In the midst of great kings: the monumentalization of text in the Iron Age Levant

: TheBar-RakibPalaceInscriptionsfromZincirlihavereceivedrelativelylit-tle attention from philologists and archaeologists alike because of their predictable and derivative content. However, these monuments provide an unparalleled insight into the monumentalization of text in the Iron Age Levant. As might be expected, Bar-Rakib’s Aramaic inscriptions and reliefs repeat themes and tropes from other monuments. They also were strategically deployed at the site so as to interact with nearby monuments left by earlier rulers. What has received less attention is the fact that Bar-Rakib’s monuments also shared many artistic tropes with small finds from Zincirli, including letters, incantation plaques, seals, and amulets. These correspon-dences suggest that monumental texts functioned by appropriating aspects of personal artifacts to be used on a communal scale. By projecting not only prestige but also intimacy, Bar-Rakib’s inscriptions invited their audience to interact with them in imaginative ways. As the audience related to the monumental texts through acts of reading, viewing, and ritual, they would in turn reconfigure their own relationships to other communicative media, places, each other, and the polity as a whole. It was this ability to relate to communities and thus reshape them that made a text monumental in the Iron Age Levant. This was accomplished through the strategic juxtaposition of text with visual and performative media in particular spatial contexts.


Introduction
The remains of the ancient Syro-Anatolian kingdom of Yādiya/Samʾal provide an unparalleled insight into the monumentalization of text in the Iron Age Levant. This Iron Age kingdom-based at the modern site of Zincirli Höyük (the ancient city of Samʾal)was active at least from 920 to 711 BC, and it preserves monumental inscriptions in four languages. These include Hieroglyphic Luwian, a member of the Anatolian branch of the Indo-European languages, as well as three Northwest Semitic languages: Phoenician, Samʾalian, and Aramaic. 1 In particular, it preserves more Northwest Semitic monumental inscriptions than any other single site. While there is currently no evidence for the drafting of specific inscriptions, there are substantial indications that the Zincirlian monuments were produced by adapting iconographic and textual themes from a variety of different media attested at the site, such as letters, seals, and jewelry. In this article, I will argue that the monumentalization of text at Zincirli involved bringing it into particular relationships with materials, images, and places, and these material relationships promoted new social relations between the texts and the communities they targeted. The targeted communities would experience these monumentalized texts not only through reading and hearing, but also through viewing the text with its accompanying iconography and through moving about the space created by the text's deployment. The text was thus monumentalized by being transformed into a piece of mixed media-an object of reading, viewing, and performing-with communal significance. This is perhaps nowhere more apparent than in the Aramaic palace inscriptions of Bar-Rakib.
Bar-Rakib was the last attested king of Samʾal, and therefore the last to erect monumental inscriptions at the site. Of the monumental inscriptions uncovered so far at Zincirli and in its environs, nearly half were erected by Bar-Rakib. His monuments represent the pinnacle of Zincirlian monumental discourse and the largest expression thereof. They also grew out of a uniquely tendentious period in Zincirli's history. During the second half of the tenth century BC, the polity that would become Samʾal seized its independence from the burgeoning empire of Karkamiš (Giusfredi 2010: 45-51, 78-9;Younger 2016: 391;Herrmann 2017: 5-6). Afterwards, apart from occasionally paying tribute to Assyria, the polity was more or less independent until the second half of the eighth century. Then, with the appearance of Tiglath-Pileser III and his restructured Neo-Assyrian empire on the international scene, Samʾal accepted vassal status under Bar-Rakib's father Panamuwa II. The latter part of Bar-Rakib's reign saw significant impositions from the Assyrians at Samʾal, including changes in monumental discourse and even a change in standard language. Simultaneously, non-royal elites in the kingdom had been steadily increasing in power, posing a greater threat to the newly weakened royal family than ever before (Struble and Herrmann 2009: 40-2;Gilibert 2011: 126-8;2013: 54). Bar-Rakib's father, Panamuwa, had accepted Assyrian vassalage in the first place in order to retake the throne from a usurper (Gilibert 2011: 16-17;Younger 2016: 416-17). Bar-Rakib's monuments spoke to this unique need to appease both native elites and foreign overlords, as the various dimensions of his monumental discourse attest.

Defining monuments and monumentality in terms of social formation and affordance
Monuments are typically understood as large, public, and permanent works of art and architecture. These defining features are increasingly falling by the wayside, however, because an artifact may possess all of these qualities and not be a monument. Alternatively, an object may possess none of these qualities and still be monumental. The Guennol Lioness from Ancient Elam, for instance, stands only 8.4cm high, and yet it has been analyzed as a monument because of its effect on its viewers (Porada 1950;Osborne 2014: 1-2). In Wu Hung's seminal study Monumentality in early Chinese art and architecture, his paradigmatic examples are the Nine Tripods. Not only were these bronze vessels not particularly large and hidden from public view, it may be that they never even existed outside of literary depictions. What made them into monuments was not their presence in public space but rather within the public imagination (Wu Hong 1995: 4-12).
Monuments are thus not defined by their size but rather by their scale (Smoak and Mandell 2019: 311), and theoretical work on monuments now generally defines them as community-scale communicative media (Gren 1994: 89-91;DeMarrais et al. 1996: 17). Monuments primarily communicate messages related to social formation. Following Nathaniel Levtow, I use social formation as a short hand for the 'construction and configuration of social relations,' which he emphasizes as a 'dynamic, constructive, relational process' (Levtow 2008: 11, 33). Similarly, Wu Hung relates monumentality to 'political, ethical, or religious obligations to a tradition' that serve to 'consolidate a community or a public' and 'to define a center for political gatherings or ritual communication' (Wu Hong 1995: 4). James Osborne thus suggests that monumentality is best understood as 'an ongoing, constantly renegotiated relationship between thing and person, between the monument(s) and the person(s) experiencing the monument' (Osborne 2014: 3, emphasis in original). Following Osborne's relational definition, Timothy Pauketat suggests that monuments primarily function by affording relationships to other objects, people, and places. His use of 'affordance' is meant to emphasize that a monument allows its users to construct meaning dynamically, rather than simply receive a pre-existing, singular meaning contained in the monument (Pauketat 2014: 442). Combining these observations, I define monuments as artifacts that afford social formation to communities. Monumentality is thus an artifact's potential to accomplish this. Monuments only function when communities engage with them and use them to relate to one another (Wu Hong 1995: 4-11;Osborne 2014;Pauketat 2014). By engaging monuments, communities develop relationships amongst themselves, to places and artifacts, and often to the elites deploying the monuments. Monuments are the material means of enacting social relations on a communal scale. This definition is admittedly vague unless it is historically situated, which is precisely why Wu recommends an approach to monuments that considers them in concert with other examples of monuments from the same culture or period or in comparison to near cognates (Wu Hong 1995: 11-14). Of course, in producing such historically and culturally specific studies of monuments, one hopes that universals will begin to emerge and take the place of failing descriptors like size, material, durability, and publicity.

Multiple dimensions of affordance
Especially in a cultural context of widespread non-literacy like the Ancient Near East, it must be emphasized that monumental texts afforded meaning in a variety of ways. Monuments are mixed media; they are simultaneously texts and images that are inextricable from the locations and performances in which they are embedded (Thomas 2014: 60-1;Da Riva 2015: 610). Kristel Zilmer thus treats monumental inscriptions as 'composite products and processes that activate multiple modes of expression in their varying settings' such that 'the capability of interpreting the materiality of an object becomes important too, if not more important than the understanding of the words per se' (Zilmer, forthcoming). In other words, an audience's experience of a monumental inscription involves acts of viewing, moving, and perhaps even touching and feeling. Ideally, reading and/or hearing would also form a part of this experience, but this mode of experience may have been inaccessible to most members of the audience unless an expert was present to read the text for them. Therefore, a monumental text must be studied in relation to the monument of which it is a part. In addition to interpreting a monumental text's words and rhetorical structure, scholars have suggested accounting also for their orthography, artistic technique, epigraphic supports, accompanying iconography and ritual implements, architectural or other spatial contexts, audiences, and attached ritual performances (Thomas 2014: 60-1;Bahrani 2014: 24-43;Smoak and Mandell 2019: 311-12).
I propose that different aspects of meaning affordance can best be considered under three broad categories or dimensions: the verbal dimension, the aesthetic dimension, and the spatial dimension. The verbal dimension accounts for a text's language, semantics, rhetorical structure, poetics, and any aspect experienced when a text is read as a piece of written media. The aesthetic dimension accounts for a text's material, orthography, iconography, epigraphic support, the other objects accompanying the monument, and any other aspect of a monumental text experienced as a piece of visual media. The aesthetic dimension is thus concerned not only with the visual aspect of the text but also of the text's support. Monumental texts were separated from everyday writing by their media and adornment but also by their strategic deployment in particular environments. The spatial dimension accounts for a monumental text's location on a local and geographic scale, its space-and place-making capabilities, its architectonic context, and the local ritual motion and actions it promotes. It is important to emphasize that these three dimensions were not experienced separately by monument users and are rather proposed for the convenience of analysis. As mixed media, inscribed monuments were experienced as written, visual, and spatial media simultaneously. There are cases where these dimensions allowed the monument to afford different meanings along different axes, 2 but often they worked together in combination to produce a unified affordance that could be experienced in multiple ways (Gilibert 2011: 109;Levtow 2014: 35-36;Harmanşah 2015: 152).

Levantine 'I Am' monuments and their monumentality
Bar-Rakib's palace inscriptions fall into a category of Levantine monumental inscriptions labeled 'I Am' monuments. These monuments have been assigned to various textual genres, but they are united by their use of the 'I Am' formula, a unique development in Levantine monumental discourse. The 'I Am' formula was used almost exclusively within the Levant and is attested perhaps just under 100 times in inscriptions dating from between 1500 BC and 200 BC. In addition to its limited geographical and historical deployment, the 'I Am' formula is also significant enough a marker to justify labeling a type of inscription with it because of its function. The 'I Am' formula that headed these inscriptions actually conjured up the imagined speaker in the minds of the monuments' users by means of deictic projection. That is, the use of the deictic element 'I' without a speaker present outside of the monument, triggered the users to imagine an encounter with the monument's implied speaker (Hogue 2019b). The formula thus reified the primary function of these monuments, which was to conjure up their implied speaker before their target audience. This allowed the conjured agent to interact with his monument's users in their imaginations, to make demands of them, and ultimately to reshape their identities by proposing a particular narrative of cultural memory as well as particular social roles and rules for them to fulfill (Sanders 2009: 118;Hogue 2019b: 339;2019c: 81). This was the primary meaning afforded by Levantine 'I Am' monuments: the conjuration of the implied speaker and the social formation he proceeded to propose. Furthermore, the formula did not accomplish this merely as a textual element but also in special relationship to visual and spatial elements of the monument.
The function of 'I Am' monuments can be briefly illustrated by the statue of Idrimi, a Syrian king who set up a new dynasty at Alalaḫ in the Late Bronze Age. Dating to 15 th century BC city of Alalaḫ, the statue of Idrimi is the oldest known example of a Levantine 'I Am' monument. The statue conjures up the presence of Idrimi in two obvious ways. First, the opening 'I Am' statement deictically projected the speaker into the minds of the text's readers and/or hearers (Hogue 2019b: 327). Second, the statue itself was likely perceived as a substitute or coextensive constituent of the person, rather than as a mere representation (Aro 2013: 236;Bahrani 2003: 121-48;2014: 43). While different mechanisms of conjuration can be described here, the statue and its inscription were a single monument. Text and image functioned as two dimensions of a single functional artifact. Idrimi is a particularly good example of this because the 'I Am' statement was carved along the statue's mouth, suggesting that it was meant to be understood as the direct speech of the statue. The placement of the text thus united its function with that of the image. While the original placement of the statue is impossible to determine, the condition of the statue and its findspot suggest that it was 'given an honorable burial' by the denizens of Alalaḫ after it was destroyed by an invading army (Longman 1991: 60-1). This ritual response demonstrates that those burying the monument treated it as though it were a person because it was a means of conjuring up Idrimi. A little over 100km north of Alalaḫ and over 700 years later, Bar-Rakib's 'I Am' monuments present a sophisticated development of this tradition stretching back to Idrimi.

The production of the Bar-Rakib Palace Inscriptions
I turn now to my primary case study on Levantine monumentalization: the Palace Inscriptions of Bar-Rakib. These three inscriptions and possibly more were carved on basalt wall reliefs in Bar-Rakib's newly built palace on the acropolis of Zincirli near the end of the eighth century. These reliefs were produced as adornments for stone orthostats, an essential architectural feature of Syro-Anatolian palaces and other important buildings more generally. Stone orthostats were first developed as a means of protecting mudbrick walls from erosion; the earliest iterations of them were even completely undressed and may have only served this practical purpose. Later, Syro-Anatolian rulers and artisans began to take advantage of the communicative opportunities orthostats afforded. From the beginning, the use of stone projected a sense of permanence which was only augmented by its use to protect other materials from the weathering effects of the environment. The use of orthostats in public and ceremonial spaces also rendered them an ideal medium for public-targeted messages. This architectural feature was thus ripe for monumentalization, and by the Iron Age they had become some of the most common carriers of relief imagery and inscriptions (Harmanşah 2007: 72-4). Nevertheless, these orthostats remained essential architectonic features of the buildings they framed; they were never employed as mere adornments. As such, Bar-Rakib's palace orthostats were produced as essential parts of Bar-Rakib's palace. It is highly unlikely that ancient viewers differentiated them as individual scenes or inscriptions serving as decoration rather than as essential pieces of the palace itself.
Basalt was plentiful in the region surrounding Zincirli, and multiple quarries may have supplied the stone utilized for Bar-Rakib's monuments. Gerçin was the nearest quarry and also the findspot of an inscribed Zincirlian statue. Though farther away, Yesemek is also a good candidate because it functioned as both a quarry and a sculpture workshop (Pucci 2008: 174 n. 879;Gilibert 2011: 73;Herrmann 2017: 261-3). After the stones were quarried, they were roughly finished in workshops using stone tools. Detailed work was probably carried out using iron chisels and was completed on-site where the orthostats were installed. 3 Monumental relief scenes have a longer history in Zincirli than do the inscriptions that accompany them in the case of the Bar-Rakib palace. When the site was first established as the kingdom's central city, its gateway was adorned with Karkamišean reliefs that were moved from the nearby site of Pancarli, which was formerly the regional seat of power under the Karkamišean kings (Herrmann 2017: 263-8). After this initial reuse of Karkamišean reliefs, the artistic style of Karkamiš continued to inform the reliefs of Zincirli. Reliefs from the two sites are remarkably similar. It has been determined that there were in fact separate workshops producing monumental sculpture for each site, but these were clearly conversant with one another (Gilibert 2011: 122).
While there is no definitive material evidence for earlier drafts of the text of the inscriptions, we do have some material indications of the techniques utilized to render them in stone. Northwest Semitic inscriptions were usually incised in stone by means of iron chisels. The best indication of this procedure is a broken stone tablet from Tell Aushariye inscribed in ink but with the initial letters cut into the stone by means of a chisel (Younger 2007: 142;Fales 2007: 106;Keimer 2015: 194 n. 5). 4 Apparently, stonemasons used some sort of guide in the production of inscriptions, chiseling out originally inked inscriptions. While Zincirli attests iron chisels that may have been used to produce its inscriptions, 5 the inscriptions are unique among the Northwest Semitic corpus in that they are carved in raised relief rather than incised. That is, the tools were used to cut away the surface of the stone and leave the rounded letter shapes. This method of carving inscriptions was clearly adapted from the standard practice of producing Hieroglyphic Luwian inscriptions, which had been inherited from the artisans of the Hittite empire (Gilibert 2011: 82).
Draft inscriptions may have been consulted at Zincirli and even traced onto the stone, but none have survived (Tropper 1993: 176). Spelling errors in the inscriptions indicate, however, that the stonemasons interpreting these drafts were not as literate as the scribes who produced them. If the inscriptions were actually traced on the stone first, this may imply that the individuals responsible for the tracing were not actually the scribes but other artisans working from a separate draft they could not expertly read (cf. Worthington 2012: 154-9). Regardless, the scribes were apparently not directly involved in the final production of the monuments.
If the composition of the Zincirlian inscriptions was carried out by a scribe and its carving in stone by a separate-likely illiterate-stonemason, it is possible that the mason was also responsible for the figurative art adorning the monuments. Images and texts interacted with one another on the monuments and were executed so as to communicate a single message together (Bunnens 2005: 22;Gilibert 2011: 79-82). It is thus probable 4 An edition of this inscription prepared by André Lemaire and Frederick Mario Fales is forthcoming. A photograph is available on the Tell Aushariye website at https://aushariye.hum.ku.dk/english/ironage/ (accessed 20 April 2022). 5 Nine iron chisels that may have been used for stone-carving have been discovered in a seventh century layer of Area 5 in Zincirli's North Lower Town. However, their scattered distribution suggests that this may not have been the phase during which they were utilized (Herrmann 2011: 382). Similar chisels were also discovered at Zincirli's South Gate and in the Southwest Lower Town. None of the artifacts have yet been published (Herrmann, personal correspondence). that these were carved at the same time by a single craftsman. Unfortunately, there are no Northwest Semitic monumental inscriptions that describe their own production. The same is not true for Hieroglyphic Luwian inscriptions, which occasionally give both scribes and 'carvers' credit for having carved inscriptions. 6 However, this leaves open the question of what role each played in the production of such inscriptions. Regardless, it is clear that multiple individuals were involved in producing monumental inscriptions in the Levant. While the possibility of a scribe who was also a stonemason remains open, it is also possible that the scribe's role only extended as far as composing the text.

The monumentality of the Bar-Rakib Palace Inscriptions
This section describes the verbal, aesthetic, and spatial dimensions of the three Bar-Rakib Palace Inscriptions separately before analyzing them in concert to understand better the relationship between these dimensions of artistic production and the monumentalization of text. In general, all three dimensions make use of earlier monumental and artistic discourse attested at Zincirli but with some significant new features. I will argue that this reuse of old features in new contexts and configurations was a significant feature of monumentalization in the Levant. Previous scholarship has broadly defined monumentalization as the transformation of some thing into a monument, and this process could include cases in which pre-existing non-monumental texts are received by later audiences as monuments (cf. Riegl 1903Riegl , 1982Harmanşah 2013: 30-1;Osborne 2017: 4-5;Christiansen 2019). However, intentionality was key to the monumentalization of texts in the Ancient Levant, so I instead emphasize the intentional production of monumental text as a monumentalization process. The monumentalization of text at ancient Samʾal involved transforming a text into a piece of mixed media with communal significance. The Bar-Rakib Palace Inscriptions took this process a step further by addressing multiple communities at once using a variety of media strategies.

The verbal dimension
I begin this section with the Aramaic text of the three Bar-Rakib Palace Inscriptions based on the transcriptions in Kanaanäische und aramäische Inschriften (KAI) (Donner and Röllig 2002). 7 For the broken portions of the transcriptions, I follow Josef Tropper's reconstructions but the translations are my own. I direct the reader to previous treatments of the inscriptions for in-depth commentary and literary analysis of individual clauses (Younger 1986;Tropper 1993: 132-46;Botha 1996 Bar-Rakib was the first Samʾalian king to erect monumental inscriptions on Samʾal's acropolis in over a century. It is thus unsurprising that these inscriptions are conversant with-and indeed aggrandizements of-the previously erected monumental inscription on the acropolis, which was still standing only a short distance away from Bar-Rakib's inscriptions. This was the inscription of Kulamuwa, one of the early kings of Samʾal who had successfully negotiated internal conflict, war with neighbouring Cilicia, and the encroachments of the Assyrian king Shalmaneser III to maintain Samʾal's independence (Brown 2008). He erected a monumental palace with a relief inscription describing his achievements on the acropolis to celebrate his reign. As Bar-Rakib's inscription implies, that palace had remained standing until the days of Bar-Rakib, and neither the palace nor its inscription had been replaced by any subsequent constructions by the other Samʾalian kings. Bar-Rakib was the first to respond to this inscription in monumental form and to attempt to surpass the achievements of his forefather.
Bar-Rakib adapted and modified Kulamuwa's phraseology in order to characterize his own reign. First and foremost, Bar-Rakib maintained the traditional 'I Am' opening of Levantine inscriptions, which was their most unique and arguably most operative feature. This formula conjured up the presence of the implied speaker in the minds of the inscription's audience by means of deictic projection. 12 Though Kulamuwa was a cendimension of an individual's social self that is given material form in the monument (Suriano 2014;Sanders 2012Sanders , 2013Hogue 2018). See also Benjamin Sommer's discussion of avatars in Ancient Near Eastern religion (Sommer 2009: 15, 40-1).
tury dead, he remained present at Samʾal in the form of his monument and its opening pronouncement of ʾnk klmw 'I am Kulamuwa' (KAI 24: 1) (Tropper 1993: 30). 13 By drawing upon this traditional feature of Samʾalian monumental discourse in his own inscriptions on the acropolis, Bar-Rakib permanently materialized himself alongside Kulamuwa as his counterpart and as his competitor.
Bar-Rakib's inscriptions also draw on the typical Zincirlian motif of the dynasty being 'in the midst of great kings'. This phraseology occurs multiple times in the Zincirlian corpus in three different languages. 14 In addition to offering a means of semantically paralleling earlier monuments, the trope of being 'in the midst of great kings' is also a significant traditional element of the poetics of Zincirlian inscriptions. This trope relies upon the strategic use of deixis to propose an ideological message. Elements of temporal, spatial, and even personal deixis in these texts are utilized to indicate a relationship to the king, who designates himself 'I' and thus the deictic center of this discourse (D. J. Green 2010: 22, 297-307;Hogue 2019c: 93-5). Kulamuwa, for instance, divided his inscription into two halves. The first half focuses on the Samʾalian dynasty 'in the midst of great kings'-that is, struggling for survival among their many enemies. Kulamuwa overcomes these enemies at the end of that portion of the inscription. The second half of the inscriptions relates Kulamuwa's domestic achievements and his various improvements to Samʾalian society closer to home (O'Connor 1977: 15-29;Fales 1979: 6-22). The inscription thus transitions from describing a negative past marked by the presence of foreign enemies (the 'great kings') to describing a positive present marked by peace and prosperity among the people of Samʾal. Kulamuwa's ideology is expressed through the negative characterization of temporal and spatial distance from himself and the positive characterization of his close proximity.
Bar-Rakib uses the trope of being 'in the midst of great kings' to develop his own ideological message, but this is surprisingly different from that of Kulamuwa. Bar-Rakib's use of the trope reveals a significant softening of that rhetoric and likely points to Assyrian pressure. The longest complete palace inscription of Bar-Rakib (KAI 216) is a bipartite inscription like that of Kulamuwa's neighbouring wall relief, but he has reversed its poetics. In the first part, Bar-Rakib pronounces his uncompromising loyalty to a foreign king-Tiglathpileser III of Assyria. When he speaks of being 'in the midst of great kings', these great kings are no longer his enemies but rather 'my brother kings', as they presumably accept the same pro-Assyria ideology as Bar-Rakib (D. J. Green 2010: 296-7).
Elements that are distant from Samʾal are thus presented in a surprisingly positive light, as opposed to Kulamuwa's use of distance to suggest a negative ideological stance. In the second part of the inscription, Bar-Rakib surprisingly takes aim at his predecessors-Kulamuwa in particular. He uses a distal deictic particle hʾ 'that' to indicate the neighbouring palace of Kulamuwa and announces that it 'was no suitable palace for my fathers'. This use of deixis is significant because it not only gestures towards a real element of the built environment of Bar-Rakib's inscription but does so in a negative way. Distal deixis was typically a metaphor for an enemy or an inept predecessor in Levantine inscriptions. By indicating Kulamuwa's palace-and undoubtedly its attendant inscription-in this way, Bar-Rakib casts aspersions on Zincirli's past as an independent polity (Davis 2019: 190-2). By contrast, he utilizes the proximal demonstrative pronoun znh 'this' to indicate his own palace and the positive prospective future of the polity as an Assyrian vassal.
Bar-Rakib's language choice also exposes his limitations as an Assyrian vassal. Kulamuwa chose the prestige dialect of his day-Phoenician-for his inscription, while Bar-Rakib abandoned Samʾal's eighth century prestige dialect-Samʾalian-in favor of the Standard Old Aramaic favoured by the Assyrians (Younger 2016: 385). This was clearly not the result of natural linguistic change because Bar-Rakib erected an earlier inscription for his father, Panamuwa II, in Samʾalian outside of the city. 15 It is also important to note in this regard that Bar-Rakib was the first Samʾalian king to style himself mlk šmʾl 'the king of Samʾal', using the name of the city and the etic-particularly Assyrian-designation for the kingdom. Prior to Bar-Rakib, Samʾalian kings such as Kulamuwa always styled themselves mlk yʾdy 'king of Yādiya' using the emic-and probably Luwian-name for the kingdom (Younger 2016: 378-83). Even in Bar-Rakib's memorial to his father, Panamuwa II is still styled mlk yʾdy 'king of Yādiya'. 16 Bar-Rakib's shift in language and title in his own inscriptions on the acropolis demonstrate a clear attempt to appease his Assyrian masters, since emissaries from Assyria were undoubtedly now among the elites visiting the Zincirli acropolis.
It is obvious that Bar-Rakib's inscriptions relied upon earlier Zincirlian monumental inscriptions. However, the new inscriptions have clearly developed their discourse in sophisticated ways and not simply copied the preceding ones. The question arises as to how this rhetoric was iterated and developed in the time intervening between the acts of inscription that produced the monuments. Over 100 years separate Kulamuwa's and Bar-Rakib's inscriptions, which are also written in two different languages. More than a generation separated Bar-Rakib's inscriptions from that of Panamuwa I, which he also appears to draw from and which is again in a separate language. The discourse used in Bar-Rakib's monuments must have been iterated elsewhere, but where? The answer lies in incantation texts and epistolary practice.
An Aramaic incantation discovered in Zincirli in 2017 sheds important light on how some elements of monumental rhetoric were practised in lieu of inscribing new monumental texts. The text has been preliminarily dated to the late ninth century based on its palaeography. The incantation noticeably opens with the phrase ʾn rḥm 'I am rḥm', utilizing the 'I Am' formula (Richey 2020: 23 n. 30). 17 To date, this is the only Northwest Semitic incantation text that has been discovered utilizing this formula, and it is also the only non-monumental text known to do so. This particular incantation was carved on a small piece of stone and so has been preserved, but it suggests a potentially wider practice of producing such incantations that relied on the conjuring effect of formulae from monumental inscriptions. Such practice may also be attested by the small-scale emulation of royal monuments by non-royal elites, as in the recently discovered Katumuwa Stele from the lower town of Zincirli, which utilizes the 'I Am' formula in addition to many other tropes from royal Samʾalian monumental discourse (Pardee 2009(Pardee , 2014Struble and Herrmann 2009;Herrmann 2014;Gilibert 2011: 95-6;Hogue 2019a). The overlap between monumental rhetoric and a personal incantation suggests a performative and specifically an oral setting for monumental writing. That is, the monumental text-like the incantation-was likely meant to be read aloud in order to have its desired effect.
Still more enticing are the significant parallels between monumental rhetoric and epistolary writing. It is worth noting in this regard that the origins of Levantine 'I Am' monuments lie in the monumental reproduction of letters. The 'I Am' formula first gained prominence as an opening for monumental inscriptions when it was adopted by the Hittite king Šuppiluliuma II at the end of the 13 th century. This formula replaced the earlier standard opening for Hittite monuments UMMA 'thus (says)', which was adapted directly from standard Hittite and Mesopotamian epistolary practice (Güterbock 1983: 21). In the same way that a letter reproduced an individual's direct speech in textual form, the Hittites conjured the voice of the king by presenting his monumental texts as though they were letters addressed to the public. This triggered the audience of such inscriptions to receive them as intimate communications from the king, who spoke to them through the monument with the immediacy of a letter. The last Hittite king replaced this with the 'I Am' formula because it essentially accomplished the same thing in a more direct way. It conjured the presence and voice of the king. After the Hittite Empire dissolved, its successor states in the Iron Age Levant and their neighbours maintained and popularized the 'I Am' formula as a standard opening for monumental texts (Hogue 2019b).
The 'I Am' formula was not the only element of Levantine monumental rhetoric to be inspired by epistolary practice, however. Monumental inscriptions also imitated the basic structure of letters. After the initial formulae introducing the direct speech of the sender, Levantine letters-much like their Mesopotamian counterparts-proceed into an account of the occasion for the letter, usually some past action or agreement that requires present action. The next section then requests an appropriate response on the part of the addressee. These two basic sections are typically divided either by means of deictic markers (most commonly 'now' or 'further') or graphically by means of horizontal lines (Fitzmyer 1974;Alexander 1978;Pardee 1978;Hawkins 2000: 538-42;Hawley 2003). Both of these methods were imitated by Zincirlian monumental inscriptions. For example, the two sections of the Kulamuwa Inscription discussed above were divided by a series of horizontal lines in clear imitation of paragraph markers on letters and other such documents (Figs. 1, 2). Even more strikingly, the two halves of the aforementioned Katumuwa Inscription are divided by the deictic marker wʿt 'and now', which is otherwise primarily encountered in Northwest Semitic letters (Pardee 2009: 63

The aesthetic dimension
Bar-Rakib's monuments attest a more complex visual repertoire than any of his predecessors. They are all wall reliefs-an epigraphic medium only shared with the monument of Kulamuwa at Zincirli. All other 'I Am' inscriptions from Samʾal and even Bar-Rakib's earlier monument to his father, Panamuwa II, are inscribed either on stelae or statues, which the Samʾalians described with the same term-nṣb 'standing monument'. This is significant because the medium of the relief allowed for more complicated adornment than did stelae and it required an entirely different kind of sculpture than that of statuary. Running the whole length of the first inscription, there is an image of Bar-Rakib processing with an attendant following close behind (Fig. 3). The king holds a fresh flower in his left hand; drooping flowers are well-known from mortuary scenes and indicated the deceased state of the individual. On the contrary, an upright flower seems to have been associated with living kings in the Levant (Van Loon 1986: 246-7). The king's right  hand is upraised in a gesture commonly associated with worship, and divine symbols extend from the king's eyeline above the inscription to indicate the gods to whom he is devoted. 19 Alongside the second inscription, we see the remains of an image of Bar-Rakib seated again holding a flower but also a cup, suggesting a banquet scene (Fig. 4).
Above the inscription, we see the same divine symbols as before. 20 The third inscription accompanies an audience scene of Bar-Rakib seated on his throne-again holding a flower-and receiving a dignitary (Fig. 5). This scene is remarkable in that the symbols for the Samʾalian gods are lacking, but a crescent on a staff indicating the Assyrian moongod Sîn is present. This last relief is also paired with an uninscribed image of Bar-Rakib seated at a banquet, holding a cup and flower as in the second relief. The first relief is 1.31m high and 0.62m wide, the last is 1.13m high and 1.15m wide, and though the second is fragmentary, the dimensions of the text and iconography suggest a size in proportion with the other two. All three reliefs were originally situated on low socles so that the eye level of the king in each would approximately match that of the viewer.
The writing of the inscriptions themselves is the first visual element we should consider.
In Zincirli-as in other Iron Age Levantine polities-writing had become an especially important part of monumental art, leading Alessandra Gilibert to conclude that it functioned as 'an image of itself ' (Gilibert 2011: 120). The majority of the monuments' viewers were non-literate, so they experienced writing primarily as visual media (Zilmer 2016: 209-10). Apart from discerning the verbal dimension of the text, viewers could still read the text as an additional image of Bar-Rakib and as an indication of his power and wisdom for having produced it (DeMarrais et al. 1996: 19;Glatz and Plourde 2011: 37-8 This horned crown is thought to represent the Storm-god Hadad, the yoke represents the dynastic god Rakib-El, the star may represent the god Rašap, the winged Sun-disc represents the sun-god, and the crescent represents the moon-god (Tropper 1993: 132). 20 In this case, one additional symbol has been added. This has been described as a 'doppelgesichtiger Januskopf ' and may represent the god El (Tropper 1993: 140). 21 Three fragmentary inscriptions of Bar-Rakib from the site attest the same character size. These appear to repeat some lines and themes from the more complete palace inscriptions and they were found in the same context, but they are too fragmentary to attach to any other inscription with confidence. It is possible that these are fragments of completely destroyed inscriptions from the palace (Tropper 1993: 147   The writing itself was not the only element of Bar-Rakib's reliefs to be influenced by Luwian epigraphic traditions. The processing figure of Bar-Rakib running the length of the first inscription is perhaps best explained as an imitation of Luwian 'amu-figures', especially those attested at Karkamiš. The amu-figure was a portrait of a king, queen, or other elite person that ran alongside a monumental text (Fig. 8). The figure clearly interacted with the text, but the text never crossed the figure. The figure could be seated, standing, or shown processing, but was always presented with the left arm raised and pointing to the nose. This is because at Karkamiš the amu-figure was actually a highly elaborate hieroglyph. The portrait was modeled on the hieroglyph EGO (Luwian amu 'I') and is typically designated EGO2. In the inscriptions it buttresses, it must be read as the first sign. The amu-figure is the 'I' of the 'I Am' formula at Karkamiš, and this is why it stands alongside the text without overlapping it (Payne 2016). It is a part of the text itself. At Zincirli, almost all of the monumental inscriptions are alphabetic, so this hieroglyph could not be adapted. The aesthetic dimension of the hieroglyph, however, could be and was emulated as an autonomous image of the king running alongside the alphabetic inscription (Gilibert 2011: 82, 87-8;Hogue 2019b: 332-3). 22 The image of Kulamuwa adjoining his inscription is undoubtedly an amu-figure, albeit one without any semantic significance (see Fig. 1). Bar-Rakib's imitation of Kulamuwa's inscription included his own iteration of an amu-figure, and one that was even more elaborate than that of Kulamuwa and actually more similar to the most sophisticated examples from Karkamiš (see Fig. 3).   The processional scene and amu-figure also indicate a significant aspect of artistic tradition at Zincirli and elsewhere in the Iron Age Levant. Monumental images were sometimes borrowed from luxury goods and vice versa. Irene Winter has demonstrated, for example, that the monumental images at Tell Halaf were mostly inspired by images from ivory furniture inlays (Winter 1989). Zincirli attests an even broader range of art objects that shared common tropes with monumental art. While it is difficult to determine whether monumental images were appropriated from small-scale luxury goods or vice versa, what is most important to note is that monumental productions at Zincirli were in conversation with other artistic productions and images were shared across different media. While the amu-figure in particular was certainly inspired by monumental writing, it was also iterated in smaller productions such as stamp seals (Fig. 9). This crosspollination of art forms at Zincirli will become even more apparent when considering Bar-Rakib's other wall reliefs.
Bar-Rakib's second wall relief is damaged, but the preserved image shows Bar-Rakib holding a cup and suggests a banquet scene. An intact but uninscribed relief from the palace likely shows Bar-Rakib in the same pose. Banquet scenes were particularly popular in the Northern Levant, where they typically adorned grave stelae. Most of these grave monuments were uninscribed (see Fig. 12), and only two examples from Zincirli include inscriptions. 23 In the first of these inscribed grave monuments-the Ördekburnu Stele (Fig. 10)-the inscription runs beneath the banquet scene at the top of the stele and does not intrude on the iconography. This stele dates to between 820 and 760 BC and may represent one of the earliest combinations of a banquet scene with an inscription (Lemaire and Sass 2013: 126). Dating to approximately a generation later between 743 and 733 BC, the Katumuwa Stele (Fig. 11) represents a new development in inscribed banquet scenes. On this stele, the text runs directly alongside the figure sitting at banquet in clear imitation of inscribed processional scenes such as that of the Kulamuwa Orthostat. This same style was repeated in the inscribed banquet reliefs of Bar-Rakib (Struble and Herrmann 2009: 20). Though mostly associated with grave monuments, the banquet scene was apparently more versatile. It is also attested on a gold amulet (Fig. 13)-perhaps once part of a necklace or bracelet-as well as on a stamp seal (Fig. 14). Bar-Rakib's appropriation 23 On northern Levantine grave monuments more broadly, see Bonatz (2000).
of the image in a political monument is unparalleled at Zincirli, and he may be attempting to draw upon the rich cultic significance of ritual feasting in establishing his palace as a centre for similar ceremonies (Denel 2007: 191-2;Gilibert 2011: 128-31). The final palace inscription is accompanied by an audience scene. Similar scenes are attested across the Ancient Near East, but those of the Levant show an important development. Scenes of a petitioner appearing before a figure on a throne often depict a god seated on the throne. In the course of Levantine artistic development, the god was eventually replaced by the enthroned king-an indication that religious art and ritual was being appropriated for political purposes (Denel 2007: 191-3). Such scenes were especially popular at nearby Karkamiš from the late tenth century onwards. Bar-Rakib's appropriation of the scene is the first such case for an inscribed wall relief at Zincirli. Nevertheless, the audience scene was already known from other artistic productions at Zincirli, such as amulets (Fig. 15) and cylinder seals (Fig. 16). Bar-Rakib may simply be incorporating a popular political image from neighboring Karkamiš, but given the prevalence of the scene as a religious icon at the site he may also be drawing upon its ritual significance as was the case with his use of the banquet scene.
The composition of the audience scene also implies a significant tension vis-à-vis Bar-Rakib's relationship to Assyria. A staff with a crescent moon-the icon of the Mesopotamian moon-god Sîn-appears directly below the inscription, which affirms that 'my lord is Baʿal-Ḥarrān'-an epithet for Sîn. Sîn was one of the chief deities of the Sargonid Assyrian kings, so these iconic and textual references likely function in part to demonstrate Bar-Rakib's loyalty to Assyria (T. M. Green 1992: 20-1). This possibility is also suggested by Sîn's role as the deity before whom Assyrian loyalty oaths were sworn (T. M. Green 1992: 39). Bar-Rakib is effectively indexing his vassalage to Assyria by drawing special attention to Sîn of Ḥarrān. Nevertheless, this accession to Assyrian interests is not all that is accomplished by the composition. Recalling that the scene on the wall relief was typically realized as one of religious devotion, the reader of the inscription might be tempted to assume that the enthroned figure is Sîn, but this is not the case. It is Bar-Rakib on the throne receiving supplicants. Though he implicitly acknowledges his subservience to Assyria, Bar-Rakib nevertheless draws upon the typical Northern Levantine artistic repertoire for demonstrating kingship, which involved even taking the traditional seat of a god. The richness of this composition and its multiple messages attest to the complexities of navigating Levantine kingship in the shadow of the Assyrian empire (Wicke 2015: 588).

The spatial dimension
As discussed above, Bar-Rakib's wall reliefs were not merely surfaces for presenting texts and images; they were also orthostats-essential architectural features of his palace. As such, they also accomplished an essential place-making function. Orthostats were traditionally employed only in ceremonial and public contexts. As such, their presence could indicate that a space was meant to be received by its visitors as a significant place of ceremonial engagement and public spectacle. This association was only strengthened when orthostats were appropriated to carry political messages in the form of pictorial and textual adornments. Orthostats marked the liminal portals and boundaries of ceremonial zones and theatres in the urban landscape (Harmanşah 2007: 80-4). Their presence indicated places where spectacles were meant to be witnessed. Once appropriated for royal inscriptions and images, orthostats became an essential means for a king to distribute his presence throughout his city's ceremonial centers (Gilibert 2011: 87-8). From this perspective, the appearance of ritual scenes on Bar-Rakib's palace orthostats was not merely an aesthetic choice. Such scenes reinforced the function of the space enclosed by the orthostats. The audience encountering the orthostats would actually process with Bar-Rakib into his palace where he received them and feasted with them (Gilibert 2011: 87). This participation with Bar-Rakib would occur in the imaginations of the audience if not also in reality.  Bar-Rakib's palace was a typical Syro-Anatolian design known as a hilani, a building consisting of a large portico and an enclosed space beyond that was separated into several rooms length-wise. The defining feature of the hilani-and in fact of Syro-Anatolian monumental architecture more broadly-was the framing of space using a portico decorated with carved orthostats (Pucci 2008: 176). Bar-Rakib's iteration of these features appear to be conscious enlargements of the structures from Kulamuwa's palace that also stood on the acropolis, with one very important difference. The portico Bar-Rakib constructed was an enlarged version of the entry façade into Kulamuwa's palace, but it opened into an open-air courtyard (Gilibert 2011: 88). Visitors could pass through this portico by means of only one portal at the westernmost end of the structure. They then had to proceed east past the portico in order to approach the hilani. Most significantly, this resulted in a segregation of space on the acropolis never attempted by the earlier kings of Samʾal (Fig. 17). Furthermore, this segregation was clearly symbolic in nature, as Bar-Rakib's new buildings were built over the top of previous citadel fortifications and actually weakened them. The dividing walls Bar-Rakib had constructed thus served no defensive purpose (Pucci 2008: 39).
The secondary segregation of the acropolis was actually a well-attested development in Syro-Anatolian architectural tradition. This is also seen at Tell Halaf and Tell Tayinat, for instance, though not at the same time as at Zincirli. Marina Pucci notes that the same division of palace space into two key sections is also encountered in Assyrian palaces, but she questions whether this could have influenced the Syro-Anatolian tradition because those developments occurred when Assyria was not dominant over Tell Halaf and Tell Tayinat (Pucci 2008: 174). This did occur during a period of Assyrian hegemony at Zincirli, however. Pucci also does not account for the fact that the Assyrians regularly received foreign diplomats at their palaces even when they were not actively pressuring the Levant. The Samʾalians and Karkamišeans, for example, sent diplomats to the Assyrian capital at Nimrud at the beginning of the eighth century, when neither polity was subject to Assyria (Aster 2016: 185). Both sites developed practices of spatial and ritual segregation during this period, however, possibly in imitation of what they had witnessed at Nimrud (Gilibert 2011: 129-31;Barjamovic 2011). So while the development undoubtedly proceeded through local autonomy, it was likely at least partially imitative of Assyrian practice.
The architectural setting for Bar-Rakib's monumental inscriptions was a significant aspect of their meaning affordance. In the first place, Bar-Rakib implied not only his legitimacy as king of Samʾal but also his superiority to his forefathers-Kulamuwa in particular-through his construction of an enlarged palace in the vicinity of Kulamuwa's (Gilibert 2011: 88). Bar-Rakib also implies the inviolability of his rule through the difficulty of access to his palace. During Kulamuwa's reign, a citizen or visitor to Zincirli would approach the acropolis along a main avenue leading from the city gate. Elite members of society could traverse this road to the northern end of the acropolis and then enter through a single citadel gate to find themselves in front of Kulamuwa's palace. During the reign of Bar-Rakib, however, after entering the acropolis, visitors would have to proceed from the eastern citadel gate to the western end of the acropolis to pass through another portal into a newly framed courtyard in Bar-Rakib's palace. They would then have to move back to the eastern end of the acropolis to reach the building where they would be received. The approach to Bar-Rakib's palace, in other words, is very indirect and progress towards it is checked by not one but two separate portals. 24 It is far more removed from the rest of the town than was Kulamuwa's palace.
The shortest Bar-Rakib palace inscription (KAI 218) and its paired uninscribed orthostat flanked the entrance to the enclosed building on the northeast side of the acropolis' southern courtyard. The other two palace inscriptions (KAI 216-217) were not found in situ, but they undoubtedly also functioned as a pair flanking a portal within Bar-Rakib's palace (Gilibert 2011: 87-8). Based on the content of the first inscription-which gestures first to hʾ byt klmw 'that palace of Kulamuwa' (KAI 216:17) and then to bytʾ znh 'this palace (of Bar-Rakib)' -I propose that these portal orthostats must have stood in a location where these deictic references actually made sense. In other words, they must have been located at a portal where both the palace of Kulamuwa and the palace of Bar-Rakib were visible. The only candidate is thus the portal on the western end of the portico dividing the acropolis that Bar-Rakib had built. This westernmost portion of the portico does attest stone socles with traces of orthostats as well as peg-holes for holding them (Gilibert 2011: 85). I suggest that this was where the first two orthostats of Bar-Rakib must have been located in antiquity, thus framing the transition between the northern courtyard marked by Kulamuwa's palace and the southern courtyard marked by Bar-Rakib's palace. This is reinforced by the processing image of Bar-Rakib: he appears to walk with the audience through the portal into his palace, where he can be found in further reliefs waiting to receive and feast with them.
The movement through the space around Bar-Rakib's monuments also created a new frame and reception for Kulamuwa's monument. Before Bar-Rakib remodelled the acropolis, Kulamuwa's Orthostat would have been the central focal point viewed by all visitors before they entered Samʾal's primary palace. Bar-Rakib's construction, however, reframed that orthostat as only a waypoint en route to the new administrative and ritual center of the city. Visitors now had to pass the Kulamuwa Orthostat and proceed to the opposite side of the acropolis. From there, the inscription of Bar-Rakib gestured dismissively toward the monumental constructions of his predecessor. The use of distal deixis in the inscription to mark Kulamuwa implies ideological distance, which is confirmed by Bar-Rakib's assessment that Kulamuwa's palace did not amount to a 'suitable palace' (KAI 216: 16). Furthermore, in gesturing towards Kulamuwa's palace, Bar-Rakib also indicates his orthostat and inscription, which were inextricable parts of the older palace. As a result, all of Kulamuwa's achievements and ideology are called into question.
The motion implied by the deictic references in Bar-Rakib's inscription is not insignificant. By gesturing towards the palace of Kulamuwa, the visitors about to pass through the northwestern portico into the southern courtyard are invited to turn back toward the palace they just passed and assess it from Bar-Rakib's point of view. The processing figure of Bar-Rakib and his proximal indication of his own palace in the southern courtyard invite the visitors to turn their backs on Kulamuwa and walk with the image of Bar-Rakib into Zincirli's new order. The transition between the two courtyards is thus transformed into the transition between the two periods of Samʾal's history. Bar-Rakib may have been a vassal king, but he claimed through his monument that he was superior to his independent predecessors. The spatial aspects of his inscriptions coax the viewers of the monuments to leave the ideology of the independent Yādiya behind and progress with Bar-Rakib into a new era of Assyrian hegemony over the vassal state of Samʾal. Moving through each successive portal on the acropolis, the visitors are led closer and closer to Bar-Rakib. This motion implies a growing intimacy with the new king and, as a result, a growing nearness to his expressed ideology. The direction of the audience's gaze and movement through the space invited them to participate in patterned ritual spectacle (Ingold 2004: 328;Hodder 2006: 82, 96;Gilibert 2013: 47-8).

Conclusion: How texts were monumentalized in the Iron Age Levant
Bar-Rakib's palace orthostats appropriated elements of previous monumental rhetoric from Zincirli and the greater region in tandem with aspects of other media in order to afford meaning to a particular community: the elites of Samʾal and Bar-Rakib's Assyrian benefactors. Bar-Rakib's monumental discourse departed in some key ways from that of earlier monuments in order to appease his overlord. Nevertheless, he maintained some very traditional elements of Samʾalian monumental discourse, allowing him to defend his legitimacy to local Samʾalian elites and even to project his superiority over his predecessors. Bar-Rakib's palace orthostats thus constitute an attempt to appeal to the two communities that most threatened his rule and who most needed to be convinced of his legitimacy. It is difficult to say whether Bar-Rakib's monuments were successful in their apparent purpose. Upon his death, Samʾal ceased to be a vassal state and was formally annexed as a province of Assyria (Younger 2016: 422). Bar-Rakib's vision of an autonomous Samʾal under the aegis of Assyria was thus only short-lived at best.
Whether or not they were successful, the concerted effort of Bar-Rakib's orthostats reveal many significant features of the monumentalization of text in the Iron Age Levant. The orthostats were monumentalized by combining the artistic features they appropriated into unified pieces of mixed media that presented a complex of messages to their target audiences. It was the juxtaposition of a text with visual media and architectonic contexts that rendered it capable of relating to communities and prompting them to renegotiate their social and material relationships. As the monumental texts were brought into new relationships with materials, images, and places, they promoted new social relationships in the targeted audiences. Furthermore, these relationships were always evaluated according to Bar-Rakib's ideology, whether by the strategic use of deixis in his inscriptions, the recontextualization of known visual motifs in his iconography, or the placement of his monuments relative to those of his predecessors. It was this ability to reshape communal relationships and prompt their ideological evaluation that rendered texts monumental in the Iron Age Levant.