomous from the text; the secondary inscription adapts to the curvature of the support, with slight compression of the signs near the junction and extension on the opposite side. The result is a technical-functional stratigraphy (primary ornament + added writing) that transforms a personal ornament into an ex-voto.
Following these observations, the relationship between the bracelets and the inscription becomes evident, with the latter clearly being secondary. First, this type of ring is not characteristic of Urartian culture. The cuneiform signs cut through the patina and show not entirely uniform depth, small burin re-engagements, and slight compressions near the overlap: typical indicators of incision on an already finished object. The herringbone ornament, by contrast, is primary: it is neither crossed nor "used" as a planned frame, but simply respected, and the incipit of the text is placed beside it for spatial convenience.
The strong homogeneity of ductus and formula among the specimens suggests a concentrated intervention in time, perhaps by the same hand or by the same workshop associated with the sanctuary, with the purpose of re-semanticizing pre-existing bracelets, transforming them into royal ex-votos. In other words, this is not an integrated decorative project, but an ideological overwriting: the official Urartian inscription "appropriates" the object and transfers it from the domain of personal ornament to that of cultural memory and political legitimation.
The typological classification of these artifacts requires an integrated approach that goes beyond formal impression ("rings") and considers metric, morphological, functional, and contextual parameters. On metric-morphological grounds, the open-circle profile with overlapping terminals and original herringbone ornament at the ends refers to the family of armlets/bracelets: the external diameter of 11–12.5 cm and the band width of ~1 cm imply an internal diameter compatible with use on the wrist/forearm/proximal part of the forearm, especially in winter or clothing contexts (worn over sleeves or wraps), rather than "digital" use. To this are added technological and use indicators (regularity of the cross-section, edge finishing, possible micro-deformations consistent with wearing), which, when present, reinforce the "ornamental" reading.
The category "ingot" (as in some readings of Marlik; see Negahban, 1996: 312; cf. Piller, 2007: 73) is not defined by the form, which remains that of the open bracelet, but rather by the economic-functional role attributed in specific contexts (funerary accumulations, metal reserves): it is therefore a diagnosis of use and not a morphological typology. For Anzaf, two elements are decisive: (1) the herringbone decoration is primary (part of the bracelet's design), while the inscription is secondary (incised ex post on a finished object), and (2) the temple deposition context with textual seriality indicates a votive re-functionalization of pre-existing bracelets, not the production of "epigraphic" rings in the strict sense. Consequently, the terminological choice follows: in the body of the text, we will use bracelets (open rings) for the material class.
Two bracelets appear intertwined. Since these are open armlets, this configuration can be obtained at any stage (collection, deposition, documentation). In the absence of explicit stratigraphic indications, the intertwining does not alter the functional reading of the group.
The inscribed votive rings from Yukarı Anzaf constitute a homogeneous group of bronze artifacts that fall within the typology of devotional objects dedicated during the reign of Išpuini and Minua. The specimens, with diameters ranging between 11 and 12.5 cm, bear nearly identical inscriptions, written in the name of Išpuini, Minua, and Inušpua, and dedicated to the god Ḫaldi (Fig. 4). The text, of a votive nature, records the offering to the god on the occasion of the conquest of the land of Amuša, an episode attributable to the phase of military expansion of the Urartian kingdom toward the east, also attested by the inscription of Ojasar-Ilandagh (CTU A 3-8) in the Nakhichevan region.
The graphic organization of the inscriptions follows a circular course along the external surface of the rings, adapting precisely to the profile of the support. Despite some irregularities and lacunae due to the condition of the metal or wear, the inscriptions reveal good technical skill in the execution of the cuneiform signs and a coherent planning of the writing space. The comparison among the different specimens (CTU B 4-1A–D; B 4-2) highlights slight graphic and phonetic variants, which do not compromise the substantial uniformity of the text but suggest the possibility of a serial execution by the same workshop or by a group of artisans active in close connection with the court or with a temple center.
The short text, repeated with minimal variations on all the rings, reads the following words:
> "To Ḫaldi, their Lord, Išpuini son of Sarduri, Minua son of Išpuini, and Inušpua, son of Minua, dedicated when they conquered the land of the city of Amuša."
Figure 4. Bracelets from Marlik (Negahban, 1996: pl. 139).
The incised formula, written in the name of Išpuini, Minua, and Inušpua, and addressed to the god Ḫaldi, fits within the official votive language of the kingdom. Still, the bronze support reveals an origin foreign to local workshops. The act of incision on artifacts of external origin can be interpreted as a symbolic gesture of appropriation and ideological re-elaboration of booty, through which the object is integrated into the Urartian cultural circuit. In this way, the rings assume not only religious value but also political significance, as they reflect a form of material and semantic transformation of the "conquered good" into a testament to victory and dynastic legitimation under the protection of the god Ḫaldi.
The toponym Amuša, attested on the rings of Anzaf Superiore (CTU B 4-1A–D), where it appears in the formula "when they conquered the land of the city of Amuša," occurs elsewhere exclusively on a stele of Argišti, son of Rusa (CTU A 11-3), found reused as a khachkar near the medieval monastery of Tanahat in the Syunik region. This circumstance points to the location of the center in the mountainous area of southern Armenia (Dan, 2020: 128). The text on the main face of the stele is deeply damaged by the medieval khachkar. Nevertheless, the name of Argišti, son of Rusa, is legible, confirming the correct attribution of the inscription. The reverse of the stele is much better preserved. The annalistic-type text reports the following words:
> "He came before me [ . . . , the king(?)] of Etiuḫi, who (or: says [scil. Argišti]) [ . . . ] the land, which villages/cities listened . . . which in battle I reached at(?) [ . . . ] I made/built; males and females I deported, villages I burned, for [tresses I destroyed]. išiaṣie šuga from everything, there was a battle marine to/on the land of ṣuluqu . . . before the city of Irdua, he came before me. The enemy I subjected powerfully, tribute I established for him(?). By the greatness of Ḫaldi, I am Argišti, the son of Rusa, mighty king, king of Biainili, lord of the city of Tušpa. By the command of Ḫaldi, Argišti says: isini of the Sun-god garrison bi... I undertook this military campaign toward the city of Amuša, men [I deported] and women to the land of Biainili. Argišti [says]: whoever [eliminates] my name and [places] his own name, may Ḫaldi, the Storm-god, the Sun-god destroy him, his name and his seed from under (the light of) the sun" (CTU A 11-3 Vo.)
The historical-geographical framework is coherent: from the very beginnings of Urartian expansion, the region between Nakhchivan, Gegharkunik, and Syunik fell within the kingdom's radius of action, as evidenced by the inscription of Ojasar–Iləndagh (CTU A 3-8), which documents an early penetration along the Araxes corridor; in this same sector, the Tanahat stele provides epigraphic confirmation of the name Amuša, reducing the need to resort to more eastern or more northern hypotheses. In these terms, it is useful to recall the proposed location in Artsakh/Nagorno-Karabakh (Salvini, 2009: 209–210; 2018: 430), which confirms the proposal of a location in southern Armenia, likely in a mountainous context controlling local transit routes (Dan, 2020: 128).
In Gilan (Marlik), excavations have yielded a series of open metal rings, classified in the publication as "bronze ingots" but morphologically comparable to bracelets/rings (Negahban, 1996: 312; Piller, 2007: 70): thickened body in the middle of the ring, tapered ends, and incised terminal ornament (chevron/zig-zag/triangles). The contexts are well defined: two pieces (Negahban, 1996: 312, no. 974) came from Tomb 25; one specimen (Negahban, 1996: 312, no. 975) from Tomb 36; while the largest group (Negahban, 1996: 312, no. 976, eight specimens: 417M, 418M, 419M, 420M, 421M, 422M, 423M, 730M, 732M) comes mostly from Tomb 36 and, for two numbers, from Tomb 40. Dimensions range between Ø ~8.3 cm (no. 974) and Ø ~16.5 cm (no. 419M), with a thicker central band and incised motifs at both ends; specimens nos. 418M, 419M, and 421M clearly show chevrons/triangles on both ends. Tomb 36, small and extremely rich (containing gaming objects), is interpreted by the excavators as a high-status burial, perhaps of a young prince; here, the concentration of rings/ingots suggests an intentional accumulation of metal value, although the form remains fully wearable and comparable to Caucasian specimens. In other words, Marlik documents both the formal tradition (open circle + terminal ornaments) and a polyfunctionality of use (ornament/metal reserve) compatible with the votive re-semanticization observed at Anzaf.
In Gegharkunik, at the necropolis of Nerkin Getashen, a small mound with a cromlech (diam. 4.8 m) covered an E–W oriented stone cist (ca. 1.9 × 0.7 × 1.5 m) containing four individuals. In the eastern sector, a woman aged 18–25 years and a child aged 1–1.5 years were deposited face to face: the child wore a bronze bracelet; among the grave goods were also a bowl, a small flask, a porcelain pendant, earrings, and carnelian beads. In the western sector, a man aged 40–45 years and a woman aged 65–70 years were found with bracelets, earrings, a pendant, and a tube; all evidence agrees on the personal use of the bracelet during life and its funerary deposition as part of the grave goods, attesting to the diffusion of this type within an intergenerational family context (typological drawing with herringbone ornament on the terminal portions) (Xnkikyan, 2002: 74-75, pl. LXVIII). This context, unequivocally "of use," completes the evidence from Marlik, where the same formal profile can take on, depending on the case, either ornamental or "monetary" value.
In more general terms, the circulation of bracelets/rings of this type falls within a broader framework of relations between southern Armenia and Gilan, active between the Final Bronze Age and the Early Iron Age, in which technological affinities (alloys, casting and finishing techniques), decorative choices (chevron, stippling), and funerary deposition practices converge. Within this network, among the most striking and recurring indicators are pendants, which serve as sensitive markers of the sharing of formal models and symbolic repertoires and corroborate, by analogy, the mobility and diffusion of the bracelets/rings under discussion (Piller, 2012; Dan et al., 2017).
This paper aimed to analyze a group of rings/bracelets of non-Urartian manufacture discovered at the fortress of Anzaf Superiore. These are open bracelets, with an original herringbone ornament at the terminals, subsequently reused through the incision of short inscriptions. The secondary nature of the writing, recognizable in the not perfectly uniform depth of the grooves, in the small burin re-engagements, and in the slight compressions near the junction, indicates a votive re-functionalization: what was originally personal ornament (or store of value) is transformed into a public ex-voto. The deposition near the entrance to the temple of Ḫaldi configures these bracelets as a truly ritual and political device: booty acquired in war, appropriated through royal writing, and consecrated in the cultual space.
The seriality of the formula, the palaeographic coherence, and the recurrent management of the writing space suggest a concentrated and probably unitary intervention (the same hand or the same workshop), capable of rendering homogeneous a group otherwise heterogeneous in terms of state of preservation and constructive details.
On the historical-geographical level, the convergence of data indicates that Amuša should be located in southern Armenia: the co-occurrence of the toponym on the rings of Yukarı Anzaf (CTU B 4-1A–D) and on the late Urartian stele of Argišti, son of Rusa, from Tanahat, together with the evidence of Urartian penetration along the upper Araxes, outlines a coherent and plausible framework (Dan, 2020: 128). Within this framework, the non-Urartian nature of the bracelets and their epigraphic reuse clarify the function of the group: objects that entered the royal circuit as booty, then re-functionalized through the votive formula in the name of Išpuini, Minua, and Inušpua to Ḫaldi, and finally deposited in the temple area as material testimony of devotion and the legitimization of power.
The deposition near the entrance to the temple of Ḫaldi shows how this re-functionalization takes place in the cultual space, transforming a movable good, worn or accumulated as value, into an institutional ex-voto. The seriality of the operation (repetition of the same formula on multiple specimens, palaeographic coherence, recurrent management of the writing space) suggests a concentrated action, perhaps entrusted to the same hand or workshop associated with the sanctuary: a material-textual device in which booty becomes public memory of the conquest and a vehicle for dynastic legitimization.
The eastern comparisons reinforce this reading. In Gegharkunik (necropolis of Nerkin Getashen), the bracelets appear in personal use and in funerary deposition; in Gilan (Marlik), open metal rings of the same morphological family can occur in tomb accumulations, also with a "monetary" dimension. The result is a shared Caucasian-Caspian formal horizon—open bracelets with geometric terminal ornaments—within which the Anzaf specimens fit with a Urartian specificity: epigraphic re-semanticization and temple consecration. Within this broader network of exchanges, the recurrence of typologically similar pendants between southern Armenia and Gilan serves as a visible indicator of the network of relations that forms the backdrop to the circulation of these same bracelets/rings.
In summary, the probable location of Amuša in southern Armenia and the operational chain of raiding → secondary inscription → votive deposition integrally explain the nature of the Anzaf rings: non-Urartian objects that, once inscribed, were incorporated into the political liturgy of the kingdom. Metallographic and isotopic analyses, together with detailed 3D surveys of the grooves, may refine the provenance profile of the metals and the relative sequence between patina and incision, verifying the degree of integration between Caucasian-Caspian craft circuits and the Urartian cultural sphere that reused them.
The authors thank the excavation team of the Anzaf fortresses, the late Prof. Oktay Belli, and ISMEO for their support.
The authors declare no competing interests.
This contribution was jointly conceived by the authors. R. Dan wrote sections 2, 3, and 5; R. Dan, A. Bonfanti, and M. Keshavarz wrote section 4; A. Bonfanti wrote section 6. The Introduction and Conclusions were jointly written. All authors approved the final manuscript.
Available from the corresponding author by email upon request.
