Reflections on the sources, distribution
and uses of water in North African churches.
Susan T. Stevens
Introduction
This paper approaches the physical evidence for the practice of baptism in North Africa from a hydraulic standpoint. Water supply and distribution is currently a hot topic in Roman archaeology, and I am curious about the variety of water use in churches reflected in fonts, fountains, basins, wells and cisterns. [1] I should say at the outset that much crucial information is not available because the 19th c. (and some more recent) excavators were interested primarily in the superficial (as opposed to the underground).[2] Nevertheless, I wondered, whether the shape of receptacles and their supply and evacuation could assist in telling sacramental from non sacramental water. And in defining this difference in treatment of water, highlight the very special power of baptism. The paper also represents the very particular point of view of an excavator of ecclesiastical sites at Carthage, striving to interpret sometimes ambiguous physical evidence in the light of the late Roman urban as well as Christian context. My discussion, as a consequence radiates from Carthage. I am field archaeologist with the salutary experience behind her, of having rushed to see a baptistery where one may not have existed.
The Baptismal Fonts of Carthage.
There are three unequivocal baptismal fonts in Carthage, at Carthagenna (the church which lies just south of the modern Supermarché at Dermech, located between decumanus 2 and 3 south and between kardo 9 and 10 east) [3], at Dermech 1/Douïmes (long presumed to be the episcopal basilica of Carthage which lies near the Antonine Baths between decumanus 4 and 5 north and kardines 13 and 14 east)[4] and at Damous el Karita, the great pilgrimage and cemetery basilica 150m outside the main north gate of the city.[5]
At Carthagenna, the baptistery is a 14.25 m. square building, octagonal on the interior (straight alternating with apsidal walls), attached to the basilica's SW side and belongs to the Byzantine phase of the church which seems to have been abandoned in the mid 7th c. In the center of the room was an octagonal colonnade supporting a ciborium or cupola over a cruciform font which was 4m from end to end. Each arm of the cross (1.25 m, oriented NS x EW) had five steps down into an octagonal basin 1.5 m. across, creating a well just over a meter deep. Both the octagonal surround supported an octagonal ciborium and the font itself were paved in white marble. The ambulatory floor had 8 mosaic panels of vegetal and geometric motifs and shell motifs in the four niches. We have no information about the water source for the font, although many cisterns were found in the adjacent ecclesiastical complex and the church. The font is presumed to have drained into the street drain of Cardo X east a few meters away.
The baptistery of the Dermech I basilica is a 12.5 m. N-S x 10.25 E-W rectangular room attached to the north flank of the basilica at the end opposite the main apse. Its plan has been variously shown. Two rows of 4 columns form a central nave 6.5m x 10 m. Sadoux' plan has a square colonnade of 12 columns envisioning an outer, then an inner ambulatory. In the center of this area is a square surround 3m on a side with columns at the four corners supporting a baldachin/ciborium. The font itself is a hexagon 1.5m side to side which steps down into a round well 60 cm in diameter. Sadoux shows two sets of three steps down cut into the square surround on the NE and SE sides of the hexagon which are not visible today. Miller's plan has a round surround with a hexagonal font at the intermediate level and round at the bottom. Depth of the font at its center is about a meter. A cistern under the floor of the north aisle ambulatory supplied t the font with water through the northwest side of the hexagon and the water ran out through a drain at the bottom of the font in the middle of the W side of the hexagon, to be collected in a reservoir under the font, perhaps evacuated to the street drain under kardo 13 E. No date is settled for this structure although the Tunisian-American team suggest that the baptistery was an addition to the complex dating to the sixth century.
At Damous el Karita the baptistery was inserted into a previously constructed hypostyle hall 35.75 x 24.55 m behind the primary apse of the first basilica dated to the late 4th c., which, according to Ennabli, was oriented NE-SW. However, the suite of two small rooms is oriented SE-NW, that is, according to Ennabli, with the secondary orientation of the basilica. The baptistery is an approximately 4 x 4 m. room with columns of a ciborium in its four corners. The hexagonal surround is about 3 m side to side with a hexagonal basin within about 2.5 m side to side. The interior well, accessible by 2 sets of stairs of three steps is approximately 2m in diameter. The whole font was revetted in marble, no information was available for sources of supply or evacuation.
These three baptismal fonts have easily identifiable common characteristics that they share with other fonts of approximately the same period from elsewhere in North Africa. Cite examples. First, all three are all Byzantine in date. Parenthetically, the dating of churches at Carthage is sketchy, based more on historic watersheds (Arian Vandals 439, Orthodox Byzantines 533) rather than the internal evidence of individual sites. One is struck however by the concidence of this general dating scheme with PA Fevrier's assertion of the change in religious mentality that produced the flowering of baptisteries in the second half of the 5th c. which culminated 100 yrs later. [6] Second, the fonts are the focal point of a room, and are ceremonialized by an anteroom, an ambulatory and/or ciborium/baldachin. Third, the fonts are stepped basins, recessed, excavated into the floor, with a specific corridor for descent and ascent. The lowest level of the font is a round or hexagonal well 2m-60cm. in diameter, the lowest level of the font. This part is called, at least in one instance the alveus fontis (hollow, cavity, deep vessel). [7] I am persuaded that the overall depth (or indeed the amount of water the font actually contained) is of less significance overall than the descent into this well. Fourth, although the evidence is circumstantial at Carthage in two of the three cases, the water source for the font was a cistern. All three fonts drained out through a small evacuation hole at the base of the tank.
Three Basins and a Well.
Are these characteristics of the Carthage baptismal fonts definitive in the absence of a clear ecclesiastical context or identifiable iconography? Do these characteristics help in interpreting the less clear-cut cases at Carthage? those basins which lack some or all of these characteristics suggested above: at Sayda[8] (underground), Bir Messaouda (supermarché),[9] Bir el Knissia,[10] and Bir Ftouha (Delattre).[11] It is not good enough, I think to say that this is no way to do an excavation because often we must make do with what what we have. Explaining all such instances away as remnants of earlier structures does not explain their incorporation into Christian practice.
The interpretation of the "underground baptistery" at Sayda is difficult because it has no clear ecclesiatical context. The font was inserted into an underground room separated from another larger room by a balustrade. The construction of the rooms seem to date to the second half of the 5th c. and were in use until at least the reign of Heraclius. The room itself is only a 1.40 m. sq. and paved with mosaic with a motif of open roses that goes up over the edge of the font surround. The surround is basically a square with beveled corners in the NE, SW, NW and SE. 12 carefully cut stone placques form the surround which supports the four columns of the ciborium. The basin itself which is quadrilobed with an alveoles in each corner of the square surround. Two steps down in the NW alveole lead into a squarish-circular basin about 80 cm in diameter. The bottom of this circle is a monolithic marble slab 72 x 59 cm in the center of which was a bung hole 25 cm in diameter. The depth of the font from the surround to the bung hole is just under 1.5 m, even deeper that the Carthagenna baptismal font but very tight at the bottom. The most interesting, but perhaps troubling aspect of this baptistery is that the well evidently served both as the source and the waste weir. Except for the last the Sayda well has all the characteristics of the Carthage baptismal fonts. In fact its excavators interpreted it as a baptistery built by orthodox Christians escaping Arian persecution in a place chosen because of the well, but it may be just a monumentalized well sacralized by the painting of Saturus, the martyr associated with Perpetua and Felicitas.[12]
On the other hand, Duval does not hesitate to discuss the basin at Bir Messaouda (between decumanus 1 and 2 south and kardo 9 and 10 E) as a baptismal font. Although it has no ecclesiatical or even Christian context to support such an interpretation. Indeed, it seems to fundamentally different from the Carthage baptismal fonts discussed above. It was built in two phases. The hexagonal surround is 3.80 m. EW x 4.20 m. N-S. The hexagonal basin of phase 1 was revetted with marble and was just over 2m from side to side. It was 50 cm deep and had a 60 cm wide evacuation channel on the SW side headed for Kardo 9. In phase 2, also revetted in marble, a single step 33 cm wide was cut into the hexagon, the drain was narrowed and the basin shape became a circle 32 cm deep and 1.57m in diameter. We have no idea what building it may be attached to, though a basilica is presumed (Rakob has since removed his basilica as a candidate). The basin is of similar size to that at Carthagenna, but it is clearly very shallow as the single step down indicates. We do not know the floor level or the configuration of the surrounding structure. The basin was originally investigated by WHC Frend and recently a British archaeologist working in the field despaired of understanding its date and context. Duval supports the notion of a baptismal font on, it seems to me, rather tenuous grounds [13]Not likely to be published again any time soon. Why assume that it is a baptismal font, what about a fountain basin or piscina in a private house?
At Bir Ftouha in the area around the modern well, about 50 m west of current excavations, initially interpreted as a baptistery by Delattre because of the remains of a basin and clear evidence of Christian occupation of the area, presence of a basilica to the E (now being excavated). Since there were actually basins and an associated hypocaust, the complex was almost certainly a bath, at least originally contructed as such.[14] The larger basin is unusual in form and has a marked drainage system associated with it.[15] Nevertheless there is a clear Christian connection in the later history of the complex because of the triconch funerary chapel (akin to that attached to the semi-circular atrium at Damous el Karita). The doe and stag motif of a mosaic found in the E end of the field may indeed have something to do with a baptistery, but certainly is not related to the basins Delattre excavated from which it is separated by some 50 m.
Ennabli makes a tenuous link between Delattre's "baptistery" at Bir Ftouha and the poem of Calbulus, a sixth century grammarian and author of a poem-inscription in the Codex Salmasianus of the Latin Anthology.[16] Without the connection to Bir Ftouha, the poem is certainly food for thought. Around the outside of the font are the words: "Marmoris oblati speciem, nova munera, supplex/Calbulus exhibuit, fontis memor, unde renatus,/per(ad) formam cervi gremium perduxit aquarum." Otherwise the poem is divided into four pieces, designated as a parte episcopi, descensio fontis, ascensio fontis, econtra episcopum [17] suggesting a cruciform font. Along the same lines Ennodius of Pavia described the water gushing in jets from the top of every column with which font was surrounded at the 6th c. church of St Stephan in Milan, and seven silver stags poured water through their mouths into the basin.[18]
At Bir el Knissia a basin in a western paved courtyard 7 x 7 m. surrounded by a peristyle which seems to have been used for burial. The basin is round, 1.9m in diameter, both sides and bottom revetted in marble. Seems to have had a lip raised 40 cm above the surrounding floor? Water circulation is suggested : a channel lead out of the basin to the SW which was about 60 cm wide. A possible source is a lead pipe 11 cm in diameter issuing through the wall of the small apse of the 'galerie exterieure'into a small channel. The basin is not stepped, its high lip means there was no intent to get into it. Although there is ambulatory space, no sign of a ciborium associated with the basin. I interpreted this space as an ablutions basin.[19] Duval says only that this is obviously the remainder of an earlier building.[20]
These three basins have very little in common except for their large evacuation channels which imply a continuous water flow. The three basins also lack the ciborium, ambulatory or any other indication of ceremonial use. While the Sayda well is a unique case, the basins at Bir Messaouda and Bir Ftouha seem, from their form, to be Roman bath basins or public water basins that may or may not have been incorporated into a later Christian context.
The most striking basin of the group is that at Bir el Knissia. While it is attached to a basilica, it is not the stepped basin with a central well recessed in the floor that seems characteristic of Carthage baptismal fonts. It seems rather to be an ablutions or fountain basin near an entrance to the basilica. In fact there is a good parallel to the Bir el Knissia ablutions basin in Carthage, the octagonal basin or fountain in the middle of the hemicycle atrium at Damous el Karita. This structure is 4m on exterior, 3 m interior, approximately. The excavators offered no discussion of its original depth or provisions in the area for supply and run off. One might mention in connection with the Bir el Knissia and Damous el Karita basins, the inscription placed over a fountain Leo I (440-461) set up in the atrium of St. Paul outside the walls: "water removes dirt from the body; but faith, purer than any spring, cleanses sin and washes souls. You who enter as a suppliant at the shrine of St Paul, wash your hands in the fountain."[21]
Such fountains might have been routine for ablutions before the faithful entered a specific kind of churches, suburban cemetery or pilgrimage complexes.[22] There is some evidence that ablutions basins may have been a regular situated in atria. While atria were not common in North African churches, a case can be made that three of five suburban churches at Carthage had atria. [23] The atria of Damous el Karita, Ste. Monique and Bir el Knissia were full of tombs of the ordinary faithful attracted by the tombs of martyrs and saints. Atria would have been the ideal location for pilgrims to perform ablutions in the interest of ritual purity before entering the church.
The Antecedents and Setting of Ablutions Basins and Baptismal Fonts
The square and rectangular basins, lacus et salientes,[24] were the mainstay of the Roman public and private water distribution system. Their simple functional forms of changed very little over time. The circular and octagonal basins at Bir el Knissia and Damous el Karita seem to me a logical extension of these forms in a circulatory space, in the forecourts of basilicas, a new Christian space. They must have served a similar function to that of their Roman counterparts inside the city.
The setting of baptism is fundamentally different from routine ablutions and the design and setting of baptismal fonts reflects the unique character of the Christian ritual bath. Just as parts of the ritual of baptism, nudity, anointing with oil after the bath, the donning of new clothes, for example, reflect routine bathing practices, the physical setting of the ritual adopted the forms of Roman baths. The antecedents of baptismal fonts in Roman bathing complexes seem to me very important because they provide a cultural sounding board for specifically Christian forms (both the setting and the font itself). In a Roman world almost universally permeated with a sense of religiosity, baths and bathing belonged securely to the secular sphere. Furthermore, the forms of bathing rooms and bath basins seem to develop within functional domestic architecture.[25] The piscinae and lavacra (among the most common words used in the West for the baptismal font) of Roman balnea were variously shaped round, hemicycle, oblong, square, polygonal, and sometimes lobed.[26] Basins were usually excavated into the floor (or set in the raised floor of the hypocaust) and were stepped on one side. Bathers typically walked into the basin and sat on the steps described by one author "as a seat and pillow." The water was typically 50-75 cm. deep. Smaller piscinae, even round ones, were typically located at the ends of a larger rooms.
Ironically, baptisterium, the source of our word baptistery, is used twice by Pliny[27] to describe the plunge pools of a Roman bath but not by Christian Latin authors to refer to a baptistery.[28] The concept of an independent building as the locus of baptism seems more or less foreign to Africa perhaps because the font was perceived as part of a larger, almost seamless ritual space. Nevertheless baptismal fonts were emphatically centralized, they were the focus of the rooms in which they were located and the fonts were set off by a surrounding ceremonial space.
The elaborate and variously-shaped rooms of Roman balneae are also reflected. The octagonal room at Carthagenna seems adapted from caldaria, the triconch chapels at Damous el Karita, BF and elsewhere in North Africa may certainly derive from the more elaborate tri-lobed rooms and pools of baths.[29] There are certainly cases where the coopting of Roman bathing forms is direct rather than indirect, when the Roman balnea at Cimiez was converted into a baptistery for example.[30] Although I do not want to make too much of a tenuous connection, Roman baths were ideal spaces for conversion into churches, for example Maktar Basilica IV.[31]The connection is also thought-provoking because John Chrysostom is said to have allowed his followers to perform baptisms in a major bath at Constantinople.[32] Could this be where baptisms took place before the 5th and 6th century fonts were in use at Carthage.
Conclusion
While Christians may consciously have adapted and in some respects inverted the Roman secular tradition of the baths, in one crucial respect, the baptismal tradition is fundamentally different. The root of the difference is the designation of the baptismal water as fons, a reference to a fountain or source, (a sealed (secret) spring, a living well.), the word often refers not to the receptacle but to the water itself. For Christians the well of the font was death and the ascent out of it a resurrection into a new life. Thus the form of baptisteries and baptismal fonts reflect how Christians transformed the social and secular act of bathing into a religious drama of death and rebirth.
Susan T. Stevens, Randolph-Macon Woman's
College
1. R.J.A. Wilson, "Recent studies on Roman
aqueducts and water supply," JRA 9 (1996) pp?, A. Wilson, "Running water
and social status in North Africa," M. Horton and T. Weidemann edd., N
African from Antiquity to Islam. Papers of a conference held at Bristol,
October 1994 (Bristol 1995) 52-56. R. Tolle-Kastenbein, Antike Wasserkultur
(Munich 1990)
2. The identification of the underground
rotunda at Damous el Karita has waffled back and forth since its discovery
between baptistery and martyrium without ever considering water systems.
What is at its center? A well, a basin? Well? Basin? Or something else?
2m interior, 3m exterior. Raised or with balustrade around it as reconstructed
by Boyadiev in the center. The underground rotunda now believed to be a
martyrium according to a recent study by Dolenz --certainly not a baptistery
as Lezine first pronounced. Obvious large water channels but where leading
to or from?
3. This basilica, identified with the basilica
Restituta, is not yet fully published, L. Ennabli, Carthage. Une metropole
chrétienne du IV à la fin du Viie siècle (Paris
1997) 62-70 and the baptistery not at all.
4. Ennabli 77-79 . For the new work undertaken
at the site, Alexander, Ben Abed -Ben Khader, Metraux "The Corpus of the
Mosaics of Tunisia: Carthage Project 1992-1994," DOP 50 (1996) 361-8,
esp. fig. 17. Which differs from Sadoux' plan and that of Vaultrin.
5. Ennabli, 121-7, identifies this church
with the basilica Fausti, Vaultrin more helpful here than Duval.
6. P.A. Fevrier, "Baptistere et ville,"
Melanges Mano-Zissis, Sbornik Narodnoi Muzeja, Belgrade 1975 211-220
7. Victor of Vita III,v,9,34 (CSEL VII
p.89) at Basilica Fausti, an Elpidoforus baptized by deacon Muritta "susceperat
de alveo fontis.
8. Duval 1959, Baptistere souterraine
9. Duval REA 1988, WHC Frend in CEDAC
10. Stevens 1993
11. Delattre 1880
12. One should perhaps take into consideration
that there is at least one other non-Christian underground buildings in
the immediate vicinity, Kobbat bent el Rey, perhaps the meeting place of
a sodality built in the 320-340 6 m. below contemporary ground level. Two
small rooms (3 x 4.5 m) of the complex were reused twin cisterns of an
earlier period. Semi-circular fountain about 1 m in diameter with play
of water--run off under the floor, no info about water source.
13. Duval, REA 34(1988)
14. Rossiter Echos du monde Classique
15. Ennabli fig. 87
16. No clear evidence links this Calbulus
to Carthage, although the collection includes poems of the Vandal court.
17. AL 373,Ennabli p. 37. Cf this with
Ennodius of Pavia's description of font with the heads of stags pouring
water into the basin, below p.
18. Milburn 206
19. BK 33-4, 38
20. Duval review of BK in REA
21. Fons, (French dictionary of antiquities)
22. Or perhaps reserved for the use of
the clergy. Perhaps we should also take another a look at the "cuve" at
Haidra II in the "martyrial apse" in this connection. Duval 2 Absides,
204. It is discarded because it does not fit the baptismal pattern.
23. Also noteworthy is the cistern/well
head at Ste Monique which was the center of that complex' atrium, though
it is unclear how this may have functioned.
24. Frontinus (Harry B. Evans)
25. Yegul p. 54, 124
26. E.g. Sabratha's and Banasa's baths
have both hexagonal and lobed hexagonal pools, Yegul p. 238, fig. 284,
fig. 288
27. Pliny Ep. 5,6,25; 2.17,11 also Sid.
Ep. 2.2
28. Sid. Ep 4,15
29. Bulla Regia, Yegul fig. 291
30. Yegul fig. 79
31. Duval, 2 Absides p.145 fig. 83
32. Palladius, Dialogus 9, Socrates, Eccl. History 6,18 , quoted in Yegul p.315