[JNES 78 no. 2 (2019)] © 2019 by The University of Chicago. All rights reserved. 0022–2968/2019/7802–0004$10.00. DOI: 10.1086/704675
“I Am”: The Function, History, and
Diusion of the Fronted First-Person
Pronoun in Syro-Anatolian Monumental
TIMOTHY HOGUE, University of California, Los Angeles*
The practice of opening monumental inscriptions
with a ﬁrst-person pronoun was popularized by the
Iron Age Syro-Anatolian polities, which inherited the
tradition from the Hittites. The ﬁrst-person pronoun
evoked the commissioner’s voice and even his image,
especially in Hieroglyphic Luwian iconography and its
adaptations. These monumental texts materialized an
imagined encounter between readers and commission-
ers that was initiated by the phrase “I am.” The ﬁrst-
person opening thus became the operative element of
these texts’ monumentality. The texts only functioned
as monuments in light of the speakers identiﬁed by
the pronominal opening. This study presents a history
of the practice and especially its employment relative
to monumental images. This reveals that the formula
had an overlapping but ultimately separable function
from that of images, allowing it to imbue texts with
monumentality in a variety of contexts. This appar-
ently made the formula an attractive object of adap-
tation during the Iron Age, leading to its diusion
throughout the greater Levant and in Mesopotamia.
During the Iron Age, a fronted ﬁrst-person pro-
noun became the standard opening for Syro-Anatolian
monumental inscriptions. In no other region or time
period did the opening formula “I am so-and-so”
rise to such prominence, though it is sparsely attested
elsewhere from the Bronze Age to the Hellenistic
period. While previous scholarship has explored top-
ics such as the formula’s origins and its function in
speciﬁc iterations, no study has yet fully elucidated
the under lying function of the fronted ﬁrst-person
pronoun that transformed it from a marginal innova-
tion into a standard feature of monumental discourse.
In every iteration, it is apparent that the pronominal
formula indicates some aspect of the monument, but
what it indicates is inconsistent. More importantly,
why the beginning of an inscription should attempt
to indicate anything in the ﬁrst place has yet to be
explored. This paper will provide an overview of the
history of the formula with a view towards elucidating
* This article beneﬁtted greatly from helpful comments from
William Schniedewind, Craig Melchert, David Hawkins, Jeremy
Smoak, Virginia Herrmann, Mary Bachvarova, and Aslı Özyar. I
would also like to thank the organizers and other participants in the
conference “Beyond All Boundaries: Anatolia in the 1st Millennium
BCE” for their feedback on my presentation of this material in June,
2018 in Ascona, Switzerland. I am also grateful to the anonymous
reviewers of this article, especially for drawing my attention to the
Kapar(r)a Inscription. Any remaining errors are solely my own. In
addition, I am grateful to the UCLA Alan D. Leve Center for Jewish
Studies for funding this research.
324 ) Journal of Near Eastern Studies
an underlying function for these seemingly inconsis-
tent employments. I argue that this function was not
only to indicate various aspects of the monument, but
to actually produce the presence of the monumental
text’s implied speaker through the strategic use of de-
ixis. The “I (am)” formula thus became a signiﬁcant
carrier of the text’s monumentality by absorbing one
of the key functions otherwise reserved for images.
Review of Previous Scholarship
Engagements with the “I (am)” formula have tended
to approach it in terms of its semantic range without
venturing too deeply into its relationship to the monu-
mental texts it opens.1 Some few exceptions have ex-
plored the other dimensions of the formula, however.
Sanna Aro argued that the formula was introduced to
create a clear interface between a monumental text and
a statue. According to this understanding, the ﬁrst-
person pronoun is a deictic reference to the statue,
and that apart from a statue, the pronoun would be
obscure.2 However, the formula cannot be aligned
with statues in every case, so Aro concludes that the
formula lost its meaning during the Iron Age when
it became a standard part of monumental discourse.3
This tends to subordinate the clearly iconic function
of the pronoun in later examples, so Annick Payne
has suggested as an alternative that the pronoun—
especially as rendered in Anatolian hieroglyphs—must
have served some iconic purpose on its own.4 Her
conclusion matches Seth Sanders’ brief comments on
the formula. He argues that the result of the formula
is that “the inscription now designates itself by the
speaker,” as opposed to by the monumental object.5
The formula thus still designates something, but not
necessarily a statue.
1 The only such study that deals exclusively with the fronted ﬁrst-
person pronoun is that of Arno Poebel, Das appositionell bestimmte
Pronomen der 1. Pers. Sing. in den westsemitischen Inschriften und
im Alten Testament (Chicago, 1932).
2 Sanna Aro, “Carchemish Before and After 1200 BC,” in Lu-
wian Identities: Culture, Language and Religion Between Anato-
lia and the Aegean, ed. Alice Mouton, Ian Rutherford, and Ilya
Yakubovich (Leiden, 2013), 241–44.
3 Ibid., 238.
4 Annick Payne, “The Hieroglyphic Sign EGO(2),” in Audias
Fabulas Veteres. Anatolian Studies in Honor of Jana Součková-
Siegelová, ed. Šárka Velhartická (Leiden, 2016), 293.
5 Seth Sanders, The Invention of Hebrew (Urbana, IL, 2009),
Apart from these attempts to locate the iconic-
ity of the “I (am)” formula, previous studies have
generally concluded that the formula serves to imbue
monumental texts with some of the qualities of di-
rect speech. Both Hans Güterbock and Payne have
previously suggested that the most basic function of
the ﬁrst-person pronoun in Anatolian inscriptions was
to produce speech.6 This was marked in an especially
explicit fashion in the Anatolian examples by the ever-
present quotative particle (-wa-).7 Similarly, Sanders
has argued that the use of the ﬁrst-person pronoun in
Northwest Semitic inscriptions “ventriloquized” the
speaker “as if he were standing right in front of us.”8
The use of the ﬁrst-person pronoun thus provided
the monumental texts with an aspect of orality that
shifted the focus of the monument from the object to
the imagined speech.
Previous approaches are united in understanding
that the formula is used to identify the monument
with the implied speaker. Where these approaches dif-
fer is in identifying how the formula relates to the
monument as a whole, and especially whether the
formula should be approached in terms of iconicity
or implied orality. In fact, the “I (am)” formula com-
bined aspects of the iconic and the oral in order to
produce the presence of the agent behind the monu-
ment, or “monumenter”—that is, the ﬁctionalized
version of the monument’s commissioner or implied
speaker that the object presents to its users through
both textual and iconic means.9 The function of the
formula is thus neither primarily iconic nor oral but
rather monumental, and it will best be elucidated by
an approach couched in a theory of monumentality.
6 Hans G. Güterbock, “The Hittite Conquest of Cyprus Re-
considered,” JNES 26/2 (1967): 74, and “Hittite Historiography:
A Survey,” in History, Historiography and Interpretation: Studies
in Biblical and Cuneiform Literatures, ed. Hayim Tadmor and
Moshe Weinfeld (Jerusalem, 1983), 21; Payne, “Hieroglyphic Sign
7 Annick Payne, Hieroglyphic Luwian: An Introduction with
Original Texts (2nd Revised Edition), 2nd ed., Subsidia et Instru-
menta Linguarum Orientis 2 (Wiesbaden, 2010), 40.
8 Sanders, Invention of Hebrew, 114; Seth Sanders, “Naming
the Dead: Funerary Writing and Historical Change in the Iron Age
Levant,” MAARAV 19/1–2 (2012): 35.
9 In what follows, “agent” refers to this person. Timothy Hogue,
“Abracadabra or I Create as I Speak: A Reanalysis of the First Verb
in the Katumuwa Inscription in Light of Northwest Semitic and
Hieroglyphic Luwian Parallels,” BASOR 381 (2019): 58.
Fronted First-Person Pronoun in Syro-Anatolian Monumental Discourse ) 325
Textual Monumentality, Cognitive
Poetics, and Deictic Projection
Monumentality may be deﬁned as the aspect or aspects
of an object by which it aords special meaning to
a community of users. In order to function, monu-
ments must be interactive in such a way as to provoke
the imagination of their audiences.10 Monuments then
are any objects which aord special meaning to the
communities in which they are embedded. Texts—es-
pecially when they become objects of communal inter-
pretation—are ripe for monumentalization.11 Textual
monuments may aord meaning through their seman-
tic content, rhetorical structure, integration into larger
monumental installations, iconic dimension, attendant
ritual actions, and even features such as orthography.12
Among other means, textual monumentality—the
aordance of specialized communal meaning—can be
produced through the strategic employment of deic-
tic elements. Kristel Zilmer has argued that Viking
runestones materialized speech through the use of
deictic referents. Speciﬁcally, references in these in-
scriptions such as “this stone” or “east/west” would
be ambiguous unless an implied speaker was imag-
ined at the location of the monument. Through the
monument, a voice is able to gesture to itself and
its surroundings using those deictic elements. The
result is an imagined face-to-face encounter with
the implied speaker of the monument.13 In Zilmer’s
10 Hung Wu, Monumentality in Early Chinese Art and Architec-
ture (Stanford, CA, 1995), 4; James F. Osborne, “Monuments and
Monumentality,” in Approaching Monumentality in Archaeology,
ed. James F. Osborne (Albany, NY, 2014), 3–4; Timothy Pauketat,
“From Memorials to Imaginaries in the Monumentality of Ancient
North America,” in Approaching Monumentality, ed. Osborne, 442.
11 Ann Rigney, “Portable Monuments: Literature, Cultural
Memory, and the Case of Jeanie Deans,” Poetics Today 25/2 (2004):
383; Ann Rigney, “The Dynamics of Remembrance: Texts Between
Monumentality and Morphing,” in A Companion to Cultural Mem-
ory Studies: An International and Interdisciplinary Handbook, ed.
Astrid Erll and Ansgar Nünning (Berlin, 2010), 349.
12 James W. Watts, “The Three Dimensions of Scriptures,” in
Iconic Books and Texts (Sheeld, 2013), 14–16; Edmund Thomas,
“The Monumentality of Text,” in Approaching Monumentality, ed.
13 Kristel Zilmer, “Deictic References in Runic Inscriptions
on Voyage Runestones,” Futhark: International Journal of Runic
Studies 1 (2010): 138; Kristel Zilmer, “Viking Age Rune Stones in
Scandinavia: The Interplay between Oral Monumentality and Com-
memorative Literacy,” in Along the Oral-Written Continuum: Types
of Texts, Relations and Their Implications, ed. Slavica Rankovic,
Leidulf Melve, and Else Mundal, Utrecht Studies in Medieval Lit-
eracy 20 (Turnhout, 2010), 152.
terms, deixis “anchors the textual reference in a real
and physical environment and creates the image of an
immediate encounter between the commissioners of
the memorial and the audience.”14 In a similar vein,
Stephen Houston and David Stuart argue that the use
of personal deixis in Mayan monumental inscriptions
“accentuates the intimate oration directed to a living
actor by a sculpted image.”15 For example, one such
inscription refers to “yourself,” “your penance,” and
“your darkness,” implying that there must be a “you”
outside of the monument who is being addressed.
The use of deixis in monumental texts thus provokes
communal interaction with the monument, which is
an essential prerequisite for the imbuing of an object
Beyond mere gesturing or indication, the process
of evoking an imagined encounter through deixis is
essentially what has been described in the ﬁelds of cog-
nitive poetics and cognitive science as deictic projec-
tion. Vimala Herman describes this cognitive process
as one in which “deictic triggers project mental space
scenarios in which dierent forms of intimacy are cre-
ated as shared deictically.”17 In other words, the use
of the personal deictic element “I” actually produces
the presence of the implied speaker by conjuring them
within the imagination of the audience. The monu-
ment’s users project themselves into an imagined space
in which both they and the implied speaker are pres-
ent in an imagined face-to-face encounter.18 Sanders
reached nearly the same conclusion when he argued
that such language in monumental inscriptions served
to “produce the presence” of agents and allow them
to “make demands for themselves.”19 The “I (am)”
formula symbolically manifested the agent before the
14 Ibid., 152.
15 Stephen Houston and David Stuart, “The Ancient Maya Self:
Personhood and Portraiture in the Classic Period,” RES: Anthro-
pology and Aesthetics 33 (1998): 88; Stephen Houston, “Imper-
sonation, Dance, and the Problem of Spectacle Among the Classic
Maya,” in Archaeology of Performance: Theaters of Power, Commu-
nity, and Politics, ed. Takeshi Inomata and Lawrence S. Coben
(Lanham, MD, 2006), 142.
16 Wu, Monumentality in Early Chinese Art and Architecture,
4; Osborne, “Monuments and Monumentality,” 3–4; Pauketat,
“From Memorials to Imaginaries,” 442.
17 Vimala Herman, “Deictic Projection and Conceptual Blend-
ing in Epistolarity,” Poetics Today 20/3 (1999): 539; Mark Turner,
The Literary Mind: The Origins of Thought and Language (Oxford,
18 Herman, “Deictic Projection”: 524–30; Peter Stockwell, Cog-
nitive Poetics: An Introduction (London, 2002), 43–49.
19 Sanders, “Naming the Dead”: 35.
326 ) Journal of Near Eastern Studies
monument’s users and aorded special meaning by
creating an imagined encounter.
As a means of deictic projection, the “I (am)”
formula played a largely similar role to that of mon-
umental images in the ancient Near East. Images
functioned by provoking collective interaction with
themselves.20 This was accomplished because images
were not merely representations but rather manifes-
tations. They acted as metonymic extensions of the
persons they purported to depict, producing aspects
of that person’s presence and even agency apart from
their physical bodies.21 Zainab Bahrani thus refers to
images as “modes of presencing,” or “ways of en-
countering that person.”22 Aro was correct to note
this function for the images accompanied by the “I
(am)” formula in Syro-Anatolia, but she did not con-
nect it to the formula itself.23 This is in fact the same
function accomplished by the formula, albeit by dif-
“I (Am)” in the Bronze Age
With the monumental function of the “I (am)” for-
mula in mind, we may turn to its history to better
elucidate this function in the context of speciﬁc ex-
amples. The “I (am)” formula likely originated as
an innovation in Akkadian based on the earlier use
of I-deixis in Sumerian inscriptions. In these inscrip-
tions, the ﬁrst clause would typically end with a ﬁrst-
person pronominal sux -me-en or copula -i-me-en,
20 Ömür Harmanşah, “‘Source of the Tigris’: Event, Place and
Performance in the Assyrian Landscapes of the Early Iron Age,”
Archaeological Dialogues 14/2 (2007): 179–204; Marian H. Feld-
man, “Object Agency? Spatial Perspective, Social Relations, and the
Stele of Hammurabi,” in Agency and Identity in the Ancient Near
East: New Paths Forward, ed. Sharon R. Steadman and Jennifer C.
Ross (London, 2010), 158–59; Irene J. Winter, “Agency Marked,
Agency Ascribed: The Aective Object in Ancient Mesopotamia,”
in On Art in the Ancient Near East, Culture and History of the
Ancient Near East, 34/2 (Leiden, 2010), 308.
21 J. N. Postgate, “Text and Figure in Ancient Mesopotamia:
Match and Mismatch,” in The Ancient Mind: Elements of Cogni-
tive Archaeology, ed. Colin Renfrew and Ezra B. W. Zubrow (Cam-
bridge, 1994), 177–78; Tom Hare, ReMembering Osiris: Number,
Gender, and the Word in Ancient Egyptian Representational Systems
(Stanford, CA, 1999), 45–46; Zainab Bahrani, The Graven Image:
Representation in Babylonia and Assyria (Philadelphia, 2003), 128;
Feldman, “Object Agency?”, 158–59; Kathryn E. Slanski, “The Law
of Hammurabi and Its Audience,” Yale Journal of Law & the Hu-
manities 24/1 (2012): 107.
22 Bahrani, Graven Image, 135, 128, respectively.
23 Aro, “Carchemish Before and After 1200 BC,” 232.
both of which may be translated “. . . (am) I.”24 This
was adapted into Akkadian inscriptions by closing the
ﬁrst clause with the pronoun anāku, a practice which
continued from the Bronze Age to the Hellenistic
period.25 In these cases, the primary function of the
formula was to manifest the agent by means of deictic
projection, imbuing the text with orality.
There are only two Bronze Age examples of the
fronting of anāku resulting in the ﬁrst true examples
of the “I (am)” formula, both dating to the 15th cen-
tury BC. These suggest that the true innovation of the
“I (am)” clause was not the use of I-deixis per se, but
rather the fronting of the deictic referent. Rather than
opening with the agent’s name as in the earlier tradi-
tion, the inscription now began with a pronoun that
demands further explanation. The inscription thus no
longer primarily designated itself by the speaker, but
by the implied speech itself. This is most purely dem-
onstrated by ZA 31, an inscription that has yet to be
brought into conversation with the Syro-Anatolian
material. It opens as follows:
(1–4)a-na-ku mk[aš]-til-ia-šu ⸢ŠAGINA? den?-líl ?⸣
DUMU m[b]ur-na-bu-ra-ri-ia-aš [D]UMU DUMU [š]
a m⸢a⸣-gu-um aš-ru [m]u-ṭi-ib ŠÀ den-líl
(1–4)I am K[aš]tiliašu, Governor of Enlil, son of
Burna-Burariaš, grandson of Agum, the humble,
who makes Enlil happy.26
The inscription has been identiﬁed as a monumental
inscription of Kaštiliašu III dating to the early 15th
century BC.27 The inscription is recorded in lapidary
script on a clay tablet of unknown provenance. Kath-
leen Abraham and Uri Gabbay argue that the tablet
likely records the same inscription as that set on a
spade and basket ritually emplaced in the temple of
Enlil, the dedication of which was explicitly narrated
in the inscription.28 The formula here functions en-
24 Alexa Bartelmus, “Restoring the Past. A Historical Analysis
of the Royal Temple Building Inscriptions from the Kassite Pe-
riod,” KASKAL: Rivista di Storia, Ambienti e Cultura del Vicino
Oriente Antico 7 (2010): 149; Poebel, Das appositionell bestimmte
Pronomen der 1. Pers. Sing., 3–6.
25 Ibid., 84–86; Susan Sherwin-White, “Aspects of Seleucid
Royal Ideology: The Cylinder of Antiochus I from Borsippa,” The
Journal of Hellenic Studies 111 (1991): 77.
26 The transcription and translation here follow the edition of
Kathleen Abraham and Uri Gabbay, “Kaštiliašu and the Sumundar
Canal: A New Middle Babylonian Royal Inscription,” Zeitschrift Für
Assyriologie 103/2 (2013): 184.
27 Ibid.: 187–89.
28 Ibid.: 184.
Fronted First-Person Pronoun in Syro-Anatolian Monumental Discourse ) 327
tirely by itself to produce the presence of the king.
The key dierence between this inscription and other
Mesopotamian monumental inscriptions from this
period is the foregrounding of the personal deictic
element. As a result, the process of deictic projection
was emphasized over the production of orality sug-
gested by the use of direct speech without a fronted
The second 15th-century example is in the Idrimi
Inscription from Alalaḫ. While there was potentially
some contact between Kassite Babylonia and Alalaḫ,
whether the Alalaḫites learned the formula from the
Kassites or arrived at the same innovation indepen-
dently is impossible to say.30 The inscription opens:
(1–2)a-na-ku mid-ri-mi DUMU mDINGIR-i-lim-ma
ARAD d⸢IM⸣ dḫe2-bat u3 diš8-tar2 NIN iria-la-la-aḫ
(1–2)I am Idrimi, son of Ilī-ilimma, servant of
the Storm-god, Hebat, and Ištar, the lady of
Alalaḫ, the lady, my lady.31
In contrast to the Kassite example, Aro concluded that
the “I (am)” formula in this inscription was meant to
indicate the statue.32 She is not the only one to have
suggested such an interface between text and image.
Tremper Longman also noted that the ﬁnal lines of the
inscription run along the statue’s face to the mouth,
implying that the inscription should be understood
as the statue’s speech.33 A connection between the
fronted ﬁrst-person pronoun and the statue on which
it is inscribed seems likely in this case. However, the
Kassite example and earlier Sumerian practice dem-
onstrate that this connection between text and statue
was an innovation at Alalaḫ. The “I (am)” formula
29 It should be noted that this was a very short-lived innovation.
Not only did the Kassites abandon the fronted ﬁrst-person pronoun,
later Kassite royal inscriptions abandoned I-deixis altogether, and
were instead narrated in the third person. Bartelmus, “Restoring the
Past,” 149–50; Abraham and Gabbay, “Kaštiliašu and the Sumundar
30 J. A. Brinkman, “Foreign Relations of Babylonia from 1600
to 625 B.C.: The Documentary Evidence,” American Journal of
Archaeology 76/3 (1972): 274.
31 The transcription and translation here follow the edition of
Jacob Lauinger, “Statue of Idrimi,” Oracc: The Open Richly Anno-
tated Cuneiform Corpus, accessed July 10, 2017, http://oracc.org/
aemw/alalakh/idrimi/X123456/html. I dier only in translating
dIM (as “Storm-god”) where Lauinger does not.
32 Aro, “Carchemish Before and After 1200 BC,” 237.
33 Tremper Longman, Fictional Akkadian Autobiography: A Ge-
neric and Comparative Study (Winona Lake, IN, 1991), 60.
did not originate as such to interact with statuary.
Rather, this usage of the formula suggests that the
Alalaḫites emphasized its manifestational function by
employing it in tandem with a monumental form that
would accomplish roughly the same thing. The statue
and formula together produced the presence of the
king in this case.
Aro argues that the appearance of the “I (am)” for-
mula in the Idrimi Inscription may explain the short-
lived Hittite adoption of the formula. The formula is
attested in Hittite in KBo 12.38, a conquest account
of Šuppiluliuma II dating to the late 13th or early 12th
century BC. In lines 4–21 of column II, Šuppiluliuma
relates having created and set up an image to com-
memorate his conquest of Cyprus. This description is
followed by a double line on the tablet, indicating that
a new text is to follow.34 This new text opens:
(22–26)ú-uk-za dUTU-ŠI Ta-bar-na-aš mKÙ.GA.
[TÚ]L-aš LUGAL.GAL LUGAL KUR uru[Ḫa]t-ti UR.SAG
DUMU mTu-ud-ḫa-li-ya LUGAL.GAL LUGAL KUR
Ḫat-ti UR.SAG [DUMU.D]UMU-ŠÚ ŠA MPA-ši-ILILI(M)
(22–26)I am My Sun, the Tabarna Šuppiluliuma,
the Great King, king of Hatti, the hero, son of
Tudḫaliya, the Great King, king of Hatti, the
hero, grandson ` Ḫattušili, the Great King, the
In this case, the formula interacts with an image in a
literary context. Güterbock argued that this must be
a cuneiform copy of a hieroglyphic inscription set up
by Šuppiluliuma, speciﬁcally the NIŞANTAŞ inscription.36
The opening of this hieroglyphic inscription was pre-
viously restored as EGO-wa/i-mi-*a “I (am) . . .,”
followed by Šuppiluliuma’s name and titles as they
appear in KBo 12.38.37 This reading has ﬁnally been
conﬁrmed by 3D scans of NIŞANTAŞ.38 Furthermore,
Gütterbock argued that column I of the same text
contained a parallel inscription devoted to Tudḫaliya,
likely originally inscribed on the statue of Tudḫaliya
that Šuppiluliuma claimed to have set up in the text.39
34 Güterbock, “Hittite Conquest of Cyprus Reconsidered”: 74.
35 Transcription and translation follow Güterbock’s edition,
36 Ibid: 81.
37 Ibid: 81; E. Laroche, “Nişantaş,” Anatolica 3 (1970): 93;
Payne, “Hieroglyphic Sign EGO(2),” 284.
38 Andreas Schachner et al., “Die Ausgrabungen in Boğazköy-
Ḫattuša 2015,” Archäologischer Anzeiger 1 (2016): 31–32.
39 Güterbock, “Hittite Conquest of Cyprus Reconsidered”: 75.
328 ) Journal of Near Eastern Studies
Aro connects these references to images to a seated
statue of Tudḫaliya, which Heinrich Otten argued
closely paralleled the statue of Idrimi.40 On this basis,
Aro concludes that the inscription of Tudḫaliya must
have also begun with the “I (am)” formula, and that
the Hittites derived this formula from the practice at
Aro’s connection of the Hittite formulae to statues
is less certain, however. No statue is preserved for
the NIŞANTAŞ inscription, and KBo 12.38 is obviously
a purely textual reference. Aro suggests that the ac-
companying statues were likely carried away near the
end of the empire period, but this argument must
be relegated to conjecture.41 Payne has alternatively
argued that the formula’s potential reference to an
image must augment the meaning the formula car-
ried on its own.42 This more conservative conclusion
is consistent with the other Bronze Age exemplars, as
well as with the great diversity seen in the subsequent
use of the formula during the Iron Age.
“I (Am)” in the Iron Age
During the Iron Age, the “I (am)” formula became
the standard opening for Syro-Anatolian monumental
inscriptions, and it also diused out into the monu-
mental discourse of the southern Levant and Mesopo-
tamia.43 The formula is mostly attested in Hieroglyphic
Luwian, with examples from Carchemish,44 Hama,45
Tabal,46 Maraş,47 Malatya,48 Tell Ahmar,49 Aleppo,50
40 H. Otten, “Neue Quellen zum Ausklang des Hethitischen
Reiches,” Mitteilungen Des Deutschen Orient-Gesellschaft 94
(1963): 17; Aro, “Carchemish Before and After 1200 BC,” 241.
41 Aro, “Carchemish Before and After 1200 BC,” 244.
42 Payne, “The Hieroglyphic Sign EGO(2),” 293.
43 Aro, “Carchemish Before and After 1200 BC,” 236.
44 ADANA 1; CEKKE; KARKAMIŠ A1b, A2+3, A5b, A6, A11a, A11b+c,
A12, A13d, A14a, A14b, A15b, A17a, A18a, A23; KELEKL; KÖRKÜN.
BEIRUT is only known from a photograph and is of unknown prov-
enance, but some stylistic features align it with the inscriptions of
Carchemish. John David Hawkins, “Corpus of Hieroglyphic In-
scriptions, I. Inscriptions of the Iron Age,” in Corpus of Hieroglyphic
Luwian Inscriptions, Volume I (Berlin, 2000), 558.
45 HAMA 1–4, 6–8; HINES; QAL’AT EL MUDIQ; RESTAN; SHEIZAR;
46 ANDAVAL; BOHÇA; possibly BOR; BULGARMADEN (non-logo-
graphic); ÇFTLK; possibly EĞREK; HSARCIK 1–2; KAYSER; KULULU
1–4; PORSUK; SULTANHAN.
47 KÜRTÜL; MARAŞ 1, 2, 4, 8, 13, 14.
48 DARENDE; SPEKÇÜR; IZGIN 1; PALANGA.
49 ALEPPO 2; BOROWSKI 3; TEL AHMAR 1, 2, 5, 6.
50 ALEPPO 6; BABYLON 1.
Amuq,51 and Cilicia.52 A smaller set are the Northwest
Semitic examples from Cilicia,53 Hama,54 Zincirli,55
and Moab.56 Tabal, Malatya, and Aleppo together
with Carchemish inherited some of the administrative
framework of the Hittite Empire, including early rul-
ers with connections to the royal family or perhaps its
cadet branch at Tarḫuntašša.57 These states continued
to develop traditions from the empire period, which
were shared among the neighboring states—a diu-
sion that included iterations of the “I (am)” formula
in new languages. Because the formula was so much
more widespread during the ﬁrst half of the ﬁrst mil-
lennium BC, I will only analyze select examples that
further illustrate the formula’s underlying function.
The 11th Century
After the dissolution of the Hittite empire ca. 1200
BC, the “I (am)” formula passed into Neo-Hittite us-
age. Potentially the earliest Neo-Hittite exemplar is
ALEPPO 6, one of the Aleppo temple inscriptions of
Taita, king of Palistin, dating to the 11th century BC.
It begins as follows:
§1 REX Itá-i-tá-sa EGO-wa/i-mi (VIR2) HEROS
(VIR2) pa-lá/í-sà-ti-[ní]-za-sa REX
§1 I am King Taita, the Hero, Palistinean
The inscription is clearly stylistically later than the
empire-period sculptures in the Aleppo temple, but
it was placed alongside the installations of that era, and
may represent an attempt to emulate Hittite style.59
This may explain the use of the “I (am)” formula. In
this case, the inscription clearly accompanies an image
of Taita in raised relief, with the “I (am)” formula and
its accompanying name and titles appearing to caption
51 ARSUZ 1–2; KRCOĞLU.
52 ÇINEKÖY; KARATEPE 1; DOMUZTEPE 1–2.
53 ÇINEKÖY and KARATEPE 1 are both bilinguals with Phoenician
versions. The Phoenician version of KARATEPE 1 is designated KAI 26.
54 KAI 202.
55 KAI 24, KAI 214, KAI 216–18, and the Katumuwa Inscription.
56 KAI 181, 306.
57 Annick Payne, Iron Age Hieroglyphic Luwian Inscriptions (At-
lanta, 2012), 4–8; Hawkins, “Corpus of Hieroglyphic Inscriptions,
I.,” 73–77, 283–87, 388–91, 429–30.
58 The transcription and translation are based on Hawkin’s edi-
tion. J. D. Hawkins, “The Inscriptions of the Aleppo Temple,” Ana-
tolian Studies 61 (2011): 44–45.
59 Ibid.: 38–41.
Fronted First-Person Pronoun in Syro-Anatolian Monumental Discourse ) 329
the image.60 The image of the king may augment the
formula in this case, but the formula is also acting
on its own to a degree. This is especially indicated
by the placement of the EGO hieroglyph. Though the
hieroglyph EGO appears to open the third word in the
inscription, Payne has argued that the inscription may
be written this way for artistic reasons rather than to
suggest that the hieroglyph was actually read in that
sequence.61 The pronominal formula is thus treated
as an independent artistic element all its own in this
inscription, and it was movable to increase its interplay
with other elements of the temple’s iconography. To
some degree, the formula itself was treated as an im-
age in this case.
Apart from the Aleppine example, the formula ap-
pears in some early inscriptions from Malatya that may
date to the 11th century BC. The dynasty that con-
trolled Malatya during this period traced its lineage
through Kuzi-Teššub, the king of Carchemish and
descendent of the Hittite emperor Šuppiluliuma I.62
Much as the kings of Carchemish were beginning to
claim elements of Hittite royal inscriptions for them-
selves during this period—most notably the title
“Great King”—the rulers of Malatya appear to have
been emulating the Hittite Empire as well.63 The Late
Hittite adoption of the “I (am)” formula was one ele-
ment of imperial rhetoric they emulated. Perhaps the
most remarkable example of this is SPEKÇÜR, which
opens as follows:
§1 (EGO?) a+ra/i-nú-wa/i-ta-sa5 REX ku-zi-
TONITRUS HEROS || (INFANS)ha-ma-si-sa5 PUGNUS-
mi-li INFANS-mu-wa/i-za?|| ⸢MAx-LIx-zi(URBS)⸣
§1 I am Arnuwantis the King, grandson
of Kuzi-Teššub the Hero, son of PUGNUS-mili,
Country-Lord of the city of Malizi.
This inscription is on a stele that includes an image of
the king in question. The image of the king appears to
60 Ibid., 40.
61 Payne, “The Hieroglyphic Sign EGO(2),” 291–92.
62 Hawkins, “Corpus of Hieroglyphic Inscriptions, I.,” 283–84;
Federico Giusfredi, Sources for a Socio-Economic History of the Neo-
Hittite States, Texte Der Hethiter 28 (Heidelberg, 2010), 37–41;
Trevor R. Bryce, The World of the Neo-Hittite Kingdoms: A Political
and Military History (Oxford, 2012), 83–85, 98–101.
63 Hawkins, “Corpus of Hieroglyphic Inscriptions, I.,” 283;
Giusfredi, Sources for a Socio-Economic History of the Neo-Hittite
States, 43, 82–88; Bryce, World of the Neo-Hittite Kingdoms, 83–87,
be modeled on the hieroglyph EGO, and Payne argues
that it should be read as an instance of EGO2—the ver-
sion of the hieroglyph in which the sign has been ex-
tended into a full portrait, also called the amu-ﬁgure.64
However, the stele also contains an image of the king’s
wife along with an inscription for her. That inscription
notably lacks the fronted ﬁrst-person pronoun.65 This
suggests that while an image may have augmented the
“I (am)” formula, the two were not a formalized pair,
and could function together or separately.
The 10th Century
The most signiﬁcant examples of the formula from
the 10th century are arguably those of Carchemish,
which David Hawkins calls “the chief heir of the Hit-
tite Empire.”66 With eighteen of the seventy total
occurrences of the formula in Hieroglyphic Luwian,
Carchemish attests the formula more than any other
site. Furthermore, the 10th century attestations from
Carchemish demonstrate the great variety of instances
in which the “I (am)” formula could be employed.
During this period, the formula was standardized to
“I (am),” expressed by the hieroglyph EGO followed
by the implied speaker’s name, titles, and genealo-
gy.67 One of the earliest examples from Carchemish is
KARKAMIŠ A14b, which opens as follows:
§1 [E]GO-mi á-sa-tú-[wa/i-lá]-ma-za-sa
[. . .?] || [k]ar-ka-mi-sà-zi+a-sa(REGIO) |REGIO
DOMINUS -i+a-sa Isu-hi-si |IUDEX-ní-sa || |(INFANS)
§1 I (am) Astuwalamanzas [. . .] Karkamišean
Country-Lord, the ruler Suhis’s son.68
64 Payne, “The Hieroglyphic Sign EGO(2),” 289; Hawkins,
“Corpus of Hieroglyphic Inscriptions, I.,” 115. On the signiﬁcance
of 10th century Carchemish, see also Giusfredi, Sources for a Socio-
Economic History of the Neo-Hittite States, 44–52.
65 Hawkins, “Corpus of Hieroglyphic Inscriptions, I.,” 301–303.
66 Ibid., 20.
67 Payne, “The Hieroglyphic Sign EGO(2),” 285.
68 Unless otherwise noted, transcriptions and translations of
Iron Age Hieroglyphic Luwian inscriptions are those of Hawkins
but updated with the new readings proposed by Rieken and
Yakubovich. Hawkins, “Corpus of Hieroglyphic Inscriptions, I.”;
E. Rieken and I. Yakubovich, “The New Values of Luwian Signs L
319 and L 172,” in Luwian and Hittite Studies Presented to J. David
Hawkins on the Occasion of His 70th Birthday, ed. Itamar Singer,
(Tel Aviv, 2010), 199–219.
330 ) Journal of Near Eastern Studies
This inscription noticeably lacks any image of the
agent and is carved on a portal lion.69 The formula
alone manifests the agent in the context of another
image without any iconic ampliﬁcation.
KARKAMIŠ A14b was apparently not the norm, how-
ever, but rather one of many variations on the for-
mula’s context. To demonstrate this, I will provide
two further early examples from Carchemish. KARKAMIŠ
A1b opens as follows:
§1 EGO-mi-i IBONUS-ti-sa Isu-hi-si-i REGIO-ní
DOMINUS-ia-i-sa |BONUS-mi-sa || FEMINA-ti-i-sa
§1 I (am) BONUS-tis the Country-Lord Suhis’s
In this case, the inscription was worked on an ortho-
stat along with an image of the agent and a goddess,
who receives no inscriptional reference. The posture
of the agent’s ﬁgure notably resembles that of the EGO
hieroglyph.70 It is highly likely, then, that the image
and the “I (am)” formula are working together here.
Such a relationship between text and ﬁgure was not
absolutely necessary, however. This can be demon-
strated by KELEKL. That inscription opens:
§1 EGO-mi-i Isu-hi-sa-´ IUDEX[. . . kar-ka-]
mi?-si-sa(URBS) REGIO.DOMINUS-ia-ix-sa || Iá-sa-
tú-wa/i-lá-ma-za-[. . . . . .]-ix-sa
§1 I (am) Suhis the Ruler [. . . Karka]miš
Country-Lord, Astuwalamanza’s [. . . so]n(?).
This inscription is on a stele accompanied by an ap-
parent image of Suhis alongside the storm-god. The
image of Suhis is not aligned with the EGO hieroglyph
in any way in this case, so there is not a clear relation-
ship between text and image. These three examples all
date to within a generation of one another, and they
suggest that there was a degree of variation as to the
alignment of the “I (am)” formula with an associated
It is also during the 10th century that the “I (am)”
formula reappeared in Akkadian, apparently for the
ﬁrst time since the 15th century. One of these—the
Kapar(r)a71 inscription from Tell Halaf—shows strik-
ing consistency with the traditions of Syro-Anatolia.
69 Hawkins, “Corpus of Hieroglyphic Inscriptions, I.,” 83–85.
70 Ibid., 91–92.
71 On the normalization and etymology of this name, see most
recently A. A. Dornauer, “Die Geschichte von Guzana Im Lichte
Der Schriftlichen Zeugnisse,” in Tell Halaf: Im Krieg Zerstörte
Denkmäler Und Ihre Restaurierung, ed. Nadja Cholidis and Lutz
Martin (Berlin, 2010), 50 n. 99; K. Lawson Younger Jr., A Politi-
The contents of this building inscription were actu-
ally copied onto a number of dierent objects in the
palace of king Kapar(r)a, one of the early rulers of the
Aramaean kingdom of Gūzāna in the Ḫābūr region.
As a building inscription, most instances of the text
open, as might be expected, with the phrase ēkalli(lì)
mKapara apil mḪadiāni, “the palace of Kapar(r)a, son
of Ḫadiānu.”72 Only one instance opens instead with
anāku mKapara apil mḪadiāni, “I am Kapar(r)a, son
of Ḫadiānu.”73 This example is inscribed on a caryatid
at the entrance to the palace, the central of three such
caryatids. The ﬁgure depicted on this central caryatid
has been identiﬁed by Alessandra Gilibert as the ﬁgure
of Kapar(r)a himself, resulting in a unique relation-
ship between text and statue that matches some in-
stances of the Syro-Anatolian monumental discourse
discussed above.74 Only the statue of Kapar(r)a,
which arguably conjured the agent in the minds of
its viewers, was equipped with the “I (am)” formula
designed to manifest him by a dierent means. The
fact that no other inscription in the palace begins
with the “I (am)” formula speaks to the strength of
this connection, as well as to the fact that the formula
and the statue were intended to accomplish the same
The Kapar(r)a inscription is not the only 10th cen-
tury example of the “I (am)” formula in Akkadian,
however, nor even the only example connected to the
Ḫābūr region. The formula also appears in a royal-
type building inscription of the Assyrian governor
Bēl-ēriš.75 It opens:
cal History of the Arameans: From Their Origins to the End of Their
Polities (Atlanta, 2016), 247 n. 84.
72 For editions of these texts, see Bruno Meissner, “Die
Keilschriftentexte auf den steinernen Orthostaten und Statuen aus
dem Tell Halaf,” in Auf fünf Jahrtausenden morgenländischer Kul-
tur: Festschrift Max Feiherr von Oppenheim zum 70. Geburtstag ge-
widmet von Freunden und Mitarbeiten, ed. E. F. Weidner, Archiv für
Orientforschung 1 (Berlin, 1933), 72–76. The normalizations used
here are adapted from William F. Albright, “The Date of the Kapara
Period at Gozan (Tell Halaf),” Anatolian Studies 6 (1956): 81–82.
73 Meissner, “Die Keilschriftentexte auf den steinernen Ortho-
staten und Statuen aus dem Tell Halaf,” 77–79.
74 Alessandra Gilibert, “Death, Amusement and the City: Civic
Spectacles and the Theatre Palace of Kapara, King of Gūzāna,”
KASKAL: Rivista Di Storia, Ambienti E Culture Del Vicino Oriente
Antico 10 (2013): 51.
75 On the erection of royal-type inscriptions by non-royal Assyr-
ian elites more generally, see Shana Zaia, “How To (Not) Be King:
Negotiating the Limits of Power within the Assyrian Hierarchy,”
JNES 77/2 (2018): 207–208.
Fronted First-Person Pronoun in Syro-Anatolian Monumental Discourse ) 331
(1)a-na-ku mdEN-APIN LÚ.ŠID ⸢ša⸣ [. . .]
(1)I (am) Bēl-ēriš, vice-regent of [. . .]76
Like the Kapar(r)a inscription, Bēl-ēriš’s inscription is
a building inscription, in this case for a temple. The
deployment of this example could not be more dier-
ent from that of Kapar(r)a, however. It was inscribed
on a cylinder and is an example of a foundation deposit
inscription. Such foundation deposits were meant to
be hidden within the foundations of the buildings they
described, and they were only to be read by future rul-
ers undertaking renovations.77 These sorts of objects
derived a great deal of signiﬁcance from the fact that
they were hidden and not intended to be viewed. Some
were even produced to be immediately buried.78 Thus,
there is no image to be viewed in concert with the “I
(am)” formula in this case, and the object was not even
intended to be viewed in the ﬁrst place. The agent
would thus be manifested by the text alone on the rare
occasions it might be read. In this case, the “I (am)”
formula was combined with elements of the traditional
monumental discourse of Mesopotamia. This was only
the ﬁrst example of such diusion of the formula.
The 9th Century
At 9th century Carchemish, we ﬁnd one of the most
interesting developments of the “I (am)” formula.
The hieroglyph EGO was occasionally extended into a
full portrait of the agent, in which case it is typically
designated EGO2. This alternate form of the hiero-
glyph was thus able to totally replace the accompany-
ing ﬁgures modeled on EGO, such as those attested
at Aleppo, Malatya, and Carchemish itself. While in-
scriptions such as KARKAMIŠ A1b (in which the image
of the agent mirrors the EGO hieroglyph), or SPEKÇÜR
and DARENDE (in which glyptic ﬁgures appear to re-
76 This inscription is labeled Aššur-rēša-iši II 2001, and an edi-
tion may be found in A. Kirk Grayson, Assyrian Rulers of the Early
First Millenium BC I (1114–859 BC), The Royal Inscriptions of
Mesopotamia—Assyrian Periods 2 (Toronto, 1991), 126–28.
77 This purpose is made explicit in lines 24–26 of this inscription
(ibid.), which may be translated: “When this temple becomes old
[and dilapidated . . .], may a later vice-regent restore its weakened
(portions). May he anoint with oil [my foundation-deposit.]” This
translation is largely adapted from Grayson, but I have chosen to
translate the term tam-me-ni-ia (l. 26), “my foundation deposit,”
rather than “my clay inscription.”
78 Gerdien Jonker, The Topography of Remembrance: The Dead,
Tradition and Collective Memory in Mesopotamia (Leiden, 1995),
84–92; Zainab Bahrani, The Inﬁnite Image: Art, Time and the Aes-
thetic Dimension in Antiquity (London, 2014), 89–92.
place the EGO hieroglyph), undoubtedly set the stage
for this development, it is ﬁrst clearly in evidence at
the end of the 10th century or the beginning of the
9th.79 One of the earliest Karkamišean examples is in
KARKAMIŠ A13d, which opens:
§1 EGO2-wa/i-mi-i Ika-tú-wa/i-sa5 |kar-ka-
mi-si-za-sa(URBS) |REGIO DOMINUS-s[a . . .]
§1 I (am) Katuwas, the Karkamišean Country-
Lord, [. . .]
For this example, as might be expected, there are no
accompanying images other than the text itself. It is
rendered on an orthostat without an associated statue,
and the only image of the king is EGO2. The text here
serves an iconic function in addition to providing se-
mantic information.80 The instances of EGO2 are thus
signiﬁcant examples of the “I (am)” formula doing
double duty as text and image.
Approximately one-sixth of all the hieroglyphic oc-
currences of EGO have been extended into full portraits
of the agent. Apart from Carchemish, the practice is in
evidence at Tabal, Maraş, and possibly Amuq, Malatya,
and Cilicia. There are also cases of EGO2 being adapted
as an iconographic element without being treated as
writing. As in Carchemish, the examples of EGO2 at
Tabal, Malatya, and Amuq were on stelae, and thus
the hieroglyph alone served as the monument’s iconic
reference to the agent.81 Similarly, the potential exam-
ple of EGO2 from Cilicia has the hieroglyph inscribed
on a portal lion without a preserved separate image
of the agent.82 EGO2 did not necessarily replace other
images of the agent, however. This was patently not
the case at Maraş, where the use of EGO2 is curiously
limited to statues, suggesting a connection between
the text and its medium that was not made in such a
way at other sites.83 This interplay suggests that EGO2
at Maraş should be understood as an iconic reference
to the monument itself within the text.84
A similar iconic development of the EGO hieroglyph
is in evidence at Hama. Most of the examples of the
79 Payne, “The Hieroglyphic Sign EGO(2),” 289.
80 Ibid., 293–94.
81 For examples from Tabal, see ANDAVAL and BOR in Hawkins,
“Corpus of Hieroglyphic Inscriptions,” 514–19. For a potential ex-
ample from Amuq, see JISR EL HADID, ibid., 379. For this treatment
of the examples from Malatya, see Payne on SPEKÇÜR and DARENDE
in “The Hieroglyphic Sign EGO(2),” 289.
82 Ho. 1, KARATEPE 1: ibid., 290, n. 28.
83 See MARAŞ 1, 4, 13, 14.
84 Payne, “The Hieroglyphic Sign EGO(2),” 290–91.
332 ) Journal of Near Eastern Studies
“I (am)” formula come from the 9th century inscrip-
tions of Urahilina and his son Uratamis. While the
rulers of Hama did not adapt the practice of using
EGO2 as a portrait, Uratamis did still personalize the
hieroglyph to a degree. The ﬁgure represented by EGO
typically has its hair curling outward, and this standard
was maintained by Urahilina. Most of Uratamis’ in-
scriptions, however, show the hair curling inwards.85
Furthermore, the inscriptions of Uratamis only use the
logogram EGO for the “I (am)” formula, whereas all
other occurrences of the ﬁrst-person pronoun in his
inscriptions are rendered syllabically as á-mu.86 The
limited use of EGO in Uratamis’ inscriptions as well
as the personalization of the hairstyle suggest that
the logographic pronoun was meant to be speciﬁcally
identiﬁed with that ruler.87
Zincirli attests uniquely complex examples of re-
lating the “I (am)” formula to imagery. Though two
Hieroglyphic Luwian monumental texts have been
connected to the site, the “I (am)” formula is only
attested in inscriptions in Northwest Semitic dialects
there.88 Curiously, all of these inscriptions are carved
in raised relief, demonstrating a clear link to the hi-
eroglyphic scribal technique at Carchemish.89 Further-
more, the sculpture at Zincirli is clearly modeled on
Karkamišean style, though it was likely produced by a
85 The inward curl is evident on HAMA 1, 3, 6, and 7. Ibid., 287.
86 Ibid., 287.
87 An interesting comparative to this situation may be observed
in Egypt. The Hyksos kings modiﬁed the classiﬁer for the hieroglyph
ḥḳꜢ, “ruler,” in their inscriptions in order to mark their Asiatic identity,
most notably by changing the ﬁgure’s hairstyle. Danielle Candelora,
“Deﬁning the Hyksos: A Reevaluation of the Title ḥḳꜢ ḫꜢswt and Its
Implications for Hyksos Identity,” JARCE 53 (2017): 213.
88 On the Hieroglyphic Luwian inscriptions, see Younger Jr.,
Political History of the Arameans, 387; Virginia R. Herrmann, Theo
van den Hout, and Ahmet Beyazlar, “A New Hieroglyphic Luwian
Inscription from Pancarlı Höyük: Language and Power in Early
Iron Age Samʾal-YʾDY,” JNES 75/1 (April, 2016): 53–70, https://
doi.org/10.1086/684835. For editions of the Semitic inscriptions,
see Josef Tropper, Die Inschriften von Zincirli, Abhandlungen Zur
Literatur Alt-Syrien-Palästinas 6 (Münster, 1993); Dennis Pardee,
“A New Aramaic Inscription from Zincirli,” BASOR 356 (2009):
89 M. W. Hamilton, “The Past as Destiny: Historical Visions in
Samʾal and Judah under Assyrian Hegemony,” The Harvard Theo-
logical Review 91/3 (1998): 222; E. J. Struble and V. Rimmer Her-
rmann, “An Eternal Feast at Samʾal: The New Iron Age Mortuary
Stele from Zincirli in Context,” BASOR 356 (2009): 20; Alessandra
Gilibert, Syro-Hittite Monumental Art and the Archaeology of Per-
formance: The Stone Reliefs at Carchemish and Zincirli in the Earlier
First Millennium BCE (Berlin, 2011), 82.
separate workshop.90 Finally, a number of elements of
the monumental rhetoric of the texts at Zincirli appear
to be derived from Hieroglyphic Luwian phraseol-
ogy—especially that of Carchemish.91 Most notably,
nearly all of the monumental inscriptions at Zincirli
open with the “I (am)” formula.
The most signiﬁcant example from Zincirli is also
one of the earliest in a Northwest Semitic dialect. The
inscription dates to approximately 830 BC, potentially
as little as one or two generations after the last monu-
mental inscription in Hieroglyphic Luwian from the
site, KARABURÇLU 2.92 KAI 24, the Kulamuwa inscrip-
tion, opens in Phoenician as follows:
(1)ʾnk . klmw . br . ḥy[’]
(1)I am Kulamuwa, son of Ḥaya.93
This inscription not only opens with the “I (am)”
formula, its iconography also includes a full-length
portrait of the king running alongside the inscrip-
tion in clear imitation of Karkamišean attestations
of EGO2.94 The same innovation is attested a century
later in the Aramaic inscriptions of Bar-Rakib (KAI
216–218), though in that case the artisan was likely
imitating the Kulamuwa inscription.95 In contrast,
the other attestations of the “I (am)” formula from
Zincirli, in the Panamuwa and Katumuwa96 inscrip-
90 Ibid., 122.
91 E. Masson, “La Stèle Mortuaire de Kuttamuwa (Zincirli):
Comment L’appréhender,” Semitica et Classica 3 (2010): 53;
H. Craig Melchert, “Remarks on the Kuttamuwa Stele,” Kubaba 1
(2010): 3–11; I. Yakubovich, “West Semitic God El in Anatolian
Hieroglyphic Transmission,” in Pax Hethitica: Studies on the Hittites
and Their Neighbours in Honour of Itamar Singer, ed. Y. Cohen,
A. Gilan, and J. L. Miller (Wiesbaden, 2010), 396; I. Yakubovich,
review of Investigationes Anatolicae: Gedenkschrift Für Erich Neu,
ed. J. Klinger, E. Rieken, and C. Rüster, and Studia Anatolica in
Memoriam Erich Neu Dicata, ed. R. Lebrun and J. De Vos, Kratylos
56 (2011): 181; Hogue, “Abracadabra or I Create as I Speak.”
92 See Younger Jr., Political History of the Arameans, 383–87, for
the dating of Kulamuwa. KARABURÇLU 2 is dated on stylistic grounds
to between 925–850 BC: D. Bonatz, Das Syrohethitische Grabdenk-
mal: Unterschungen Zur Entstehung Einer Neuen Bildgattung in Der
Eisenzeit Im Nordsyrischsüdostanatolischen Raum (Mainz, 2000), 19.
93 Transcriptions of Northwest Semitic inscriptions follow KAI,
but the translations are my own. Donner and Röllig, Kanaanäische
und aramäische Inschriften; see also Tropper, Die Inschriften von
94 Gilibert, Syro-Hittite Monumental Art and the Archaeology of
95 Ibid., 87–88.
96 On the vocalization of KTMW as Katumuwa, see K. Lawson
Younger Jr., “Two Epigraphic Notes on the New Katumuwa In-
scription from Zincirli,” MAARAV 16/2 (2009): 159–66.
Fronted First-Person Pronoun in Syro-Anatolian Monumental Discourse ) 333
tions, are accompanied by a statue and a relief image
respectively.97 Though the scribes writing in a Semitic
alphabetic tradition could not combine the so-called
amu-ﬁgure and the ﬁrst-person pronoun, they nev-
ertheless appear to have attempted to emulate the
full range of expression aorded by the Karkamišean
versions of EGO and EGO2. Zincirli provides arguably
the clearest example of Syro-Anatolian monumental
discourse being iterated in new languages as more
than just the semantic dimension of the formula was
carried over. This suggests a clear attempt to align the
Zincirlian tradition with the Karkamišean one, but it
also illustrates the divergence aorded by rendering
the formula in new script tradition. While a ﬁgure
resembling the hieroglyph EGO2 has been preserved,
it is necessarily divorced from the “I (am)” formula.
These examples represent an interaction between text
and image, but not a combination such as could be
accomplished by Anatolian hieroglyphs.
Another split adaptation of Syro-Anatolian mon-
ument-making traditions may be observed in Moab
in the 9th century BC. The formula is attested in two
inscriptions, one from Dibon (KAI 181) and one from
Kerak (KAI 306). These inscriptions also serve as the
southernmost examples of the formula’s use in the
Levant. The two may be copies of each other, and
both open as follows:
(1–2a)ʾnk . mšʿ . bn . kmšyt . mlk . mʾb . hdbny
(1–2a)I am Mesha, son of Chemoshyat, king of
Moab, the Dibonite.98
The Kerak Inscription was inscribed on a free-stand-
ing statue of the agent, similar to the examples from
Maraş.99 Furthermore, a pair of portal lions was dis-
covered at Kerak, demonstrating Moab’s participation
in the artistic and intellectual koiné that characterized
the Syro-Anatolian states of the Iron Age and their
neighbors.100 The Mesha Inscription from Dibon is
perhaps more interesting. The inscription is carved on
a stele with no associated iconography whatsoever. In
this case, the inscription alone adorns the monument;
97 Tropper, Die Inschriften von Zincirli, 54–55; Gilibert, Syro-
Hittite Monumental Art and the Archaeology of Performance, 95.
98 This transcription is from KAI 181; KAI 306 ﬁts this infor-
mation onto one line. The translation is my own.
99 Martin Weber, “Two (?) Lion Reliefs from Iron Age Moab:
Further Evidence for an Architectural and Intellectual Koiné in
the Levant?” BASOR 377 (2017): 98, https://doi.org/10.5615/
100 Ibid.: 100.
if the “I (am)” formula has an iconic dimension here,
it is as text.101 In this case, the formula is perhaps aug-
mented by the southern Levantine tradition of erect-
ing aniconic standing stones in order to produce an
The 8th Century
The use of the text by itself for an iconic function is
also suggested by the 8th century Luwian-Phoenician
bilingual inscriptions from Cilicia, which exhibit per-
haps the most complex relationship between text and
image. ÇINEKÖY is inscribed on a statue of a bull with
a deity standing on top.103 The Luwian text is unusu-
ally incised in cursive script, rather than the expected
raised relief script of earlier Hieroglyphic Luwian
monuments.104 Though a statue accompanies the text,
if it really is a statue of a diety it may not depict the
implied speaker at all. The image and the “I (am)”
formula thus seem disconnected. Additionally, the
Luwian version of the inscription is spread around
the monument in an almost haphazard manner, while
the Phoenician version is centered at the front of the
KARATEPE 1 provides an even more extreme exam-
ple of the divorce of text and image in Cilicia. Like
ÇINEKÖY, the inscription is incised rather than carved
in relief. The inscription appears in ﬁve copies—two
in Hieroglyphic Luwian, and three in Phoenician—
four of which appear on orthostats and portal beasts
with no associated images of the agent.106 One of the
Phoenician copies—designated “C,” and somewhat
101 Nathaniel B. Levtow, “Monumental Inscriptions and the Rit-
ual Representation of War,” in Warfare, Ritual and Symbol in Bibli-
cal and Modern Contexts, ed. Brad E. Kelle, Frank Ritchel Ames,
and Jacob L. Wright, Ancient Israel and Its Literature 18 (Atlanta,
102 Elizabeth Bloch-Smith, “Massebot Standing for Yhwh: The
Fall of a Yhwistic Cult Symbol,” in Worship, Women, and War: Es-
says in Honor of Susan Niditch, ed. John J. Collins, T. M. Lemos,
and Saul M. Olyan, Brown Judaic Studies 357 (Providence, RI,
103 This statue may be an example of a depiction of a king as the
storm-god. Guy Bunnens, Tell Ahmar II: A New Luwian Stele and the
Cult of the Storm-God at Til Barsip-Masuwari (Leuven, 2006), 128.
104 Recai Tekoglu et al., “La bilingue royale louvito-phénicienne
de Çineköy,” Comptes rendus des séances de l’Académie des Inscrip-
tions et Belles-Lettres 144/3 (2000): 968, https://doi.org/10.3406/
105 Ilya Yakubovich, “Phoenician and Luwian in Early Iron Age
Cilicia,” Anatolian Studies 65 (2015): 48.
106 Hawkins, “Corpus of Hieroglyphic Inscriptions,” 45.
334 ) Journal of Near Eastern Studies
divergent from the other inscriptions—was inscribed
on a statue of a god.107 While there is some disagree-
ment about whether the Phoenician version is the
original or a translation, what is agreed upon is that
the Phoenician receives some form of pride of place.108
This is illustrated especially by its prominent, centered,
and clear display as opposed to the chaotic distribu-
tion of the Hieroglyphic Luwian inscriptions. Perhaps
because of this emphasis on the Phoenician text, the “I
(am)” formula was not clearly associated with imagery
in Cilicia. The texts opening with the formula (at least
the Semitic versions) were nevertheless prominently
displayed as key parts of the monumental installations.
The texts do not repeat any of the iconic motifs of
the accompanying reliefs, but rather seem included as
images themselves, albeit ones that were potentially
decipherable as texts. The display of the formula in
Phoenician script thus had an iconic dimension per-
haps greater than that of the accompanying Luwian
text, which could have better aorded an additional
connection of the formula to a separate image. As
for the Luwian version itself, its function appears to
have been purely iconic, given its apparently haphaz-
ard deployment, and it may have been included not
to be read, but rather to visually legitimate the other
elements of the monumental installation.109
The divorce of the “I (am)” formula from imag-
ery is also exempliﬁed by later Hieroglyphic Luwian
examples, which began to transition to a phonetic
rendering of the ﬁrst-person pronoun rather than a
logographic one. This practice is attested in PALANGA
from Malatya, as well as in BULGARMADEN and HSARCIK
1 from Tabal, all of which date to the late 8th century.
HSARCIK 1 is inscribed on a boulder with no associated
imagery.110 BULGARMADEN is a rock inscription with no
apparent link to any iconography.111 PALANGA alone was
carved on a statue, but the ﬁgure is broken, and there
is no clear reference to it in the inscription.112 How-
107 Payne, Iron Age Hieroglyphic Luwian Inscriptions, 20.
108 Annick Payne, “Multilingual Inscriptions and Their Audi-
ences: Cilicia and Lycia,” in Margins of Writing, Origins of Cultures,
ed. Seth Sanders, OIS 2 (Chicago, 2006), 123–27; Yakubovich,
“Phoenician and Luwian in Early Iron Age Cilicia,” 44–48.
109 Aslı Özyar, “The Writing on the Wall: Reviewing Sculpture
and Inscription on the Gates of the Iron Age Citadel of Azatiwataya
(Karatepe-Aslantaş),” in Cities and Citadels in Turkey: From the Iron
Age to the Seljuks, ed. Scott Redford and Nina Ergin, Ancient Near
Eastern Studies, Supplement 40 (Leuven, 2013), 131–33.
110 Hawkins, “Corpus of Hieroglyphic Inscriptions,” 483.
111 Ibid., 521.
112 Ibid., 325–26.
ever, the rendering of the syllabogram á is somewhat
reminiscent of EGO, and there may be an intentional
link between the two in that inscription.113 The incon-
sistency of employing images with these examples as
well as the phonetic writing of the “I (am)” formula
suggests that the formula was functioning separate
from other icons in these cases.
It is during the 8th century that we also ﬁnd per-
haps the most unusual semantic developments of the
“I (am)” formula in Hieroglyphic Luwian in an in-
scription from Tabal. KULULU 4 opens as follows:
§1 EGO-wa/i-mi ru-wa/i-sa4 IUDEX-ní-sa
§ I was Ruwas the Ruler, the sun-blessed
Most notably, this inscription sees the verb á-sá-ha, “I
was,” inserted in the midst of the agent’s titulary. Two
concepts can be illustrated by this shift in the standard
formula. First, the past-tense formula makes explicit
that the speaker in the inscription is the deceased per-
son to whom the monument is dedicated. The agent
and the monument commissioner are clearly not the
same individual.115 Secondly, the fact that the “I (am)”
formula has been clearly complemented with the ver-
bal phrase á-sá-ha, “I was,” suggests that the formula
was here intended on a purely semantic level. There is
no iconic component to amplify the use of the formula
in the inscription. As for the monument containing
the inscription, it is a stele with no iconography or
associated statuary whatsoever.
The split between text and image is also exempli-
ﬁed by the 8th century Akkadian adaptations of the
“I (am)” formula. This is perhaps most notable at Til
Barsip, which had an established tradition of using
the formula in Hieroglyphic Luwian, perhaps due to
the inﬂuence of the practice at neighboring Carchem-
ish.116 In fact, the formula was so standardized that
it may have continued to be used even after the As-
113 Payne, “The Hieroglyphic Sign EGO(2),” 284 n. 15.
114 This transcription and translation are based on Hawkins,
“Corpus of Hieroglyphic Inscriptions,” 445.
115 In fact, the commissioner of the monument is identiﬁed sepa-
rately in §15 of the same inscription as Hulis, Ruwas’ nephew. Ibid.,
446; Payne, Iron Age Hieroglyphic Luwian Inscriptions, 50–51.
116 Some of the monumental artwork from the site so closely re-
sembles that of Carchemish that it has previously been proposed that
both were produced by the same craftsmen. D. Ussishkin, “On the
Dating of Some Groups of Reliefs from Carchemish and Til Barsip,”
Anatolian Studies 17 (1967): 190–92; Irene Winter, “Carchemish
ša kišad puratti,” Anatolian Studies 33 (1983): 181–82; Bunnens,
Fronted First-Person Pronoun in Syro-Anatolian Monumental Discourse ) 335
syrian conquest of the site. The governor of the site,
Ninurta-bēlu-uṣur, set up a trilingual inscription in
Akkadian, Aramaic, and Hieroglyphic Luwian at the
nearby city of Ḫadattu (modern Arslan Tash). The “I
(am)” formula has been reconstructed for the Akka-
dian version of the text, having possibly been calqued
back into the language from which it originated. The
inscription’s opening closely follows Syro-Anatolian
models, though it lacks a genealogy.117 It opens as
(1)[a-na-ku? mdMAŠ-EN]-⸢PAP⸣ EN.NAM URU.
(1)I am Ninurta-bēlu-uṣur, governor of
Though the openings of the Aramaic and Hiero-
glyphic Luwian versions of the inscription are poorly
preserved, they likely began with the same formula.119
This trilingual was inscribed on a pair of portal lions
without any accompanying ﬁgure of the agent. With-
out a preserved form of the EGO hieroglyph, it cannot
deﬁnitely be determined to what degree the formula
may have been iconically personalized. If the formula
was present in the Akkadian and Aramaic inscriptions,
however, it would have been divorced from accompa-
nying imagery, and perhaps acted as an icon on its own.
The Akkadian adaptation of the formula at Til Bar-
sip may have set the stage for further adaptation of the
formula in Mesopotamia. Some of the earliest attesta-
tions of this are in the inscriptions of Ninurta-kudurrī-
Tell Ahmar II, 46–47; Gilibert, Syro-Hittite Monumental Art and
the Archaeology of Performance, 121–22.
117 In this, it resembles the Karkamišean inscriptions of Yariris,
who also did not give his genealogy. This may be due to the fact
that both rulers were technically governors or regents who were
legitimated by the relationship to another ruler rather than by their
own descent. Elif Denel, “Ceremony and Kingship at Carchemish,”
in Ancient Near Eastern Art in Context: Studies in Honor of Irene
J. Winter by Her Students, ed. Jack Cheng and Marian H. Feldman
(Leiden, 2007), 194–97.
118 The transcription follows that of RINAP 1, Tiglath-Pileser III
No. 2001 but the translation follows that of Röllig. Hayim Tadmor,
The Inscriptions of Tiglath-Pileser III King of Assyria (Jerusalem,
1994); Wolfgang Röllig, “Die Inschriften Des Ninurta-Bēlu-Uṣur,
Statthalters von Kār-Salmānu-Ašarēd. Teil I,” in Of God(s), Trees,
Kings, and Scholars: Neo-Assyrian and Related Studies in Honour
of Simo Parpola, ed. M. Luukko, S Svärd, and R. Mattila, Studia
Orientalia 106 (Helsinki, 2009), 265–78.
119 Röllig, “Die Inschriften Des Ninurta-Bēlu-Uṣur”; Guy Bun-
nens, “Assyrian Empire Building and Aramization of Culture as Seen
from Tell Ahmar/Til Barsip,” Syria 86 (2009): 78–79; Younger Jr.,
Political History of the Arameans, 362–65.
uṣur of Sūḫu.120 The appearance of the formula at
Sūḫu may be a native innovation, but it may also be
explained by the state’s interactions with the polities
of Syro-Anatolia, such as its trade relationship with
Hama.121 One example should serve to demonstrate
the nearness of the formula at Sūhu to that of Syro-
Anatolia. Ninurta-kudurrī-uṣur no. 1 opens:
(1)a-na-ku md MAŠ-NÍG.DU-ÙRU LÚ.GAR kursu-ḫi
u kurmá-rí DUMU mdUTU-SAG-PAP LÚ.GAR kursu-ḫi
(1)I am Ninurta-kudurrī-uṣur, governor of the
land of Sūḫu and the land of Mari, son of Šamaš-
rēša-uṣur, governor of the land of Sūḫu and the
land of Mari.122
This opening is strikingly similar to the model known
from Syro-Anatolia. Like the Akkadian example from
Til Barsip, the examples of the formula from Sūḫu
lack a clear connection to separate images. Of the in-
scriptions containing the formula, no. 9 is inscribed
on a stele and nos. 1, 2, and 14 are all recorded on
Mesopotamian interaction with the “I (am)” for-
mula may also be suggested by an 8th century inscrip-
tion from Hama. Perhaps as little as twenty years after
the reign of Uratamis, a king named Zakkur gained
control of the site after a brief period of domination
by Aram-Damascus.124 He erected a statue at Tel Aphis
with an inscription that opens as follows after a brief
(2)[ʾ]nh . zkr . mlk . ḥmt . wlʿš . ʾš . ʿnh . ʾnh
(2)I am Zakkur, king of Hamath and Luʿash.
A man of Ana(t) am I.125
120 For the relevant inscriptions from Sūḫu, see Ninurta-kudurrī-
uṣur nos. 1, 2, 9, and 14 in Grant Frame, Rulers of Babylonia. From
the Second Dynasty of Isin to the End of Assyrian Domination (1157–
612 BC), The Royal Inscriptions of Mesopotamia, Babylonian Pe-
riods 2 (Toronto, 1995).
121 Younger Jr., Political History of the Arameans, 474–75.
122 The transcription follows RIMB 2 (Frame, Rulers of Babylo-
nia, 291) but the translation is my own.
123 Ibid., 291–300, 315–17, 322; K. Lawson Younger Jr., “An-
other Look at the Nomadic Tribal Arameans in the Inscriptions of
Ninurta-Kudurrī-Uṣur of Suḫu,” in Marbeh Ḥokmah. Studies in the
Bible and the Ancient Near East in Loving Memory of Victor Avigdor
Hurowitz, ed. S. Yona et al. (Winona Lake, IN, 2015), 614–30.
124 Younger Jr., Political History of the Arameans, 474–76;
Stefania Mazzoni, “Tell Aﬁs in the Iron Age: The Temple on the
Acropolis,” Near Eastern Archaeology 77/1 (2014): 51, https://doi.
125 The transcription follows KAI 202, but the translation is my own.
336 ) Journal of Near Eastern Studies
Zakkur claimed to be from Ana(t), one of the chief
cities of Sūḫu and the location of all of the inscriptions
of Ninurta-kudurrī-uṣur.126 This may be one motiva-
tion for Zakkur’s use of the “I (am)” formula. The
other motivation for his adaptation of it is likely the
fact that Zakkur was restoring the kingdom of Hama.
He notably uses the titulary of the former kings of
Hama as attested in Hieroglyphic Luwian, only adding
Luʿash as part of a combined kingdom.127 The most
unusual feature of Zakkur’s inscription is his use of
Aramaic, but this may be explained by the fact that he
captured the territory of Hama from Aram-Damascus.
The paleography of the Zakkur inscription is so simi-
lar to that of the fragmentary inscriptions of Hazael
from the site as to suggest that the inscriptions were
produced by members of the same scribal school.128
The Aramaean scribal network at Tel Aphis may have
simply been absorbed by Zakkur upon his conquest
of the region. Whether the “I (am)” was also a carry-
over from the Aramaean administration is currently
impossible to conclude, but it was clearly aligned with
Zakkur’s claimed homeland of Ana(t), as well as the
region he had conquered.
As for the potential connection between the for-
mula and the image in the Zakkur inscription, the re-
lationship between the two may be the most complex
encountered thus far. The inscription is on the base of
a statue that has not been preserved. The inscription
claims only that the monument was erected lʾlwr, “for
Elwer,” the deity to whom the monument is dedicat-
ed.129 A similar monumentalization clause occurs in
the Panamuwa inscription from Zincirli, in which the
agent claims hqmt . nṣb. zn . lhdd, “I raised this monu-
ment for Hadad.”130 In the case of the Panamuwa
inscription, the statue has been identiﬁed as the deity
Hadad rather than the agent, Panamuwa.131 It is thus
126 Younger Jr., “Another Look at the Nomadic Tribal Arame-
127 See, for comparison, HAMA 1, an inscription of Uratamis. It
opens as follows: §1 EGO-mi MAGNUS+ra/i-tà-mi-sa u+ra/i-hi-li-ni-
sa |INFANS.NI-za-sa [i-ma-tú-wa/i-ni REGIO REX]||: §1 I am Uratamis,
son of Urahilina, king of Hama.
128 Maria Giulia Amadasi Guzzo, “Tell Aﬁs in the Iron Age: The
Aramaic Inscriptions,” Near Eastern Archaeology 77/1 (2014): 55;
Younger Jr., Political History of the Arameans, 475–76.
129 KAI 202:1 reads in full: [n]ṣbʾ . zy . šm . zkr . mlk [. ḥ]mt .
wlʿš . lʾlwr [. mrʾh] “The monument that Zakur, king of Hamath
and Luʿash, placed for Elwer, his lord.”
130 KAI 214:1. On my translation of nṣb as “monument” rather
than “statue,” see Hogue, “Abracadabra or I Create as I Speak,” 58.
131 Tropper, Die Inschriften von Zincirli, 54.
likely that the ﬁgure in the statue erected by Zakkur
was Elwer rather than Zakkur. Similarly, the statues
associated with ÇINEKÖY and KARATEPE 1 that were dis-
cussed above have been identiﬁed as representations
of the storm-god.132 It may thus have been a common
practice to inscribe the “I (am)” formula on divine
images in the 8th century, but only the Panamuwa
inscription gives a clue as to why.
The Panamuwa inscription speciﬁes that when sac-
riﬁces are oered before the statue and the name of
Hadad is invoked, the ritual participant is to say: [tʾ]
kl . nbš . pnmw . ʿmk . wtš[ty . n]bš . pnmw . ʿmk, “May
the ‘soul’133 of Panamuwa eat with you, and may the
‘soul’ of Panamuwa drink with you.” The inscription
then goes on to explain this ritual practice as follows:
ʿd . yzkr . nbš . pnmw . ʿm . [hd]d, “Henceforth, may
he invoke the ‘soul’ of Panamuwa with Hadad.”134
These ritual prescriptions clarify that the monument
was intended to manifest Panamuwa together with
Hadad; the statue acts as a manifestation of Hadad,
while the inscription and its prescribed ritual manifest
Panamuwa together with him.135 As a result, not only
are the agent and deity present with the monument’s
users, but more importantly the agent is present with
the deity. In this case, the formula played a role per-
haps not unlike that of votive statues and inscriptions
known especially from Mesopotamia. Placing some
material representation of the supplicant near a di-
vine image allowed them to enter a state of constant
ritual interaction.136 In this case, rather than placing
a votive inscription or statue near the divine statue,
the inscription—with its manifestational “I (am)”
formula—is placed directly on the divine statue. The
“I (am)” formula does not indicate the statue in cases
132 Bunnens, Tell Ahmar II, 128.
133 I follow Sanders in placing this dicult-to-translate term in
scare quotes as a placeholder: Seth Sanders, “The Appetites of the
Dead: West Semitic Linguistic and Ritual Aspects of the Katumuwa
Stele,” BASOR 369 (2013): 48–49 n. 62. For more on the nbš,
especially as it relates to ritual practice, see Sanders, “Naming the
Dead”; Matthew Suriano, “Breaking Bread with the Dead: Katu-
muwa’s Stele, Hosea 9:4, and the Early History of the Soul,” JAOS
134/3 (2014): 385–405; Struble and Herrmann, “An Eternal Feast
at Samʾal”; Virginia Rimmer Herrmann, “The KTMW Stele from
Zincirli: Syro-Hittite Mortuary Cult and Urban Social Networks,”
in Redeﬁning the Sacred: Religious Architecture and Text in the Near
East and Egypt 1000 BC–AD 300, ed. Elizabeth Frood and Rubina
Raja, Contextualizing the Sacred 1 (Turnhout, 2014), 153–81.
134 This translation follows Sanders, “Appetites of the Dead”:
135 Ibid.: 49.
136 Postgate, “Text and Figure in Ancient Mesopotamia,” 177.
Fronted First-Person Pronoun in Syro-Anatolian Monumental Discourse ) 337
like this, but rather allows the agent to be manifested
continuously within the presence of the deity. This is
a very sophisticated development of the formula, as it
simultaneously allows the agent to enter the presence
of the deity without need of a separate monument,
while also pointing even more eectively to the inti-
mate relationship between the two.
The 7th Century
The Mesopotamian iterations with the “I (am)”
formula likely set the stage for the Neo-Assyrian ad-
aptation of it, but this also may have been motivated
more generally by Assyrian interactions with Syro-
Anatolian artistic traditions. Assyrian monuments
adapted a number of elements from the monument-
making practices of Carchemish and other Syro-
Anatolian states.137 Curiously, the practice of framing
inscriptions with EGO2 at Carchemish was adapted by
the Assyrians as early as the reign of Assurnasirpal II
in the 9th century on his Banquet Stele, but the ﬁgure
of the ruler was interpreted as an autonomous image
with no semantic value.138 The “I (am)” formula itself,
however, was not adopted by the Assyrian kings until
the 7th century.
The “I (am)” formula ﬁrst appeared in the inscrip-
tions of Sennacherib, attested eight times. It appeared
a further ﬁve times in the inscriptions of his son Es-
arhaddon. It only appears to have become a standard
opening for Assyrian royal inscriptions, however, dur-
ing the reigns of Ashurbanipal and Sîn-šarra-iškun. It
is also preserved in one inscription of Aššur-etel-ilāni,
but this scarcity is most likely related to the short-
ness of his reign. The Assyrians primarily adapted the
practice of fronting the ﬁrst-person pronoun in their
version of the formula, but they retained their tradi-
tional titulary rather than matching the briefer titles
and longer genealogies of Syro-Anatolia. In Assyria,
the formula was used regularly on prisms and cylinders
137 Irene J. Winter, “Art as Evidence for Interaction: Relations
Between the Assyrian Empire and North Syria,” in Mesopotamien
Und Seine Nachbarn. Politische Und Kulturelle Wechselbeziegungen
Im Alten Vorderasien von 4. Bis 1. Jahrtausend v. Chr., ed. H. Kühne,
H. J. Nissen, and J. Renger, Berliner Beiträge Zum Vorderen Ori-
ent 1 (Berlin, 1982), 357; Sanna Aro, “The Origins of the Artistic
Interactions between the Assyrian Empire and North Syria Revis-
ited,” in Of God(s), Trees, Kings, and Scholars, ed. Luukko, Svärd,
and Mattila, 13.
138 Guy Bunnens, “From Carchemish to Nimrud Between Vi-
sual Writing and Textual Illustration,” Subartu XVI (2005): 22–23.
in addition to other monumental objects; there is no
iconic dimension to these examples other than the
monumental use of Cuneiform script.139 For example,
a prism inscription of Ashurbanipal opens:
(1–2)a-na-ku mAN.ŠÁR-DÙ-A LUGAL GAL LUGAL
dan-nu LUGAL ŠÚ LUGAL KUR AN.ŠÁR.KI LUGAL kib-
(1–2)I am Ashurbanipal, great king, strong
king, king of the world, king of Assyria, king of
the four quarters.140
Curiously, the formula was standardized in Assyria
proper. In contrast, Assyrian royal inscriptions set up
in Babylonia tended to maintain the southern Meso-
potamian tradition of the “. . . (am) I” formula, the
Akkadian precursor to the “I (am)” formula.141 The
successive Neo-Babylonian and Achaemenid empires
mostly maintained the Babylonian practice, and the
“I (am)” formula then slipped into obscurity, at least
as a part of royal rhetoric in Mesopotamia and Syro-
Anatolia.142 During this brief period of Neo-Assyrian
139 On this understanding of cuneiform, see especially Irene J.
Winter, “Text On/in Monuments: ‘Lapidary Style’ in the Ancient
Near East,” in Sign and Design: Script as Image in a Cross-Cultural
Perspective (300–1600 CE), ed. B. M. Bedos-Rezak and J. F. Ham-
burger (Washington, D.C., 2016), 197–218.
140 The transcription follows Ashurbanipal No. 3 in RINAP 5,
but the translation is my own. For other Neo-Assyrian examples of
the “I (am)” formula, see Sennacherib nos. 133–34, 177, 180–82,
184–85; Esarhaddon nos. 64, 74–75, 94–95; Ashurbanipal nos.
2–5, 7, 9, 10–11, 13, 19, 33, 36, 41, 44–45, 49, 52–56, 71, 73, 105,
112; Aššur-etel-ilāni no. 1; Sîn-šarra-iškun nos. 1, 6, 10–13, 19.
Erle Leichty, The Royal Inscriptions of Esarhaddon, King of Assyria
(680–669 BC), The Royal Inscriptions of the Neo-Assyrian Period
4 (Winona Lake, IN, 2011); James Novotny and Joshua Jeers, The
Royal Inscriptions of Ashurbanipal (668–631 BC), Aššur-Etel-Ilāni
(630–627 BC) and Sîn-Šarra-Iškun (626–612 BC), Kings of Assyria,
Part 1, The Royal Inscriptions of the Neo-Assyrian Period, 5/1
(Winona Lake, IN, forthcoming); James Novotny, Joshua Jeers,
and Grant Frame, The Royal Inscriptions of Ashurbanipal (668–631
BC), Aššur-Etel-Ilāni (630–627 BC) and Sîn-Šarra-Iškun (626–612
BC), Kings of Assyria, Part 2, The Royal Inscriptions of the Neo-
Assyrian Period, 5/2 (Winona Lake, IN, forthcoming).
141 For examples, see the Babylonian inscriptions of Ashurbani-
pal in RINAP 5, nos. 101–104, 111, and 113.
142 The “I (am)” formula was brieﬂy resurrected by Naboni-
dus during the Neo-Babylonian period, but even he maintained
the standard Babylonian practice within Babylonia proper. See Na-
bonidus 23, 49, 56, and 2003 in Frauke Weiershäuser and James
Novotny, The Royal Inscriptions of Amēl-Marduk (562–560 BC),
Neriglissar (560–556 BC), and Nabonidus (555–539 BC), Kings of
Babylon, Royal Inscriptions of the Neo-Babylonian Empire 3 (Wi-
nona Lake, IN, 2019). The Achaemenids also used an “I (am)”
formula in some of their Old Persian inscriptions set up in Iran.
338 ) Journal of Near Eastern Studies
history, however, royal inscriptions of the Assyrian
heartland adapted the formula of the neighboring re-
gions to the west, perhaps because of its clear connec-
tion to royal authority there, and its ability to evoke
the royal voice and image.
Attestation in the Hebrew Bible
A ﬁnal Iron Age example relevant to the complex rela-
tionship between the formula and images comes from
the Hebrew Bible. The phrase “I am Yahweh” became
an extremely productive part of biblical literature on
its own. The fullest version of this formulation appears
at the head of the Decalogue in Exodus 20:2, which
reads as follows:
ʾnky yhwh ʾlhyk ʾšr hwṣʾtyk mʾrṣ mṣrym mbyt
I am Yahweh, your god, who brought you out
from the land of Egypt, from a house of slaves.
Previous scholarship has connected this verse in par-
ticular to the monumental “I (am)” formula.143 This
is a particularly attractive place to ﬁnd the formula
because the Masoretic version of the book of Exodus
links the Decalogue to the erection and inscribing of
stelae.144 Given the Moabite adaptation of the formula
discussed above, it is less of a stretch to see the formula
in Israelite literature as an adaptation of Syro-Anato-
lian monumental practice. The editors of the Deca-
logue were attempting to pass it o as a monumental
text of Yahweh, so they framed it accordingly with
features such as the “I (am)” formula and an account
of stelae erection.145
The most signiﬁcant example of this is the Bisitun Inscription. A
small number of Phoenician inscriptions attest the formula during
the Persian and Hellenistic periods, but its deployment appears to
have changed substantially. See, for example, the Tabnit Sarcopha-
gus Inscription (KAI 13) and the Yehawmilk Inscription (KAI 10).
The question of whether or not these examples were related to the
earlier formula will be taken up in a future study.
143 Poebel, Das appositionell bestimmte Pronomen der 1. Pers.
Sing., 53–57; Cassuto, Commentary on the Book of Exodus, 76, 241;
Sarna, Exodus Commentary, 15, 109; V. Hurowitz, “What Can Go
Wrong with an Idol?,” in N. May, ed., Iconoclasm and Text Destruc-
tion in the Ancient Near East and Beyond (Chicago, 2012), 299;
Levtow, “Monumental Inscriptions and the Ritual Representation of
War,” 37–38; Demsky, “The Interface of Oral and Written Traditions
in Ancient Israel,” 21.
144 Dennis J. McCarthy, Treaty and Covenant, Analecta Biblica
21 (Rome, 1978), 174.
145 Timothy Hogue, “The Monumentality of the Sinaitic Deca-
logue: Reading Exodus 20 in Light of Northwest Semitic Mon-
It is now possible to answer the challenge posed
by Payne and “bridge the gap in time between the
Domuztepe stela (early 9th century BC) and the book
Exodus (6th century BC).”146 Hawkins previously
hinted at a link between DOMUZTEPE 2, where the name
of a deity may be represented by the EGO hieroglyph,
and Exodus 3:14, wherein Yahweh declares ʾhyh ʾšr
ʾhyh, “I am that I am.”147 In fact, there is probably
no relationship. The formulation in Exodus 3:14 is
phrased signiﬁcantly dierently from other Northwest
Semitic attestations of the formula, including the one
in Exodus 20:2.148 Nevertheless, the formula in the
Decalogue is at least related to the Syro-Anatolian
formula more broadly, if not to DOMUZTEPE 2 in par-
ticular. As for the gap in time, the Decalogue has
previously been connected to traditions from the 8th
century, even if the redaction of Exodus was later.149
This is further supported by its use of the “I (am)”
formula, which was most widespread in the 10th–8th
The biblical adaptation of the “I (am)” formula
also points to a signiﬁcant aspect of its underlying
function. It is curious that a formula previously so
closely connected with iconography was inserted into
a text that shortly follows it with an apparent ban on
images (Ex. 20:4) that gave rise to idolatry polemics
in later interpretive traditions.150 In fact, this ban is
probably only on unauthorized images of Yahweh.
ument-Making Practices,” Journal of Biblical Literature 138/1
146 Payne, “The Hieroglyphic Sign EGO(2),” 294 n. 36.
147 Hawkins, “Corpus of Hieroglyphic Inscriptions,” 71.
148 The scholarship on Ex. 3:14 is vast, but for a recent summary
of the debate, see William M. Schniedewind, “Calling God Names:
An Inner-Biblical Approach to the Tetragrammaton,” in Scriptural
Exegesis: The Shapes of Culture and the Religious Imagination: Essays
in Honour of Michael Fishbane, ed. Deborah A. Green and Laura S.
Lieber (Oxford, 2009), 79–84.
149 William Johnstone, Exodus (Sheeld, UK, 1990), 94. Of
course, it must be admitted that the date of the Decalogue is a very
contentious issue within biblical studies, and that its transmission
was a complex one dependent upon more inﬂuences than just the
monumental rhetoric it imitated. For a review of various studies of
the text’s “biography” and a new proposed date of composition in
the 7th century, see Erhard Blum, “The Decalogue and the Compo-
sition History of the Pentateuch,” in The Pentateuch: International
Perspectives on Current Research, ed. Thomas B. Dozeman, Konrad
Schmid, and Baruch J. Schwartz (Tübingen, 2011), 289–302.
150 Robin Cormack, “Looking for Iconophobia and Iconoclasm
in Late Antiquity and Byzantium,” in Iconoclasm and Text Destruc-
tion in the Ancient Near East and Beyond, ed. N. N. May, OIS 8
(Chicago, 2012), 476.
Fronted First-Person Pronoun in Syro-Anatolian Monumental Discourse ) 339
The Decalogue, on the other hand, is in fact treated as
a legitimate manifestation of the deity.151 I have argued
elsewhere that the Decalogue was in part capable of
functioning analogously to an image in the literary
environment of Exodus—apart from any descriptions
of iconography—precisely because it opened with the
“I (am)” formula. This announced that it was to play
a similar role to an image.152
Conclusions: The Monumentality of the “I (Am)”
Formula between Iconicity and Orality
The frequent use of the “I (am)” formula and images
to augment one another, the striking use of the “I
(am)” formula as an image, iterations in which the im-
age and text mismatch, and the occasional use of the
“I (am)” formula without an image where one would
otherwise be expected all point to the same underlying
function. The “I (am)” formula functioned similarly
to an ancient Near Eastern image, but it accomplished
this in a unique way. Therefore, its degree of over-
lap was variable and culturally speciﬁc. The “I (am)”
formula may not have originated to interact with im-
ages, but its overlapping function allowed it to engage
with images in increasingly sophisticated ways. This
same underlying function allowed the formula to be
separated from images again as it was iterated in new
contexts. What remained constant, however, was its
production of presence through deictic projection, a
manifestational function that was accomplished by im-
ages as well through a dierent imaginative means.
The function of manifesting the agent through de-
ictic projection is why the fronting of the ﬁrst-person
pronoun was so signiﬁcant an innovation. While the
use of “I”-deixis in the earlier “. . . (am) I” formula
arguably produced the same eect, that eect was de-
layed. The fronted ﬁrst-person pronoun, on the other
hand, initiated deictic projection from the beginning
151 Hurowitz, “What Can Go Wrong with an Idol?,” 300; Na-
thaniel B. Levtow, “Text Destruction and Iconoclasm in the He-
brew Bible and the Ancient Near East,” in Iconoclasm and Text
Destruction, ed. May, 330–31; Levtow, “Monumental Inscriptions
and the Ritual Representation of War,” 37–38.
152 Hogue, “Monumentality of the Sinaitic Decalogue,” 86–90.
of the inscription. Beginning with the ﬁrst-person pro-
noun creates a short-lived tension in which the users
are given an empty discursive space that must be ﬁlled.
The name and titulary then allow the agent to step
into that space and become present with the users.
The “I (am)” formula was therefore the most opera-
tive clause in a monumental inscription for producing
that text’s monumentality—that is, the aordance of
special meaning through provoking communal en-
gagement with an object.153 As a result, the formula
became a potential source of authority by virtue of
its ability to provoke communal interaction with the
texts it headed. The “I (am)” formula evoked a socially
The spread of the “I (am)” formula illustrates a net-
work of artistic interactions in the eastern Mediterra-
nean during the Iron Age that resulted in the formation
of a monumental koiné. The formula was an attractive
element of monumental rhetoric to adapt, developing
a new strategy of manifesting people through monu-
ments. Images could now be augmented by a formula
that would specify their function and in fact outstrip
them by also manifesting the agent’s voice and verbal-
ized agency. If necessary, images could now be forgone
altogether, or else used for other ends because the “I
(am)” formula was already performing their duty. This
was why the “I (am)” formula was so closely aligned
with images, at various times either becoming one itself
or replacing them altogether. This is why inscriptions
with no attendant iconography such as the Mesha In-
scription, BULGARMADEN, or HSARCIK 1 could still func-
tion. And this is why the composers of the Decalogue
chose to literarily create a theophany with the “I (am)”
formula. The formula manifested the persons it des-
ignated in a way that allowed them to address a com-
munity with an authoritative voice. It created a sense
of authoritative orality which was materialized in ways
meant also to evoke or replace iconicity.
153 Wu, Monumentality in Early Chinese Art and Architecture,
11; Osborne, “Monuments and Monumentality,” 3–4; Pauketat,
“From Memorials to Imaginaries in the Monumentality of Ancient
North America,” 441–42.