SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

Points of View: The Theodosian 
Obelisk Base in Context 

Linda Safran 

THE OBELISK with its sculpted base in the former Hippo­
drome of Constantinorle (PLATES 1-3) is well known as 
a rare datable work 0 Late Antique art. 1 A sixth-century 

source puts the raising of the obelisk in the year 390,2 and Greek 
and Latin epigrams on the plinth (the lower part of the base) 
credit Theodosius I and the urban prefect Proclus with this 
feat.3 The inscriptions lack punctuation and their precise trans­
lation is difficult, but the sense is clear enough. The Latin 
(PLATES 2, 4) may be rendered as: 

Formerly difficult, I was ordered to obey the peaceful masters 
and to raise 

I The only major monograph on the oblelisk base is G. Bruns, Der Obelisk 
und seine Basis auf dem Hippodram zu Konstantinopel (=1 stForsch 7 
[Istanbul 1935]), reviewed by H. Kahler, PhilWoch 59 (1939) 93-99; J. 
Kollwitz, Gnomon 13 (1937) 423-27; E. Weigand, BZ 37 (1937) 452-58; T. G. 
Allen, AJA 41 (1937) 157ff; see also A. W. Byvanck, De Obelisk van 
Constantinopel (=MedAmsterdam N.R. 23 11 [Amsterdam 1960]); R. Janin, 
Constantinople byzantine 2 (Paris 1964); G. DAGRON, Naissance d'une capi­
tale. Constantinople et ses institutions de 330 a 451 (Paris 1974: hereafter 
'Dagron'); E. Iversen, Obelisks in Exile (Copenhagen 1972) II 9-33. The 
obelisk base figures in most surveys of Early Christian and Byzantine art as 
well as many textbooks of history. Bibliography on specific aspects of the base 
is given in succeeding footnotes. 

2 Marcellinus Comes, in T. Mommsen, cd., Chron. Min. II (=MGH, AA 11) 
62. 

3 CIL III 737 (ILS 821; PLATE 4): difficilis quondam, dominis parae serenisl 
iussus et extinctis palmam portare tyrannis.l omnia Theodosio cedunt 
subolique perenni.l ter denis sic victus ego domitusque diebus,l iudice sub 
Praclo superas elatus ad auras; CI G IV 8612 (W. R. Paton, The Greek 
Anthology III [Loeb edition (London 1925)] 9.682; PLATE 5): Klova 'tE"tp<i-
1tAEUPOV ad x86vt KEi~Evov o.x8oc) ~ouvoc; avacr'tTlcrm 8EuoocrtoC; ~acrtAd)c) 'to A­
~Ttcrac;' II POKAoC; £1tEK£KAE'tO, Ka 1 'tocroc; Ecrnll KtooV 1i EAtotC; £v 'tpt<lKov'ta 01>00. 
For the literature on the inscriptions see esp. C. Mango, "The Byzantine 
Inscriptions of Constantinople: A Bibliographical Survey," AJA 55 (1951) 62; 
A. Cameron, "A Biographical Note on Claudian," Athenaeum 44 (1966) 
32-40. 

409 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

410 THE THEODOSIAN OBELISK BASE 

the palm after the tyrants were dead. Everything gives way to 
Theodosius 

and his eternal offspring. In this way conquered and 
mastered in three 

times ten days, under Proclus the Prefect, I was raised to the 
high air. 

The Greek inscription (PLATES 3, 5) records: 

Only Emperor Theodosius dared to erect the four-sided 
column which had 

lain heavy on the earth for a long time. Proclus was summoned, 
and this 

so enormous column was put up in thirty-two suns. 

The content of the inscriptions is investigated in greater detail 
below, but for now it is sufficient to note that the name Proclus 
has been effaced and then rather clumsily re-inscribed. This 
emendation must have occurred between 392, when Proclus 
was removed from office, and 395, when his opponent fell 
from power. 4 There is thus no significant debate about either 
the date of the work, early in the last decade of the fourth 
century, or its motivation, celebration of Theodosius' victory 
over "the tyrants," no doubt Maximus Magnus and his son 
Victor.:> The lack of specificity in the inscriptions is not 
surprising, as unsuccessful usurpers were never mentioned by 
name in inscriptions put up by the victors. 6 

Nor is there any real confusion about the essential 
iconography: all four sides of the upper base show members of 
the imperial house in an architectural setting generally identified 
as the kathisma, the imperial loge of the Great Palace adjoining 
the Hippodrome. The imperial suite, flanked by various 
functionaries, overlooks either simple tiers of spectators on the 
southwest (PLATE 6) and northeast (PLATE 7), or spectators 
supplemented by musicians and dancers on the southeast 

4 On the controversy among Proclus, his father Tatian, and Rufinus see S. 
Rebenich, "Beobachtungen zum Sturz des Tatianus und des Proculus," ZPE 
76 (1989) 153-65. 

5 See C. V. E. Nixon, tr., Pacatus, Panegyric to the Emperor Theodosius 
(Liverpool 1987). 

6 "It was an almost universally observed principle that one never mentioned 
an unsuccessful usurper by name. It was always enough to refer to him 
contemptuously just as 'tyrannus"'; this principle is well illustrated by 
Claudian's panegyrics to Theodosius: A. Cameron, "Some Prefects Called 
Julian," Byzantion 47 (1977) 62, citing R. MacMullen, "The Roman Concept 
Robber-Pretender," RJDA SER.3 10 (1963) 221-225. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

SAFRAN PLATE 1 

Old photograph of the Theodosian obelisk and upper part 
of base from the southeast. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

PLATE 2 SAFRAN 

Old photograph of the Theodosian obelisk base from the south, 
showing southwest (left) and southeast (right) sides. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

SAFRAN PLATE 3 

Old photograph of the Theodosian obelisk base from the north, 
showing northeast (left) and northwest (right) sides. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

PLATE 4 SAFRAN 

Latin inscription on the southeast side of the base of the 
Theodosian obelisk. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

SAFRAN PLATE 5 

Greek inscription on the northwest side of the base of the 
Theodosian obelisk. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

PLATE 6 SAFRAN 

Southwest side of the base of the Theodosian obelisk, upper part. 
Photo by P. Steyer, 1964. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

SAFRAN PLATE 7 

Northeast side of the base of the Theodosian obelisk, upper part. 
Photo by P. Steyer, 1964. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

PLATE 8 SAFRAN 

Southeast side of the base of the Theodosian obelisk, upper part. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

SAFRAN PLATE 9 

Northwest side of the base of the Theodosian obelisk, upper part. 
Photo by D. Johannes, 1993. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

PLATE 10 SAFRAN 

Old view of the Hippodrome at Constantinople from the southwest, 
showing the Built Obelisk, the Serpent Column, and the Theodosian 
obelisk and base, with Hagia Sophia in the distance. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

SAFRAN PLATE 11 

'" .. 

Plan of the Hippodrome at Constantinople and surrounding area in 
the 1920s. Surviving Byzantine masonry, including the sphendone 
and parts of the cavea, is in black. 

• 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 411 

(PLATE 8) or by kneeling barbarians on the northwest (P LATE 
9). The sides of the base with these more narrative elements, 
the southeast and northwest, are above the two sides of the 
plinth that bear inscriptions in Greek or Latin. The sides of the 
base that depict only functionaries and spectators (the south­
west and northeast) are above the sides of the plinth that 
contain narrative scenes of chariot-racing and obelisk-raising. 
Whether or not such a two-level base was planned from the 
outset/ the two parts were certainly installed simultaneously in 
the last decade of the fourth century, when the obstinate 
obelisk was finally conquered by Theodosius and raised by his 
urban prefect. 

The obelisk base has often been studied from the perspective 
of style. It has served as the key monument in identifying a so­
called Theodosian court style, which is usually described as a 
"renaissance" of earlier Roman classicism. 8 Both more and less 
hieratic modes of representation, however, are used on the 
base, even on a single side. An example is the southeast side of 
the upper base (PLATE 8), with the stiffly posed imperial suite 
above and the sketchily rendered musicians and dancers below. 
The reliefs on the northeast and southwest plinth (PLATES 2, 3) 
are also done in this less hieratic, even 'impressionistic', style. 
Therefore, if there was a Theodosian court style, it varied so 

7 According to A. Effenberger's unpublished theory, the plinth was 
originally the sale base for the obelisk, but when the obelisk was broken, the 
carved base was interposed to make up for the lost height. This suggests a 
precise desired height for the whole obelisk-cum-base ensemble. lowe this 
information to Dr Albrecht Berger. 

8 See A. Taylor, "Stylistic Variety in Constantinopolitan Stone Relief of the 
Theodosian Period," ByzSt 10.2 (1983) 184-201; E. Kitzinger, Byzantine Art in 
the Making (London 1977) 32ff, and" A Marble Relief from the Theodosian 
Period," DO P 14 (1960) 17-42; J. Kollwitz, Ostromische Plastik der 
theodosianischen Zeit (Berlin 1941); S. Sande, "Some New Fragments from 
the Column of Theodosius," ActaAArtIlist N.S. 1 (1981) 1-78, with some 
serious misstatements about the spectators' relationship to the obelisk base; H.­
G. Severin, "Ostromische Plastik unter Valens und Theodosius I," jBerlMus 
12 (1970) 211-52; B. Brenk, "Zwei Reliefs des spaten 4. Jahrhunderts," 
ActaAArtHist 4 (1969) 51-60; P. Angiolini-Martinelli, "Tradizione e rin­
novamento nelle sculture della base dell'obelisco di Teodosio a Costan­
tinopoli," CorsiRavenna 22 (1975) 47-62. The base was originally dated on 
stylistic grounds to the time of Constantine: A. J. B. Wace and R. Traquair, 
"The Base of the Obelisk of Theodosius," JHS 29 (1909) 60-69. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

412 THE THEODOSIAN OBELISK BASE 

widely in its representation of figures that the label IS not very 
usefu1.9 

The base has received similar attention in terms of iconog­
raphy. It has been of great interest to those who interpret some, 
or indeed many, of the sculpted figures as portraits. Scholars 
have attempted with differing results to identify the imperial 
figures and their neighbors as specific members of the Theo­
dosian house or the Constantinopolitan aristocracy.lo Various 
Hippodrome and civic functionaries have been categorized on 
the basis of their clothing and attributes. The base has also been 
examined from the viewpoint of realistic architecture, with the 
reliefs used to reconstruct the appearance of the kathisma and 
the Hippodrome. II It has been considered a reservoir of infor­
mation about material culture, depicting as it does different 
clothing options, a variety of musical instruments, and even 
tools demonstrating Roman technological know-how. 12 Finally, 
ethnographic information has been read into the base, with the 
kneeling barbarians on the northwest side (PLATE 9) and the 
imperial Gothic bodyguards on all four sides distinguished by 
their costume and physiognomy. A historiographic inquiry into 

9 The best stylistic comparison for the narrative style on the plinth certainly 
postdates it, according to all recent scholarship. This is the fifth- or sixth­
century" Kugelspiel" found in the Hippodrome and now in Berlin. It depicts 
circus scenes and perhaps the kathisma door, but the patronage and audience 
for this portable work cannot be ascertained. See A. Effenberger and H.-G. 
Severin, Das Museum fur spatantike und byzantinische Kunst (Mainz 1992) 
116ff; A. Cameron, Porphyrius the Charioteer (Oxford 1973) 33f, 58ff. 

10 Most recently S. Rebenich, "Zum Theodosiusobelisken in Konstan­
tinopel," IstMitt 41 (1991) 447-76; Kollwitz (supra n.l) 423-27; H. Wrede, 
"Zur Errichtung des Theodosiusobelisken in Istanbul," I stMitt 16 (1966) 
178-98; H. P. L'Orange, "Zum Alter der Postamentreliefs des Theodosius­
Obelisken in Konstantinopel," Likeness and /con. Selected Studies (Odense 
1973) 206-209, and Studien zur Geschichte des spatantiken Portrats (Oslo 
1933) 66-70; J. C. Baity, "Hierarchie de l'empire et image du monde. La face 
nord-ouest de l'obelisque theodosien a Constantinople," Byzantion 52 (1982) 
60-71; E. Demougeot, "Obelisques egyptiens transferes a Rome en 357 et a 
Constantinople en 390," in id., L'Empire romain et les Barbares d'Occident, 
IV'- V II' siecie. Scripta varia (Paris 1988) 253-72. 

II R. GUILLAND, Etudes de topographie de Constantinople byzantine I 
(Berlin 1969: hereafter 'Guilland'); Cameron (supra n.9); A. Piganiol, "La loge 
imperial de I'Hippodrome de Byzancc et Ie problcme de I'Hippodrome 
couvert," Byzantion 11 (1936) 383-90; G. Egger, "Die Architekturdarstellung 
im spatantiken Relief," JKSW 55 (1959) 1-30. 

12 H. Kahler, "Der Sockel des Theodosiusobelisken in Konstantinopel als 
Denkmal der Spatantike," ActaAArtHist 6 (1975) 45-55. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 413 

the scholarly literature on the base would doubtless yield inter­
esting insights into scholarly fashions, trends, and agendas over 
the past century. 

This paper will not address any of the stylistic or iconographic 
issues that have fueled scholarship to date. Instead, it will 
attempt to redress a serious omission by placing the obelisk 
base in its original context, the late fourth-century Hippodrome 
of Constantinople. With a capacity estimated at 100,000 
people,13 the Hippodrome was the city's public gathering place 
par excellence. It will be shown that the obelisk base was 
designed and positioned in a particular way to relate to its 
heterogeneous public, to communicate certain messages and to 
elicit certain responses. Today recontextualization is difficult, 
given the state of preservation of the Hippodrome, but it is an 
essential task if one hopes to approach an understanding of the 
messages and combinations of meaning on the four sides of the 
base. Some of these messages were constructed by the 
emperor or his advisors and their craftsmen; other meanings 
were brought to the base by the audiences on each side. Both 
authorial intent and audience response deserve consideration in 
attempts to grasp the meanings of a Late Antique work. 

To understand intent and response, it is first necessary to 
establish as closely as possible the composition of the audiences 
in the late fourth century. For this we have a few literary 
sources as well as limited archaeological and art historical 
evidence. Unfortunately, much later material is irrelevant: the 
role of the Hippodrome in urban life changed markedly in the 
late fifth and sixth centuries, and later sources have to be used 
with great caution, if at all, in reconstructing the fourth-century 
situation. We can supplement the meager sources by applying 
some general principles derived from reader-response criticism 
and reception aesthetics as formulated in recent decades in the 
field of literary criticism. Studies in visual theory and the 
psychology of perception offer additional perspectives. Al­
though these methods constitute useful new tools for the 
historian of art, the objective remains a traditional one: a better 
understanding of the monument, including its original context, 
its patronage, and its audiences. 

13 A. Cameron, Circus Factions (Oxford 1976) 236. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

414 THE THEODOSIAN OBELISK BASE 

I 

The obelisk base is almost always discussed and photographed 
in close-up, corresponding to modern interests in its iconog­
raphic and stylistic details. But the base was never intended to 
be seen that way; probably only a small team of sculptors ever 
had that perspective and only for a short span of time. The Late 
Antique audiences in the Hippodrome would have seen the 
monument in completely different ways. First of all, they saw 
an entire monument-the obelisk and its two-part base­
dominating the other objects on the barrier that divided the two 
tracks of the circus (PLATES 1, 10). The obelisk base was only 
one among many possible visual foci in the Hippodrome, and 
its visibility was regularly obscured by chariots racing past. An 
individual spectator could see only one or two sides of the 
carved and inscribed base, never three or four. It was far away 
from some viewers and nearer, but never really close, to many 
others. 

The four sides of the obelisk base faced different sides of the 
circus, and there were different spectators on each side (PLATE 
11). This fact probably dictated certain considerations in the 
production of the work. While this seems logical, even obvious, 
it has not been discussed-it has barely even been noted-in 
the secondary literature. 14 This is especially surprising if, as 
seems probable, the sculpture was carved in situ in the 
Hippodrome, with full cognizance of the encircling seating 
patterns.15 At the very least the inscriptions, which record how 
many days it took to raise the obelisk, could only have been 
carved once the obelisk was up. 

The first problem is to differentiate the audiences and deter­
mine who sat where in the Hippodrome. There are no con­
temporary sources, but the sixth-century historian John Malalas 
records a change in seating patterns only a few years later in the 
reign of Theodosius II; there are scattered references in other 
historians; and the Book of Ceremonies includes earlier as well 
as contemporary tenth-century information on Hippodrome 

14 A. Grabar, L 'empereur dans l'art byzantin (Paris 1936) 66, noted that two 
sides of the base were more visible than the others. 

15 Suggested by Allen (supra n.1) 157ff; Kahler (supra n.12) 46. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 415 

ceremonial. I6 In addition, the archaeological evidence from 
circuses throughout the Roman world, and artistic depictions of 
these in many media, prove the consistency and conservatism 
of the circus as a building type. 17 The Circus Maximus in Rome 
was clearly the chief exemplar and model, but consistent 
features are recorded in third- and fourth-century mosaics, as 
well as in reliefs and on such portable objects as gems and 
lamps. Some later ivory diptychs also show circus scenes. Only 
a few of these, however, depict the cavea, the seating area of the 
circus, with any detail (Humphrey 243). 

On the basis of textual, art historical, and archaeological 
evidence, we can be sure that the emperor and his retinue sat 
on the southeast side in the kathisma box connecting the 
Hippodrome with the Great Palace (Fig. 1). In Rome's Circus 
Maximus, the equivalent royal box was similarly located along 
the left-hand track, looking from the viewpoint of the carceres 
(starting gates; Humphrey figs. 53-54). In the fourth-century 
Circus of Maxentius in Rome, the royal box was also on the left 
side, although closer to the starting gates (Humphrey fig. 278). 
This was also the location of the tribunal in smaller provincial 
circuses.18 

The precise location of the kathisma at Constantinople is un­
known, and it is not clear whether it directly faced the Theo­
dosian obelisk, which may have marked the physical center of 
the Hippodrome or perhaps the cen ter of the barrier dividing 
the two tracks (PLATE 11).19 Judicial archives were housed 

16 Malalas, ch. 351, in E. and M. Jeffreys and R. Scott, trr., The Chronicle of 
John Malalas (=Byzantina Australiensia 4 [Melbourne 1986]) 191; Guilland 
371-595 passim. 

17 See now the comprehensive study by J. H. H UMPHREY, Roman Circuses. 
Arenas for Chariot Racing (London 1986: hereafter 'Humphrey') with earlier 
bibliography. 

18 On the kathisma see Cameron (supra n.9) 50££; Guilland 462-98; 
Humphrey 78ff. At Sirmium, Milan, and Thessaloniki, exceptionally, the royal 
box seems to have been situated on the right side of the circus: Humphrey 
636ff. 

19 Augustus' obelisk in the Circus Maximus (now in the Piazza del Popolo) 
marked the midpoint of the barrier; Constantius' erected in 357 (now in front 
of the Lateran) marked the midpoint of the entire Circus. On these obelisks 
see Iversen (supra n.l) I 55-75. A barrier must have existed despite the absence 
of any traces (see S. Casson, "The Excavations," in Casson et aL., Preliminary 
Report Upon the Exca'lJations Carried Out in the Hippodrome of Constan­
tinople in 1927 [London 1928J 9£), but its height in the fourth century cannot 
be determined; in the thirteenth century according to Robert of Clari it was 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

416 THE THEODOSIAN OBELISK BASE 

under the kathisma and, because the urban prefect and sena­
torial commissions sometimes made judgments in the Hippo­
drome, it is likely that they had offices, and therefore seats, 
somewhere along the southeast side.20 

On the northwest side, facing the kathisma, the partisans of 
the four racing colors sat in separate wooden grandstands (Fig. 
1). At the end of the fourth century the Blues were directly 
opposite the kathisma, as proved by the reshuffling of seats in 
400 by Theodosius II, who preferred looking across at his 
fellow Green supporters (Malalas ch. 351 Ueffreys 191J). The 
details of the so-called factions and their roles in later Byzantine 
history need not concern us here. But as Cameron has shown, 
these fans were not just ordinary citizens, but young men with 
the leisure to indulge their intense interest in chariot racing 
and-especially in later centuries-in mischief. Loud as they no 
doubt were, they represented but a tiny proportion of the race­
viewing population as whole. In 602, the only year for which 
precise numbers are recorded, there were 1,500 Greens and 900 
Blues in a population numbering between 500,000 and 750,000. 21 

There is no reason to suspect a larger number of partisans two 
centuries earlier, when the population of Constantinople was 
smaller, probably closer to 250,000. 22 

The fans were a part, to meros, of the whole demos. That the 
young men had a distinct seating area, at least in later centuries, 
is further suggested by a reference to sculpture located "in the 
place called N eolaia," the "young men's area" of the Hippo­
drome. 23 Also on the northwest side, seated alongside the circus 
partisans, were members of the garrison troops; they occupied 
six sections of the grandstands (Malalas ch. 351). Appropriately 
enough, these most avid fans were closest to the finishing line, 
which in most circuses was near the center of the right-hand 

"a good fifteen feet" high, and this has been widely repeated (e.g. Cameron 
[supra n.9] 181). See also C. Mango, "L'euripe de l'hippodrome de Con­
stantinople," REByz 7 (1949) 180-93. 

20 Malalas ch. 340 Ueffreys 185); Dagron 239,317. 
2! Cameron (supra n.13) 20, 120, citing Theoph. Sim. Hist. 8.7.11. 
22 Dagron 518-25; see also D. Jacoby, "La population de Constantinople a 

l'epoque byzantine: un probleme de demographie urbaine," Byzantion 31 
(1961) 81-109. John Chrysostom says there were 100,000 Christians in the city 
ca 400: In Act. Apost. Hom. 11.3, Migne, PC LX 96ff. 

23 A. Cameron and J. Herrin, trr., Constantinople in the Early Eighth 
Century: The Parastaseis syntomai chronikai (Leiden 1984) 6of, 171 f. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

Greens :. 
j> 

Circus Partisans 
..... / ..... . 

Sphendone 

LINDA SAFRAN 417 

Carceres 

1/ I; / 
/ / / Circus Functionaries 

// 
// 

k<iihisma 
.... .... / 

r"'"'' 
. .. / 

.: Government Functionaries 

Judicial Archives 
Senatorial Offices 

Key 

Great 
Palace 

@@ Theodosian Obelisk Base 

@ Serpent Column 

t8J Built Obelisk 

Figure 1: Diagram of the Hippodrome at Constantinople, showing 
parts of the circus, relative location of monuments on the central 
barrier, seating arrangements (boldface), and activities by location 
(italics). Drawing by R. A. Fellerman. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

418 THE THEODOSIAN OBELISK BASE 

track; such early third-century mosaics as those in Lyon and 
Silin (Libya) show the finishing line. 24 The adjacent white line 
depicted in the Lyon mosaic is the break line, at which the 
chariots could leave the lanes assigned to them at the starting 
gate in order to move to the inside position. 

To the southwest, along the curved end of the Hippodrome 
called the sphendone, executions were occasionally staged; the 
emperor Valentinian watched his chief eunuch burn alive there 
in 370. 25 But mostly the chariots would pile up in serious 
accidents here,26 as depicted in many circus mosaics, including 
Silin and Piazza Armerina.27 The sphendone, including the 
extremes of the two long flanks of the Hippodrome, probably 
housed the 'cheap seats', as in a modern football stadium. These 
spectators were far away from the emperor, from the most avid 
partisans, from the starting gates, and from the finish line (cf 
Guilland 376); nevertheless, their occupants no doubt voiced 
loud support for their favorite charioteers. The highest seats 
were surely reserved for the lowest classes of the population, 
the non-citizens and slaves (Humphrey 122). 

Finally, it is unlikely that any sizeable group watched from the 
carceres to the northeast, housing the starting gates at ground 
level and utilitarian spaces for the charioteers, circus 
functionaries, and organizers of the races above. In the fourth­
century Piazza Armerina mosaic, the cavea is interrupted by 
the carceres and a circus magistrate is depicted alone in his 
upper-story box (Humphrey fig. 112). Only an early third­
century mosaic from Carthage shows seating completely 
around the circus, including over the carceres, but this mosaic 
seems heavily influenced by the iconography of amphitheaters 
in which the seating completely encircled an oval arena 
(Humphrey 210 and fig. 63). 

From the Theodosian and Justinianic law codes it can be 
inferred that women and clergy were sometimes a part of the 
Hippodrome audience. In the fourth century they might have 

H Humphrey 84ff, figs. 36 (Lyon), 107 (Silin). 
25 Malalas ch. 340 a effreys 185); Cameron (supra n.l3) 172; later examples of 

executions and mutilations in Guilland 375ff. 
26 Guilland 375; Dagron 333. 
27 Humphrey 211 ff (Silin), 223ff, fig. 112 (Piazza Armerina) with additional 

bibliography. The Piazza Armerina mosaic surely represents the Circus 
Maximus, whereas the Silin example shows more generic circus iconography: 
Humphrey 230. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 419 

been seated on any of the three sides for spectators, although 
females of imperial rank probably viewed the proceedings from 
the upper levels of the kathisma.28 Women are not identifiable, 
however, in any of the surviving artistic representations of 
circus scenes, including the relatively detailed mosaics at Piazza 
Armerina and at Gafsa, a late fifth-or early sixth-century site in 
Tunisia (Humphrey fig. 72). 

II 

The obelisk base confirms the textual and material evidence 
for seating in the Hippodrome. Most significantly, the two in­
scriptions on the southeast and northwest sides correspond to 
the most important and most distinctive groups of spectators. 
This can hardly be accidental. Why orient the inscriptions in 
this manner, or indeed have inscriptions at all, if they could not 
be seen and appreciated in some way? And in fact, the 
languages of the inscriptions appear to correspond with the 
literacy of the two audiences. 

On the southeast, facing the kathisma, is the inscription in 
Latin (PLATE 4 and Fig. i)-the language of power, the language 
used in the fourth century only by the court, the bureaucracy, 
and the army.29 Latin was also the language of the law, which 
would correspond to the presence of judges on the southeast 
near the judicial archives. Most people in Constantinople spoke 
Greek, the lingua franca of the eastern Roman Empire and the 
language of the inscription on the northwest facing the circus 
partisans (PLATE 5 and Fig. 1). The absence of inscriptions on 
the southwest and northeast sides may well reflect the illiteracy 
of the masses, which was certainly common in the fourth 
century.3D There is another possibility as well. Inhabitants of 

28 Women sat with men in the Circus at Rome: ]. P. V. D. Balsdon, Life 
and Leisure in Ancient Rome (New Yark 1969) 258f. Women and priests 
evidently infiltrated the Hippodrome. In the sixth century a wife's presence 
there was grounds for divorce, and ecclesiastics had to be explicitly prohibited 
from attending: Cod. lust. 1.4.34.3, Nov. lust. 117, cited in O. Pasquato, Cli 
spettacoli in S. Ciovanni Crisostomo (=OrChrAn 201 [Rome 1976]) 93. 

29 G. Dag~on, "Aux origines de la civilisation byzantine: Langue de culture 
et langue d'Etat," RIIist 241 (1969) 23-56, esp. 25, 36, 39. 

30 In general see W. V. Harris, Ancient Literacy (Cambridge [Mass.] 1989), 
who notes that illiterates could serve as decurions by the end of the third 
century (292) and that all types of honorific and commemorative inscriptions, 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

420 THE THEODOSIAN OBELISK BASE 

Constantinople, a cosmopolitan city, spoke Gothic, Coptic, 
Armenian, Georgian, Syriac, and probably other tongues in 
addition to Greek; all these languages were used in preaching. 31 

Perhaps this multilingualism militated against an inscription in 
one of the two most prominent local languages. In addition, the 
presence or absence of inscriptions indicates an awareness of 
the setting, including the spectators' relative distances from the 
obelisk base (further discussed below). 

It has been argued that the content of the Theodosian inscrip­
tions echoes that on Constantius' obelisk erected at Rome in 
the Circus Maximus in 357. 32 Although both Latin inscriptions 
allude to an unnamed tyrant, the texts are not that close. Very 
few would have recognized competition between the Old and 
New Rome in the form of textual emulation. More significant is 
the change of language in the later obelisk base. The Roman 
inscription in Latin contrasts with those in Latin and Greek 
intended for a polyglot audience. Further, images supplement 
the Constantinopolitan inscriptions and may explain their 
relative brevity. 

The Latin inscription of the base, written from the viewpoint 
of the obelisk, almost takes the form of a riddle. This is good 
classical epigrammatic style: a quick look at the Anthologia 
Graeca reveals a large number of inanimate objects-houses, 
fountains, baths, even mosquito nets-presenting themselves in 
the first person.33 The text communicates power with the repeti­
tion of such words as "ordered," "obey,» «conquered,» and 
"mastered." The alliteration of dominis and domitusque echoes 
and reinforces this sense of mastery. It also argues against a 
suggestion that domitusque is a cutter's error for duobusque, 
which would have reconciled the thirty days needed to raise 
the obelisk here with the thirty-two days recorded in the 
Greek inscription. 34 We should interpret this inconsistency 

including otherwise numerous papyri, declined in number during the fourth 
century (298, 3 I 7). 

)1 R. Browning, "The Language of Byzantine Literature," in id., History, 
Language and Literacy in the Byzantine World (Northampton 1989) XV 
(108). 

)2 Wrede (supra n.10) 188ff; Iversen (supra n.I) I 57£. 
)) Paton (supra n.3) 9.631, 651 ff, 667, 764 et al. 
34 Suggested by Cameron (supra n.3) 34, but Janin (supra n.1: 190) already 

replaced the clearly legible domitusque in line 4 with duobusque. lowe this 
observation to Lee F. Sherry. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 421 

partly as a product of choosing oblique and pedantic Latin hex­
ameters that are epic in form and content. Such stock phrases as 
"three times ten days" have been adopted from famous Latin 
authors, including Vergil. 

Where the Latin text is generalized, the Greek text is specific. 
Instead of the epic conceit of the speaking obelisk, we find a 
straightforward third-person narrative in pentameter about a 
four-sided column. Although the translation is somewhat un­
clear, the specificity of thirty-two suns, or thirty-two days, or 
the morning of the thirty-second day is very different from the 
Vergilian "three times ten." Both the epic and elegiac meters 
would be familiar to such a highly literate audience as the im­
perial court, but the factual specificity of the Greek text may 
have had greater appeal to a less epically inclined audience. It is 
worth noting that the name of Theodosius is centralized both 
visually and textually in both the Latin and Greek inscriptions 
(PLATES 4, 5). 

Regardless of the degree to which its contents could be under­
stood, the mere fact of an inscription was significant. MacMul­
len has demonstrated that "the epigraphic habit ... was taken 
seriously," and although the height of the epigraphical fashion 
antedated the Theodosian obelisk base by some two centuries, 
the implications of a formal inscription in stone must still have 
been clear. The ability to erect an inscription was proof of social 
status in antiquity, a status naturally held by the emperor and his 
representative, the urban prefect. An inscription was important; 
it commanded attention and respect; and it was usually ad­
dressed to a whole community, 35 even if its contents were not 
accessible to all or even most members of its audience. The 
same must be even more true for a large-scale work of public 
art that so prominently incorporates inscriptions. 

Besides the language of the inscriptions, the artistic language of 
each side of the base can be linked to its intended audience. As 
noted above, the iconography is repetitive but it is also dif­
ferentiated on each side. The quality of the reliefs also differs on 
each side. This is not a question of diverse styles or of stylistic 
'modes', which are indeed present on the base; nor is it a 
product of different degrees of wear or weathering. There is no 
such thing as an objective aesthetic judgment, but generally 

35 R. MacMullen, "The Epigraphic Habit in the Roman Empire," AJP 103 
(1982) 244, 246. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

422 THE THEODOSIAN OBELISK BASE 

agreed-upon qualitative criteria do exist and observations made 
in situ can be supplemented by old photographs (PLATES 6-9). 

The southeast side facing the imperial box-surely the best 
carved (PLATE 8)-has both the greatest number and highest 
degree of individualized figures and faces. Here we see the 
most varied draperies, the sharpest relief, and the most solid­
looking architecture with the most inventive grillework. The 
next-best-carved side, the northwest facing the circus partisans 
(PLATE 9), also has some sensitive portraits. Only on these two 
sides is there some compositional modulation-some of the 
Gothic guards turn away from the emperor, adding variety to 
the otherwise centrifugal and static poses. Nevertheless, the 
carving of the architecture on the northwest is a little less fine 
than on the opposite side of the base: the kathisma arch seems 
squeezed into the available space and the balustrade curves up 
at the left side. 

In contrast to the high artistic quality of the southeast and 
northwest sides, the figures on the southwest-all with identical 
wavy-cap hairstyles-are very monotonous (PLATE 6). But the 
carving on the northeast is the least accomplished of all (P LA TE 
7): the relief is very flat and the figures outside the kathisma are 
all cookie-cutter repetitions, frontal and stiff. The presence of a 
vertical channel, which made the base into a fountain, also 
shows that this was perceived as the least important side, even if 
the channel was originally less conspicuous. This hydraulic 
function was probably planned along with the reliefs and was 
not a later addition, because the overall upper-base composition 
of kathisma and flanking functionaries has been maintained and 
no figures have been destroyed. 36 

Byvanck noted that the northeast face differed from the 
others and attributed it to a change in style during the reign of 
Theodosius' son Arcadius. 37 But the idea that the sides of the 
obelisk base were carved at different times is problematic. A 
change in ruler is unlikely to alter sculptural quality or even 
style. The obelisk base was not such a huge project that its 
completion required many years; we know of no specific 
historical circumstances that might have mandated a delay; and 
it is so centrally and publicly located that its unfinished state 

36 Scholars disagree about when this was done and how the fountain would 
have looked: cf Bruns (supra n.1) 18ff with F. Krauss' appendix (85f). The 
circus barrier was typically composed of water basins and fountains, hence 
the name" euripos" channel. 

37 A. W. Byvanck, L 'art de Constantinople (Leiden 1977) 45. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 423 

might well have been a dynastic embarrassment. It is difficult to 
agree with Bruns that the southeast and southwest sides were 
carved under Theodosius and the northwest and northeast 
sides under Arcadius. This assumption led Bruns to label the 
south sides (PLATE 2) as Al and A2 and the north sides (PLATE 
3) as Bl and B2.38 These labels introduced highly tenuous 
notions of style and patronage that have attained the status of 
fact through frequent repetition in the secondary literature. The 
sculptor of the high-quality southeast side, however, is un­
likely to have executed the mediocre southwest face and the 
sculptor of the northwest did not carve the weak northeast 
side. Instead, better artists were most likely responsible for the 
southeast and northwest sides, which face the most important 
spectators. 

III 

At this point we must return to the fundamental question of 
context. How visible were the iconographic and qualitative 
differences to their respective audiences? Could the spectators 
on different sides of the circus see and appreciate the nuances 
of epigraphical and artistic language? First we shall address the 
question of visibility and legibility, then assess the different . , 
vIewers responses. 

The best visibility is clearly from the long sides of the Hippo­
drome, the sides closest to the base-in other words, the sides 
occupied by the imperial retinue on the southeast and the 
circus partisans on the northwest. In his recent study of Roman 
circuses, Humphrey calculated (588) that the minimum ade­
quate single-track width was just under 30 m., which can sup­
port twelve chariot teams at a breadth of 2.5 m. per team. The 
average width of Late Antique circuses was 67-79 m., including 
the two tracks plus the width of the central barrier with its 
water basins and statuary (Humphrey 635). The interior width 
of the Hippodrome at Constantinople, based on the surviving 
dimensions of the sphendone established by twentieth-century 
excavation and supplemented by Hero of Byzantium's tenth­
century calculations, is approximately 80 m. 39 Dividing by two 
to determine the width of each track and subtracting the width 

38 Bruns (supra n.1) passim, already criticized by Kollwitz (supra n.1) 427 
and Wrede (supra n.10) 192. 

39 Dagron 328; cf Casson (supra n.19) Plan II. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

424 THE THEODOSIAN OBELISK BASE 

NORTHWEST SOUTHEAST 

Key 

~ Theodosian Obelisk Base 

@ Serpent Column 

I8J Built Obelisk 

SOUTHWEST 

Figure 2: Diagram of the Hippodrome at Constantinople, showing 
dimensions of the arena and cavea, and relative distances between the 
base of Theodosian obelisk and spectators on all sides. Drawing by R. 
A. Fellerman. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 425 

of the obelisk base, we can deduce that the nearest spectators­
on the southeast and northwest-sat about 38.5 m. from the 
base at track level (Fig. 2 and PLATE 11). 

Spectators seated in the upper tiers of seats along those sides 
are obviously farther away. To calculate this distance we must 
know the depth of the seating area, which is the width of the 
preserved sphendone, approximately 22 m.,40 and the height of 
the seats. Robert of Clari, who saw the Hippodrome as a 
crusader in the early thirteenth century, attests thirty or forty 
rows of seats. 41 We may estimate a height of approximately 35 
cm. per row, based on the relatively intact and well-excavated 
circus of Lepcis Magna and the Circus of Maxentius at Rome 
(Humphrey 33, 591). Both of these were smaller than the Hip­
podrome at Constantinople, but this would affect the number 
of rows rather than the heights of individual tiers. The simple 
calculation a 2 + b2 =c 2 yields the distance of the seats farthest 
from the obelisk base at the center of the southeast and north­
west sides: these spectators were approximatcl y 61 m. from the 
base (Fig. 2). 

Can the reliefs and inscriptions be seen from a distance of 38.5 
m., let alone 61 m.? Investigation is hampered by the impos­
sibility of obtaining a clear view of the base and its plinth from 
those distances today; they are partly below the current ground 
level and further obscured by protective fences (PLATE 10). 
Nevertheless, field testing, proportional observations, and pro­
fessional manuals on visibility indicate that the reliefs would 
have been visible to fourth-century audiences on the southeast 
and northwest, and that these spectators should also have had 
little trouble reading the inscriptions. 42 By analogy with other 

40 Dagron 328; Casson (supra n.19) Plan II. 
41 Janin (supra n.l) 186; Dagron 327. 

42 Many physical and psychological factors can have an impact on visibility 
and perception, but I assume, arguendo, an 'ideal' viewer with normal 
eyesight: see J. Follis and D. Hammer, A rchitectural Signing and Graphics 
(New York 1979) ISf. To test the legibility of the inscriptions the following 
experiments were performed. (1) A series of smaller carved inscriptions at 
Dumbarton Oaks were surveyed at proportionately reduced distances. These 
were all readable, especially when sunlight produced strong contrasts of light 
and shade. (2) In Istanbul the first word of the Greek inscription was drawn 
to scale and read by several people without difficulty from 38.5 m.; the glare of 
the sun on white paper was the only problem. Illustrating this, however, is 
difficult: what was visible to the eye was small and far away in the camera 
lens, and this distortion is only increased by reproducing photographs or pro-



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

426 THE THEODOSIAN OBELISK BASE 

ancient relief sculpture, the reliefs on the obelisk base were 
certainly augmented with paint and possibly with gilding or 
metal attachments. Such coloristic treatment would render all of 
the carvings significantly more legible, and sculptural details are 
still visible today at 38.5 m. The inscription fields (PLATES 4, 5), 
in the form of a tabula ansata, measure about 100 cm. high and 
the letters, approximately 10 cm. tall, are separated enough to 
offer clear legibility at 38.5 m. In all likelihood these too were 
originally colored, making them more legible at even greater dis­
tances. Spectators seated in the upper tiers and those who were 
less literate could still see the epigrams and recognize their 
power and weight as inscriptions even if they could not read 
them. 

What about the more distant audiences, those in the sphen­
done seats and the charioteers and others in the carceres ? The 
average interior or arena length for Late Roman circuses is close 
to 450 m.43 The length of the Hippodrome in Constantinople is 
not certain because the carceres have not been excavated, but it 
was between 370 and 450 m.; M uller-Wiener gives a length of 
400 to 420 m.44 From the inner edge of the sphendone to the 
obelisk base is 242 m., according to the excavation plan of the 
1920s (ef PLATE 11). By other calculations the distance is less, 
but in any case too far for the inscriptions to be legible even if 
the sphendone audience were literate, which is doubtful. At 242 
m. the reliefs can still be seen, though not very well, and not by 

jecting slides. (3) Because many art historians prefer to see first-hand obser­
vation bolstered by reliance on a text, manuals on graphics standards used by 
architectural signage firms and by such organizations as the U.S. Department 
of Transportation were consulted. One such manual indicates that one-inch 
(2.54 em.) letters can be read from a distance of 50 ft. (15.24 m.); it seems 
reasonable to extrapolate that letters of 10 cm. should be readable from a 
distance of 61 m. (Follis 18, 21; I am grateful to George Sexton and Jerry 
Moore for discussion of this topic). Theoretically, then, the closest and most 
important spectators, those on the southeast and northwest, should have had 
little trouble reading the inscriptions. And their distance presented no 
problem at all in 'reading' the reliefs: even small sculptural details are clearly 
visible at 38.5 and 61 m. 

43 Roman circuses range in exterior length from 244 m. at Gerasa to the 620-
m.-Iong Circus Maximus: see Humphrey passim. 

H Dagron 328; W. Muller-Wiener, Bildlexikon zur Topographie Istanbuls 
(Tubingen 1977) 64. With conviction not borne out by evidence, Pasquato 
(supra n.28: 88) gives dimensions of 500 x 125 m. for the Hippodrome. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 427 

those whose VIew was blocked by other monuments on the 
barrier. 45 

Compensating somewhat for the distance from the sphen­
done and the carceres, the width of the obelisk base on the 
southwest and northeast is about 36 cm. greater than that of the 
two flanking sides. If this is deliberate-that is, if the discrepan­
cy does more than just echo the irregular shape of the lower 
edge of the obelisk-then it belongs to a long Roman tradition 
of optical compensation for distance. For example, Trajan's 
Column, erected in 113, has spiral relief bands 1.25 m. high at 
the top versus 0.89 m. at the bottom; on the column of Marcus 
Aurelius (ca 180) the spiral relief bands are all the same height, 
but the depth of carving is greater at the top. Both of these 
solutions help compensate for the increasing distance from the 
viewers below.46 

Another significant way in which the obelisk may have re­
lated to specific Hippodrome audiences is in the shadow it cast 
at different times. Races went on all day and throughout the 
year, but at certain times the obelisk shadow must have pointed 
to, or even fallen on, the emperor in the kathisma. This 
potential is evident in situ in the Hippodrome and its effect of 
pointing at the emperor can perhaps be connected with the 
possible use of the obelisk as a sundial, a function of obelisks in 
the Roman world since the time of the first Roman emperor. 
Augustus' obelisk in the Campus Martius was purportedly the 
gnomon of a sundial that pointed to the altar of the Ara Pacis on 
the emperor's birthday.47 If we knew the precise location of the 
kathisma in Constantinople, we might be able to determine the 
significant anniversary of Theodosius. 

45 Visibility from the sphen done in situ cannot be tested because of the 
presence of a modern edifice halfway between the obelisk and the end of the 
sphendone. Two hundred forty-two meters in the other direction, however, 
towards the original carceres, corresponds to a vantage point just past the 
fountain of William II (cf PLATE 11). 

46 Platner-Ashby 243; R. Bianchi-Bandinelli, Rom. Das Zentrum der Macht 
(Munich 1970) 240f. 

47 E. Buchner, "Solarium Augusti und Ara Pacis," RM 83 (1976) 319-65, and 
"Horologium Solarium Augusti, Vorbericht tiber die Ausgrabungen 1979/80," 
RM 87 (1980) 355-73. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

428 THE THEODOSIAN OBELISK BASE 

IV 

The particular placement of the obelisk base clearly shows an 
awareness of the social status, the literacy, and the distances of 
its audiences. But what did the programmers or patrons of the 
work want to communicate, and what did the audiences bring 
to it? Of course we shall never know the entire range of respon­
ses to a Late Antique work. Surely there were individual dif­
ferences, just as there would be today; an example might be the 
furor over the Louvre pyramid in Paris or the Vietnam 
memorial in Washington, both works that are highly praised by 
some and harshly criticized by others. All we can hope to do is 
generalize about patterns or middle ranges of response from 
each group of spectators. 

Jauss' formulation of reception aesthetics suggests that each 
social group has a distinct "horizon of expectations. "48 In the 
fourth century this horizon was affected by familiarity with 
sculpture in Constantinople in general and in the Hippodrome 
in particular, including imperial and non-imperial imagery. Thus 
all viewers were familiar with large, impressive monuments like 
the column of Constantine and the new monuments erected in 
Theodosius' Forum Tauri between 386 and 394. These no 
longer survive, but they included an equestrian statue and a 
column with spiraling reliefs. The Theodosian column, the first 
such monument in Constantinople, must have been a striking 
addition to the urban topography, whether or not observers 
recognized the Roman antecedents.49 In a similar way, the new 
obelisk in the Hippodrome would command particular atten­
tion as the first such 0 bj ect in the ci ty. 

48 H.-R. Jauss, Aesthetic Experience and Literary Hermeneutics, tr. M. 
Shaw (Minneapolis 1982), and Toward an Aesthetic of Reception, tr. T. Bahti 
(Minneapolis 1982); reiterated by M. Kemp, "Seeing and Signs. E. H. 
Gombrich in Retrospect," Art History 7 (1984) 239: "There is actually no way 
in practice that we can operate as artists or judge as spectators outside a 
context of expectation. It simply is not possible .... The artist and spectator 
inevitably work within a series of cultural contexts, in which theoretical, 
historical and institutional factots all impinge upon and help articulate our 
judgments." See also J. P. Tompkins, ed., Reader-Response Criticism 
(Baltimore 1980). 

49 On the column of Theodosius, soon supplemented by that of his son 
Arcadius, identical in conception and similar in iconography, see G. Becatti, 
La colonna coclide istoriata (Rome 1960) 83-150, and Sande (supra n.8). 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 429 

Sculpture was definitely a part of the everyday environment 
in Constantinople. 50 Relief sculpture may have been excep­
tional, however, as most works listed in the eighth-century Para­
staseis syntomoi chronikai were evidently freestanding. In the 
absence of any contemporary inventory, determining what was 
present in the fourth-century Hippodrome is impossible. The 
Serpent Column dates from Constantine's era, 51 but the date for 
erection of the cut-stone obelisk at the southern end of the 
barrier (PLATE 10) is unclear. 52 There were statues of Herakles, 
Augustus, and other deities and emperors,53 but all of them 
could not have been on the barrier. In general the Hippodrome 
became more crowded with sculpture as the Byzantine cen­
turies progressed; for example, the four bronze horses over the 
carceres installed under Theodosius II (probably the group that 
now dominates the entry to San Marco in Venice) were not a 
distraction when the obelisk and base were erected. 54 

At the beginning of the twentieth century, scientists studying 
how the visual world is perceived invented a means of photo­
graphing eye movement. 55 This technology showed that the eye 
goes to the center, then jumps about to fixate on areas of 
greatest information. We can imagine this unconscious process 
at work in the Hippodrome. Whether the Theodosian obelisk 
was at the center of the Hippodrome or the center of the 
barrier, it was a monument to catch the eye. From any seat in 
the circus its centrality was powerful: the vertical axis of the 
obelisk and base created a tension with the horizontal axes of 
the barrier and the circus. The power of the center would be 

50 See C. Mango, U Antique Statuary and the Byzantine Beholder," DOP 17 
(1963) 53-75. 

51 Euseb. V. C. 3.54.2; see now T. F. Madden, "The Serpent Column of 
Delphi in Constantinople: Placement, Purposes, and Mutilations," ByzMGrSt 
16 (1992) 111-45. 

52 Muller-Wiener (supra n.44: 65), Wrede (supra n.10: 187ff), and Dagron 
(324) suggest that it was put up by Constantius II or Theodosius 1. 

53 Cameron and Herrin (supra n.23) 48-51, 136£, 248ff. S. Guberti Bassett, 
"The Antiquities in the Hippodrome of Constantinople," DOP 45 (1991) 88, 
summarizes the antique statues in the Hippodrome as apotropaia, victory 
monuments, public figures, and images of Rome. 

54 Cameron and Herrin (supra n.n) 160f. 
55 R. Dodge, "Five Types of Eye Movement," American Journal of 

Physiology 8 (1902) 307-29; more recently, A. L. Yarbus, Eye Movements and 
Vision (New York 1967); K. T. Spoehr and S. W. Lehmkuhle, Visual 
Information Processing (New York 1982) esp. 163-66. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

430 THE THEODOSIAN OBELISK BASE 

reinforced by the well-known cosmological symbolism of the 
circus, in which the obelisk represents the sun-it was said to 
be shaped like a sunbeam-while the chariots, sporting four 
different colors representing the four seasons, race seven laps 
suggestive of the days of the week, and so on.56 The power of 
the center also operates on the obelisk base: on every side of 
the base the viewer's eye surely goes first to the emperor, who 
is centralized and framed by architecture. The center is em­
blematic of authority (PLATES 1,6-9).57 

Objects in the visual field that attract the most attention-in 
other words, the longest involuntary fixation of the eyes-are 
those that are the least expected. As is well known, repeated 
exposure to a stimulus renders it practically invisible and 
everyday things fail to be noticed. But spectators on their initial 
view of the obelisk base were probably surprised to see the 
emperor in the kathisma implicitly presiding at the games. In 
fact, no earlier examples of this subject survive (Dagron 311). 
This striking innovation imitates reality by integrating the 
spectators with the setting and fusing real and represented 
space. This fusion or elision of image and prototype underlies 
the power of images in Late Antiquity, when insulting a statue 
of the emperor was equivalent to insulting the emperor 
himself.58 Despite the novelties of form and presentation, much 
of the visual and textual information presented on the base 
would have seemed compellingly familiar to the Hippodrome 
audiences. 

The obelisk's indisputably imperial and triumphal character is 
reinforced on every side of the base. 59 Such tangible evidence of 
victory was an especially important aspect of imperial propa­
ganda after the Goths' crushing defeat of the Romans at Adri­
anople in 378. As McCormick has shown, six triumphs were 

56 Humphrey 269ff; Dagron 330ff. This symbolism was especially well 
developed in the sixth century: P. Wuilleumier, "Cirque et astrologie," 
MelRome 44 (1927) 184-209; A. Cameron, "Corippus' Poem on Justin II: A 
Terminus of Antique Art?" AnnPisa 5 (1975) 156 with n.D8. 

57 R. Arnheim, The Power of the Center (Berkeley 1988). 
58 Basil. In Isaiam 13, Migne, PC XXX 589A-B; E. Kitzinger, "The Cult of 

Images in the Age before Iconoclasm," DOP 8 (1954) 83-150 (=id., The Art of 
Byzantium and the Medieval West [Bloomington 1976] 90-156); see also D. 
Freedberg, The Power of Images. Studies in the History and Theory of 
Response (Chicago 1989) 318ff; 392. 

59 See esp. S. MacCormack, Art and Ceremony in Late Antiquity (Berkeley 
1981) 56-67; BaIty (supra n.l0); Wrede (supra n.10). 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 431 

celebrated in the ten years after Adrianople, compared with less 
than two per decade in the preceding seventy years. Even 
minor victories, like that of Theodosius over the unnamed 
tyrants of the Latin inscription, were magnified to mitigate the 
Romans' psychological devastation. 60 

Most viewers were probably unaware of the Platonic and 
Aristotelian view that artistic works were supposed to inspire 
certain reactions and behaviors, but the principle of mimesis 
was certainly operating in the obelisk base. Looking from the 
kathisma at the southeast side (PLATES 1-2, 8), the emperor saw 
himself immortalized as victorious, regardless of whether he is 
giving or receiving the crown of victory.61 He saw before him 
representatives of all social classes from government officials to 
dancers; his courtiers also saw him and recognized themselves. 
The emperor could read the Latin inscription (PLATE 4) that 
names him and stresses his conquest and mastery of the tyrants 
and the obelisk: both had formerly been obstinate, but the one 
has now fallen and the other was successfully and very visibly 
raised aloft. In addition, the emperor saw the best-quality 
carving available; perhaps he appreciated the different stylistic 
'modes' employed for the different classes of figures. The 
'front' of the obelisk with its original dedication to Pharaoh 
Thutmosis III-probably not by accident-faced its new 
master on the southeast: the most important side of the 
monument in 1471 B.C. retained its significance in the 390s.62 

On the northwest side (PLATES 3, 9), the circus partisans and 
militia would have seen the emperor to whom they owed 
loyalty. In both life and art he looks out at them, a direct 
communication between master and subjects. Eye contact 
between figures in a work of art and its spectators creates an 
important bond; indeed Arnheim (supra n.57: 44) posits a kind 
of visual energy emanating from the work to the viewer. The 
spectators on the northwest saw the representatives of the 
emperor's orderly government and subjugated barbarians from 

60 M. McCormick, Eternal Victory (Cambridge 1986) 41H. 
61 This debate is summarized in MacCormack (supra n.59) 299 n. 220. 
62 To what degree the hieroglyphs were comprehensible to a fourth-century 

audience is uncertain. Some recognition or at least interest appears in Amm. 
Marc. 17.4, who provides a Greek translation of the hieroglyphs of Con­
stantius' obelisk in Rome. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

432 TI-IE THEODOSIAN OBELISK BASE 

all over the world as evidence of his power. 63 Perhaps some 
spectators had fought such foreigners and would share a sense 
of successful involvement that could mitigate the still vivid 
memory of Adrianople. Others might recognize the message of 
submission to authority. In addition, many could both see and 
read the Greek inscription (PLATE 5), in which references to 
the tyrants were probably omitted because the western 
usurpers had never touched the lives of most Constantino­
politans. Instead the inscription encourages a focus on dig­
nitaries present in the Hippodrome and responsible for the 
erection of its largest and most distinctive monument. The 
specific, almost didactic, information in the epigram identifies 
this enormous four-sided column as the one raised in precisely 
thirty-two days. Even if the circus partisans were no more 
literate than their fellow citizens in the sphendone, they would 
recognize the power of the written word coupled with the self­
evident power of the emperor. 

The repetitive figures of government functionaries on the 
southwest (PLATES 2, 6) and northeast (3, 7) sides also com­
municated triumph, stability, and order. Orderliness was hardly 
the impression of the Hippodrome given by contemporary 
critics like John Chrysostom and Amphilochius of Iconium, 
who wrote vicious diatribes against the circus. 64 Although these 
sermons perhaps kept some potential spectators away, circus­
goers probably remained quite unconcerned about the effects 
of races on their souls. 

The distant viewers in the sphendone were brought into 
closer connection with the events before them because the 
silentiaries depicted on the southwest stairs seem to address 
their remarks directly to these spectators (PLATE 6). Those on 
the southwest with very good eyesight may have appreciated 
the lively narrative quality of the reliefs on the plinth (2) and 
their realistic and familiar scenes. Indeed, circus workers in the 
carceres probably helped raise the obelisk, as depicted on the 
northeast plinth (3); viewers on the southwest saw a chariot 
race, echoing the real ones unfolding regularly before their 

63 H. Gabelmann, Antike Audienz- und Tribunalszenen (Darmstadt 1984) 
20M. 

64 E.g. J. Chr. Contra circenses, ludos et theatra, Migne, PC LVI 263ff; J. 
Bareillc, tr., Oeuvres completes de S. Jean Chrysostom X (Paris 1867) 484-94; 
Amphilochius of Iconium, Iambi ad Seleucam (=E. Oberg, cd., Patristische 
Texte und Studien 9 [Berlin 1969]) lines 150-80. 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 433 

eyes. It is unclear whether the two obelisks seen on the barrier 
in the racing scene represent the two now visible in the Hippo­
drome-the one supported by the base and the cut-stone ver­
sion at the southern end (10-11 )-or whether the artist merely 
reproduced from some model the disposition of two obelisks 
in the Circus Maximus at Rome. The nuances of the higher­
quality sculpture found on the flanking sides of the base would 
have been lost on the sphendone audience because of the 
distance of these seats. 

v 

The Theodosian obelisk base is a fully realized work of public 
art and all evidence points to the careful placement of the base 
to correlate with its expected viewership. The programmers of 
the base were aware of both the general context for the work, 
the Hippodrome, and the specific audiences on each side. The 
commemorative inscriptions, and possibly the narrative reliefs 
on the plinth, were carved just after the obelisk was erected. 
Differences of the Greek and Latin inscriptions in content and 
character reinforce the likelihood that different audiences saw 
each side and were not disturbed by discrepancies they could 
not see. The discrepancy in the number of days needed to raise 
the obelisk is troubling to modern viewers who look at the base 
as a four-sided unit available for examination at close range, 
whereas the original audiences could never compare the two 
inscriptions. The emperor and his retinue always watched from 
the southeast side, the partisans from the northwest; they never 
exchanged seats. 

An interest in audiences is evident in fourth-century and later 
mosaic pavements that depict the circus (Humphrey 246). 
Whereas earlier representations show only the race and the 
barrier in any detail, in later works the seating areas and 
especially the spectators became legitimate artistic subjects, 
beginning with the Piazza Armerina mosaic in the first half of 
the fourth century and continuing through the example at Gafsa 
in the late fifth or early sixth. One wonders whether the Late 
Antique artists' interest in the circus audience parallels the 
interest in the audience evinced by the planners of the obelisk 
base. Similarly, the awareness of and exploitation of different 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

434 THE THEODOSIAN OBELISK BASE 

vantage points is a general characteristic of Late Antique mosaic 
pavements. 65 

The meaningful placement of large-scale sculpture, clearly im­
portant to the planners of the obelisk base, has some precedents 
in earlier Roman art. The Arch of Trajan at Benevento has 
reliefs relating to town and country facing the appropriate sides, 
and the Arch of Constantine in Rome shows the sun rising on 
the east and the moon setting on the west, providing a cosmic 
setting for Constantine's actions. The obelisk base in Constan­
tinople takes these precedents further by exploiting the multi­
faceted context of the Hippodrome in a more particularized 
way, making the audiences on each side responsible in some 
measure for the content. Later audiences would have very 
different "horizons of expectation" in looking at the obelisk 
base, because Hippodrome spectacle and ceremonial changed 
over the centuries, and so did the quantity and types of 
sculpture in the city. These later responses need to be 
distinguished from those of the original viewers. 

Such public art as the obelisk base served a variety of func­
tions. It provided a place and a prop for imperial messages; it 
recorded events; and it worked as a mimetic agent, capable of 
recreating events in the minds of viewers.66 Although it cannot 
be proved, a panegyric delivered at the unveiling, like that 
presented to Theodosius at Rome in 389, probably further 
mediated the messages on the obelisk base. 67 The erection of 
the obelisk seems a likely occasion for a panegyric. 

The obelisk base was meant to be seen, and could be seen, 
from a number of distinct points of view. This fact has im­
portant implications for the visibility and the interpretation of 
other large public monuments of the Late Roman and Byzan­
tine eras. Art historians and other scholars err in looking at 
sculptural details divorced from their spatial, temporal, and 
social contexts. A consideration of the viewpoints of the 

65 See C. Dauphin, "Mosaic Pavements as an Index of Prosperity and 
Fashion," Le'1Jant 12 (1980) 112-34. 

66 These remarks on function are derived from C. Kondoleon and B. 
Bergmann's introduction to the session on "Ancient and Medieval Spectacle," 
College Art Association Conference, Seattle, February, 1993. 

67 See Nixon (supra n.5). 



SAFRAN, LINDA, Points of View: The Theodosian Obelisk Base in Context , Greek, Roman and 
Byzantine Studies, 34:4 (1993:Winter) p.409 

LINDA SAFRAN 435 

original audiences can help inform our own and bring new 
meaning to familiar works of art. 68 

CATI-IOLIC UNIVERSITY OF AMERICA 

October, 1993 

68 This paper has benefitted from critical comments by Adam S. Cohen, 
Lioba Theis, and Joachim Wolschke-Bulmahn. I am grateful to Lee F. Sherry 
for discussion of the inscriptions and assistance with their translation and to 
the Director and staff of the German Archaeological Institute in Istanbul for 
their kind support. Earlier versions of this paper were read at the Eighteenth 
Annual Byzantine Studies Conference, Urbana-Champaign, the Kunsthis­
torisches Institut, Universitat Bonn, and the Catholic University of America.