The portal of the medieval church was uniquely symbolic. Jesus's words provided the key to it: "I am the door. By me, if any man enter in, he shall be saved" (John 10:9).1 This figure of speech became one basis for the allegorization of the church that was current in the earlier Middle Ages. The Venerable Bede attempted a synthesis of the Alexandrian tradition of allegorical exegesis2 and Augustine's theory of signs: "The temple of the Lord in the literal sense is the house which Solomon built," he wrote; "allegorically, it is the Lord's body...or his Church...; tropologically, it is each of the faithful...; anagogically, it is the joys of the heavenly mansion...."3 It is my contention that the art and symbolism of the Romanesque portal reflect a widely held assumption on the part of artists, patrons, and worshipers that all four senses, the literal, the allegorical, the tropological, and the anagogical, adhered essentially to any church building that they were "literally" aspects of its reality. The symbolism of the portal was not merely conventional; the portal was in a real sense Christ and the entrance to heaven.

    In the twelfth century church portals and their sculptural programs were frequently inscribed with verses. When the inscription spoke in the voice of the door or the building or a sculpted image above or beside the door, it employed the rhetorical figure of prosopopoeia or personification. Far from being naive, as at first sight it may seem, the use of prosopopoeia reflects the allegorization of the church in a sophisticated way. In this paper, after passing in review a selected group of Romanesque portal inscriptions in verse, most of them voiced in the first person, I inquire into their status as speech acts, and conclude that some of them were probably intended and received as "performative utterances." This conclusion rests on the argument stated above that the allegorization of the church was understood in a "literal" sense. I also suggest that Abbot

    Portal inscriptions often directed the faithful to adopt a penitential state of mind before entering the church. The lintel of the Royal Portal (the eastern nave portal of the south facade) of the Cathedral of S. Geminiano of Modena (Emilia-Romagna) is inscribed with this full leonine hexameter verse:4 "Hinc vos pergentes cum corpore flectite mentes" (As you proceed from here, bow down your mind with your body).5 A similar sentiment is recorded at the former abbey church of Ste.-Marie of Cassan, near the village of Roujan (Hérault) in southern France. A square marble plaque has been placed in the center of an otherwise blank tympanum above the north portal of Cassan, which is now walled up. Three common leonine hexameters arranged in six lines are carved on the plaque. The verses read:

Quisquis homo sceleris funesti mole gravaris

Pre foribus Domino merens prosternere summo,

Haud secus intratur quia ianua Christus habetur.

(Whoever you are a man with a burden of deadly sin who, sorrowing before these doors, hesitate to prostrate yourself to the supreme Lord, not otherwise is the gate entered because it is considered to be Christ.)

    When the portal spoke to medieval Christians who approached the church with the intention of entering, it spoke as Christ or as the Church. The voice of the portal was often expressed in the subjunctive or imperative mood. "Enter," "do not enter," "change your ways," etc. Dante's imitation at the beginning of Canto 3 of the Inferno is exact, although demonically inverted. On the gates of hell over the portal are inscribed the awful words:

Per me si va ne la città dolente,

Per me si va ne l'etterno dolore,

Per me si va tra la perduta gente....

Dinanzi a me non fuor cose create

Se non etterne, e io etterno duro.

Lasciate ogne speranza, voi ch'intrate.6 (Inferno 3, 1-3; 7-9)

(Through me you enter the woeful city,

Through me you enter eternal grief,

Through me you enter among the lost....

Before me nothing was created

If not eternal, and eternal I endure.

Abandon every hope, you who enter.)

Each detail the voice in the first person, the repeated phrase per me, the use of the imperative in the last line, the verb intrate--shows how thoroughly Dante had absorbed the convention. This is the voice of hell.

    If an imperative utterance of this kind were to be spoken by a living person in an appropriate social context, it would be, to use J. L. Austin's term, a "performative." Austin makes two observations about performatives which are pertinent here. He says, "a performative utterance will . . . be in a peculiar way hollow or void if said by an actor on the stage, or if introduced in a poem, or spoken in soliloquy."7 And he also says, "Unless a certain effect is achieved, the illocutionary act will not have been happily, successfully performed. . . . I cannot be said to have warned an audience unless it hears what I say and takes what I say in a certain sense."8 These comments focus the problem precisely. I want to consider whether certain inscriptions on Romanesque portals might be considered to be genuine performatives (as opposed to mimetic representations of performatives) and, if so, in what sense the directives of the portals might called successful performative utterances.9

    The Last Judgment tympanum of Ste.-Foy of Conques (Aveyron) crowns a double portal in which the two doors are separated by a massive central pier. On the lower edge of the lintel beneath the Heavenly City and the Weighing of the Souls, a couplet consisting of two full leonine hexameters extends in a single line from the left edge to a point approximately two-thirds of the way across to the right. It addresses the faithful in an imperative voice:

O peccatores, transmutetis nisi mores,

Iudicium durum vobis scitote futurum.

(Sinners, if you do not change your ways, know that a hard judgment will be upon you.)

This portion of the lintel is above the left door. The inscription ends directly beneath the gates of hell; it does not extend over the right door above which hell is depicted. Although it differs from some other portal inscriptions in not explicitly referring to itself as a gate or to the act of entering the church, its position above the door and its use of the vocative and imperative associates it with the type I am describing. It is the voice of the portal. From its location, I infer that medieval Christians entered by the left door, which was, therefore, the earthly analogue of the gate of heavenly city just above it. Could they have even imagined passing through the right door beneath hell?10

    Whereas at Conques the portal addressed sinners collectively, the liminal inscriptions of the Cathedral of Notre-Dame of Le Puy (Haute-Loire) spoke to the individual sinner. On the risers of the grand stairway leading up to the "Golden Portal" of the west facade there is inscribed this full leonine elegiac couplet:

Ni caveas crimen caveas contingere limen.

Nam Regina poli vult sin[e sorde c]oli.11

(Unless you avoid impurity, avoid touching this threshold, because the Queen of heaven wishes to be worshiped without sordidness.)

A similar couplet is known to have been placed by the south transept portal, which read:

*Lubrica si vita fuerit, tunc limina vita,

Sanctaque ne violes dum mala carnis oles.12

(If your life has been lecherous, then avoid this threshold, and do not profane its sanctity as long as you smell of the evils of the flesh.)

At Conques, where the torments of hell are graphically depicted on the tympanum, the warning has to do with the life to come. At Le Puy the concern is rather that the church not be contaminated by sin. The voice of the Church in the liminal inscriptions of Le Puy speaks in the accent of the local authorities. The tone of the verses is defensive, the imperatives unyielding. These do not sound like decorative formulas. Were they meant to be obeyed, literally? If the verses of the west portal were carved on the risers where they now are in the twelfth century, would people who thought of themselves as sinners have found it difficult to step over them? I will return to these questions.

    In the allegorical symbolism of the portal, the imagery of the gate of heaven and the fountain of life plays an important role. Characteristic Romanesque uses of these images can be traced to the last third of the eleventh century in Spain to the Aragon of Sancho Ramírez (1063-1094) and his son Pedro I (1094-1104). The image of the gate of heaven is implicit as early as 1072 in the portal inscription of Nuestra Señora of Iguácel (Huesca). Iguácel is a small eleventh-century church in the Pyrenees east of the pilgrimage road coming down from the Col du Somport to Jaca. Kingsley Porter first called attention to its importance in 1928.13 Between the arch of the portal and the corbel table above it is a horizontal inscription band, which reads in part:

This is the gate of the Lord [porta Domini] through which the faithful enter into the house of the Lord [domus Domini] which is a church [eglesia] founded in honor of Saint Mary. It was built on the order of Count Sancho together with his wife named Urraca. It was finished in the era 1110 [= A.D. 1072] in the reign of King Sancho Ramírez in Aragon. . . . Amen.14

Ten years after the church was completed, Count Sancho gave it to the monastery of San Juan de la Peña.15 The eglesia on the literal level is the physical structure built by Count Sancho and Urraca in honor of the Virgin Mary. The

    The castle of Loarre (Huesca) is about twenty miles south of the pilgrimage road from Jaca to Puente la Reina. After a successful battle against the Muslims near its site, Sancho Ramírez built the castle c. 1070 as a frontier outpost and installed in it, with the assistance of the abbot of San Juan de la Peña, a community of Augustinian Canons.16 Since the chapel is high up in the castle, the south portal is at the level of the crypt. A flight of stairs leads up from the portal to the level of the church in a design which is similar to that of Sacra di San Michele in Italy. A frieze of Christ in Majesty with angels, with an inscription around the mandorla, is set above the arch of the portal. The top half of the frieze was destroyed in the eighteenth century; the bottom half is mutilated. Not only is the frieze difficult to see from any normal viewing point, but the inscription is so deteriorated that it can no longer be read. However, early in this century Don Ricardo del Arco deciphered the words fons ego sum vita(e), "I am the fountain of life."17

    This image, with its reference to the sacrament of baptism, was associated with the portal on the basis of three passages in the Bible. In his second apocalyptic vision John says: "And he showed me a river of water of life, clear as crystal, proceeding from the throne of God and of the Lamb" (Apocalypse 22:1). John's vision was interpreted with the help of two prophetic texts from the Old Testament. One was the vision of Ezekiel: "And he brought me again to the gate of the house [ad portam domus], and behold waters issued out from under the threshold of the house toward the east: for the forefront of the house looked toward the east: but the waters came down to the right side of the temple to the south part of the altar" (Ezekiel 47:1). The other was the words of the prophet Joel: "And it shall come to pass in that day, that . . . waters shall flow through all the rivers of Juda: and a fountain shall come forth of the house of the Lord, and shall water the torrent of thorns" (Joel 3:18).

    In the monastic church of Santa Cruz de la Serós (Huesca), the allegorical symbolism of the portal as the gate of heaven and the fountain of life was combined. Santa Cruz is situated between Loarre and the pilgrimage road from Jaca to Puente la Reina. The church, constructed for the Benedictine sisters of Saint John and dedicated to the Virgin Mary, was finished and decorated c. 1100. A primitive tympanum displays a Chi-Rho monogram, or chrismon, in the center, which is flanked by two lions, each wearing a delicate chain collar and a ring about his tail just above the tuft. Four leonines appear on the tympanum. Three run clockwise from the top around the rim of the chrismon; the fourth is on the lower edge of the tympanum beneath the lions' feet. Since these words are inscribed on the rim of the chrismon, the identification of the

Ianua sum perpes; per me transite, fideles.

Fons ego sum vite; plus me quam vina sitite.

Virginis hoc templum qui vis penetrare beatum,

Corrige te primum, valeas quo poscere Christum.

(I am the eternal door; pass through me, faithful ones. I am the fountain of life; thirst for me more than for wine. You who wish to enter this blessed church of the Virgin, reform yourself first, in order that you may be able to call upon Christ.)

A shift from plural imperatives in verses one and two to singular forms in verses three and four indicates that the latter two verses belong together as a single utterance. The plural imperatives, which are addressed to all Christians, give them a general command to pass through the door and to thirst for the fountain. The singular imperative which is addressed to any one of the faithful who is about to carry out the general command directs him or her to reform before entering in order to be able to pray to Christ after entering.

    It seems unlikely that the audience for these commands was intended to be the sisters of the order of Saint John, or at any rate the sisters only. They might be thought to be in some sense within the church already. The route between Jaca and Puente la Reina over which pilgrims to Santiago de Compostela moved in great numbers in the eleventh and twelfth centuries passed within two or three miles of Santa Cruz, which was in turn the gateway to the Cluniac monastery of San Juan de la Peña in the hills behind it. Pilgrims visitingSanta Cruz may have applied the commands to themselves.

    The monastery of San Juan de la Peña (Huesca) is dug into the side of a great overhanging cliff. It was founded in the eighth century by a hermit named Juan. About 1014 Sancho el Mayor brought the monastery under the rule of Cluny. The Archbishop of Bordeaux assisted by the bishops of Jaca and Maguelone consecrated the upper church on December 4, 1094.18 A portal in the south wall of the nave communicates with the cloister. The portal is capped with a Mozarabic horseshoe arch which may have been reemployed from the ninth-century lower church.19 At some point an inscription was carved on it in Visigothic script . The inscription faces the cloister so the words are visible as one enters the church. It is a full leonine elegiac couplet, which reads:

Porta per hanc caeli fit pervia cuique fideli,

Si studead fidei iungere iussa Dei.

(Through this gate the gate of heaven becomes accessible to each believer who strives to combine the commandments of God with faith.)

In syntax it differs from the inscriptions of Santa Cruz and Loarre. The main clause of the couplet employs the third-person indicative, and the injunction in the pentameter is equally impersonal. Yet the descriptive statement ("the gate of heaven is accessible through this gate"), which can only be validated by the active faith of the believer, is arguably the voice of the Church. The phrase "this gate" implies a speaker who is on the scene (as the statement, "the gate of heaven is accessible through the portal of the monastic church of San Juan de la Peña," would not).

    The voice of the lintel inscription of the late eleventh-century tympanum of the Cathedral of San Pedro of Jaca (Huesca) does not explicitly identify itself as the gate of heaven or the fountain of life, yet the imperative directive to the viewer to cleanse him or herself and the use of the verb venire and the adverb huc imply these themes:

Vivere si queris qui mortis lege teneris,

Huc suplicando veni, renuens fomenta veneni.

Cor viciis munda, pereas ne morte secunda.

(If you who are bound by the law of death seek to live, come hither in prayer, renouncing the fomentations of poison. Cleanse your heart of vices, lest you perish in the second death.)

    The symbolism of the portal as the gate of heaven and the fountain of life did not originate in Aragon.20 But these portals are the oldest surviving Romanesque exemplars of the two themes, which quickly spread beyond Spain. In France, by the second half of the twelfth century, the themes had merged into a conception of the portal as the gate of life. The Benedictine priory church of St.-Marcel-lès-Sauzet (Drôme) was founded in 985 by Lambert, the Count of Valence, and his wife in the hills along the east bank of the Rhône just outside of Montélimar. It was subsequently attached to Cluny. The existing church is a work of the second half of the twelfth century.21 Above the door is a Christ in Majesty tympanum with the Four Animals. The Four Animals are in an unusual chiastic order: Matthew is at the upper right; Mark at the lower left; Luke at the lower right; and John at the upper left. The Christ has been replaced at some later date by a Virgin and Child. All the heads, except for the Eagle's, have been chiseled off. The lower edge of the tympanum is inscribed with three full leonine hexameter verses:

Vos qui transitis, qui crimina flere venitis,

Per me transite quoniam sum ianua vite.

Ianua sum vite. volo parcere . . . venite.

(You who are passing through, you who are coming to weep for your sins, pass through me since I am the gate of life. I am the gate of life.

Here the voice of the portal is the voice of Christ, speaking as the gate of life. The gate of eternal life is, of course, the gate of heaven, but the persistence of images of thirsting and weeping makes clear that the fountain of life is an underlying theme. There can be little doubt that these verses, not only in their general conception, but even in details of wording, derive ultimately from Santa Cruz de la Serós. The portal (ianua), speaking in the first person, commands the faithful to pass through it (per me transite).

    A similar conception governs the west portal of the Benedictine priory church of Sant Pau del Camp in Barcelona. Sant Pau was a monastic dependency of Sant Cugat del Vallés. It was probably founded c. 1117. The present church is thought to be the work of the beginning of the thirteenth century.22 Although it is terribly weathered, the salient features of the portal decoration can still be made out . A tympanum of Christ in Majesty rests on a massive lintel. Saint Paul kneels on Christ's right hand; Saint Peter on his left. The Four Animals are arranged on the facade in counterclockwise order: Matthew is above the archivolts on the upper left and the Eagle representing John is on the upper right; the Lion and Bull representing Mark and Luke project from the base of the outer arch. Seven circles with a variety of interior designs are carved on the imposts. They seem to complement the simple chrismon in the center of the lintel. A narrow inscription band frames the four sides of the lintel. This carries two full leonine hexameter verses and a prose statement possibly to the effect that Renard placed seven Arab workers (Moabitinos) in the construction project for his own sake and for the sake of the soul of his wife Raimunda.23 The verses which run along the whole top edge and the left half of the bottom edge of the lintel read:

Hec Domini porta via . . . est omnibus horta.

Ianua sum vite, per me gradiendo venite.

(This is the gate of the Lord, the way . . . is open to all. I am the gate of life, come and walk through me.)

While identities in phrasing (ianua sum vite) and leonine rhyme (vite / venite) do not necessarily imply a direct connection between St.-Marcel-ès-Sauzet and Sant Pau, they suffice to show that the image and the rhyme had become commonplace by the beginning of the thirteenth century.

    Whatever the relationship between Sant Pau and St.-Marcel, the portal of Sant Pau must have been influenced by the west portal of Maguelone. The former Cathedral of St.-Pierre of Maguelone (Hérault), more or less deserted now, occupies a spot midway along the slender barrier island that shelters the mainland around Montpellier from the open Mediterranean. It stands on a tiny peninsula an island at the time the cathedral was built in the twelfth century

    Although this hodgepodge might not lead one to suppose that any unified conception of the portal governed its final arrangement, comparison with Sant Pau del Camp offers reason for believing that the arrangement reproduces an earlier, more coherent one. The inscriptions which enclose the floral design of the lintel on all four sides are of two kinds. There is a prose inscription which gives us the name of the maker (in what sense is uncertain), Bernard de Tréviers, and the date, A. D. 1178. And there are four full leonine hexameter verses. This layout of the inscription bands around the four edges of the lintel, combining leonine verses with a prose inscription including a person's name, is identical to the layout of Sant Pau. The Sant Pau tympanum was arguably an imitation of the design of an early twelfth-century tympanum of Maguelone the tympanum that was apparently broken up and used to provide the kneeling figures of Paul and Peter which are now below the lintel on either side of the door.25

    The verses of Maguelone read:

Ad portum vite sitientes quique venite;

Has intrando fores vestros componite mores.

Hinc intrans ora, tua semper crimina plora;

Quicquid peccatur lacrimarum fonte lavatur.

(You who are thirsting, come to the port of eternal life; in entering these doors, correct your behavior. Pray as you enter here, always weep for your sins; whatever is amiss is washed by the fountain of your tears.)

The shift from the plural imperatives of the first two verses to the singular of the second two parallels the shift in the verses of Santa Cruz de la Serós. Maguelone's location on an island in the inland waterway with access to the Mediterranean though a narrow channel cut in the eleventh century by Bishop Arnaud perhaps inspired the elegant pun in the first verse. The faithful are urged to come, not ad portam, to the gate, but ad portum vitae, to the port or harbor of life. Here in embryonic form is one of the organizing metaphors of the poetry of courtly love from the twelfth to the sixteenth centuries.

    The author of the verses on the north portal of the abbey church of Notre-Dame of Bourg-Dieu of Déols (Indre) explored other metaphorical implications of the gate of life. There the portal addressed the pilgrim as an exile held captive in Babylon (earth) and offered passage to the road which led back to his or her native land (heaven).26 The abbey was founded by Ebbe le Noble, Prince of Déols, in 917, and given by him to the monks of Cluny. In the twelfth century, the abbey church, measuring some 370 feet in length, was one of the great structures of western Europe. Its grandeur offered no protection from devastation during the Hundred Years' War and the Wars of Religion. Later, it served for two centuries as a stone quarry. Finally, in 1830, most of its remains, including the north porch, were auctioned off to a road contractor and demolished.27 Fortunately, the ruins of the abbey, including the decoration of the north portal, were described before the demolition by Father Dubouchat and by a traveler who visited the ruins about 1825.

    From their descriptions we learn that the north portal of Déols was ornamented with a tympanum of Christ in Majesty with the Four Animals.28 As at the Royal Portal of Chartres, to which this scheme was evidently related, the apocalyptic image of Christ in Majesty was accompanied by references to worship, to the secular arts and sciences, and to manual labor in three recessed orders of archivolts decorated with figures. The inner archivolt displayed angels in adoration; the keystone represented the Lamb of God. The middle archivolt displayed nine statues of the Arts and Sciences. At the summit of the archivolt was Philosophy with one banderole in her right hand which she offered to Grammar and another in her left hand which she offered to Dialectic. Standing beside Grammar were Rhetoric, Astronomy, and Arithmetic. Beside Dialectic stood Geometry, Music, and Physics. Each of these figures was probably identified by an inscription.29 The outer archivolt figured the Works of the Months.30 A statue of the Virgin was mounted above the ensemble. Only a fragment of the tympanum and the Lamb of God remain. These are displayed in the Musèe Bertrand in Châteauroux.

    On the lower edge of the tympanum the voice of the portal spoke in a hexameter couplet with disyllabic end rhyme:

Qui captivaris miser in misera regione,

Ad patriam r[editus per me patet ex Babylone].31

(You wretch who are held captive in a wretched place, the way back from Babylon to your homeland lies open through me.)

Faith and the Church were the necessary conditions for passage through the gate and onto the road home, but at Déols work and study seem to have been offered as palliatives to the misery of the human condition in exile. Philosophy addressed Grammar and Dialectic in verses which were inscribed on the banderoles in her hands. She said to Grammar:

*Auri grata sonent sicut sacra dictio partes.

Per te doctrine relique noscuntur et artes.32

(May the parts of speech sound as pleasing to the ear as sacred Scripture. Through you the other sciences and arts are known.)

And to Dialectic she added:

*Notaque sint nostris per te sophismata nostra.33

(May our faults of logic be made known to us through you.)

    Philosophy's verses have the same form hexameter couplets with disyllabic end rhyme as the portal verses. By stressing the value of grammar and dialectic for the pursuit of knowledge in the arts and sciences, the poet may have intended to imply the quality of instruction, in those subjects at least, of the monastic school of Déols. They were the gates to higher education as the wordplay on the standard portal formula per me in both couplets makes clear. And yet there is a crucial difference between Philosophy's verses and the couplet spoken by the Portal. The words of the Portal are addressed the viewer, while the words of Philosophy are addressed the statues on either side of her. From the point of view of the worshiper, the former speech is direct address, while the latter is a representation of direct address.

    What can we make of the communication between the portal and its audience? Its message was conveyed in Latin the sacred language of the western church, but not a language any longer accessible to common people. Would the typical visitor have been able to read the words of the portal? The question is really twofold, because it involves not only the matter of comprehension, but also of visibility.

    To take the latter point first, many inscriptions are today badly worn and degraded. Even with the aid of binoculars they can be hard to read. But, when they were newly carved and perhaps painted, would a viewer with normal eyesight have been able to distinguish all the letters without serious difficulty (the question is really whether they were intended to be legible, but there is no way of answering this question directly). We cannot always be sure that present viewing conditions correspond to the original ones. The inscriptions of Ste.-Foy of Conques can easily be read with the naked eye from the parvis in front of the basilica. But if the tympanum was once enclosed within a narthex, or if, as one theory has it, it has been displaced from its original location within the church, would the inscriptions have been as visible as they now are? The answer to that question would depend on our knowing both the amount of light that was available and the viewer's line of sight. It seems safe to assume that the penitential couplet on the lintel, at least, would have been visible to entering pilgrims. At Santa Cruz de la Serós the portal is relatively low and the

    It is difficult to estimate how many worshipers might have been functionally literate in Latin. No doubt most of the clergy was literate, but certainly not all of them. There was some lay literacy among both the aristocracy and the merchant and artisan classes, particularly in southern Europe. The great mercenary warrior chieftain, El Cid, whose peregrinations may have taken him along the Jaca-Pamplona road, and who died in Valencia (1099) just about the time the portal of Santa Cruz de la Serós was being decorated, could read and write, and so, apparently, could his wife Jimena.34 But for many (the majority perhaps?), the words of the tympanum would have been unintelligible without the aid of an interpreter.

    It may appear that, in any logical sense of the term, the performative utterances of these inscriptions could only be described as being either "hollow" or "unhappy" or both. Insofar as we are dealing, not with spoken language, but with words memorialized in stone, the utterances might be said to have the same empty status as any speech composed by an author for a character to speak. And insofar as the performatives are not, or were not, seen or comprehended by their audence, they might be said necessarily to fail in their effects.

    Still in all, there is something unsatisfying about this preliminary judgment. Portal inscriptions themselves do not all share the same status as speech acts. Inscriptions that employ the first or second personal pronouns can be distinguished from those that do not. It does not matter finally whether the speaker uses only the first person or only the second person or both, since I can only use the pronoun "I" when I am addressing "you," and conversely the use of the pronoun "you" implies an "I" who uses it. Émile Benveniste observes,

...it is a fact both original and fundamental that these "pronominal" forms ["I" and "you"] do not refer to "reality" or to "objective" positions in space and time but to the utterance, unique each time, that contains them.... The importance of their function will be measured by the nature of the problem they serve to solve, which is none other than that of intersubjective communication.35

In Benveniste's sense, these voices all speak in performative utterances, apart from the question of whether they give specific commands, merely in utilizing the first or second personal pronouns. They effect "the conversion of language into discourse." Benveniste goes on to say: "Thus the indicators I and you cannot exist as potentialities; they exist only insofar as they are actualized in the instance of discourse, in which, by each of their own instances, they mark the process of appropriation by the speaker."36 Benveniste's formulation applies, if I understand it correctly, to spoken discourse. As phrased, it cannot literally be true of written texts, still less of words frozen in stone. Writing would seem to convert the pronouns I and you to "potentialities" and the statements of which they are a part to mimetic representations of discourse. But there must be a class indeed, a large class of written utterances that have the potential to be reconverted to discourse at the moment they are read by a receptive reader or reported to a receptive listener.37 In these cases it is the "you" that actualizes the discourse. If the reader or listener accepts the role of "you," then it would appear that "intersubjective communication" has taken place.

    On these grounds, it seems to me possible to distinguish the voice of the portal (in the first or second person) from a mimetic utterance, even one in the first or second person that happens to be inscribed on the portal. The voice of the portal is, or purports to be, the voice of the Church or of God. It was also, of course, the work of a human author (a monk in the local monastery, for example), writing in Latin in a conventional form of verse. No one would have doubted the human authorship of these inscriptions. However, it seems equally certain that they would not have been received by their audiences nor intended by their authors as scraps of a fictional monologue. The ontological status of the putative speaker (God or the Church) was radically different from any fictional character, or even from any historical person. Even though the form of the words (the use of Latin, the verse structure, the diction, the syntax) was the responsibility of their human author, their content was an expression of the divine message of God, ratified by the Church. In this sense, the human author functioned as a scribe or as a kind of simultaneous translator.

    Fiat lux has been described as "the supreme performative."38 This phrase comes to us as reported by the author of Genesis, in the Latin of Jerome's translation of the Hebrew text. It is necessarily a mimesis of a speech act which occurred once in principio, at the beginning of time. But if the words inscribed on the church were felt to be the words of the Church or of God, those words might have been thought immediately operative whether anyone "heard" them or not. Moreover, the dual problem that some inscriptions (at least) were difficult or impossible to see and that some viewers (at least) could not read Latin could have been and, as is becoming increasing apparent, almost certainly would have been solved by the intervention of "guides" or interpretators, whose function would be parallel to the human author of the sacred words.39

    Thus, it is possible to make a distinction between inscriptions which are or were performatives when they are actualized by their viewers and inscriptions which are mimetic performatives or merely descriptive utterances. At Déols, for example, the performatives spoken by Philosophy were mimetic, but the portal inscription was potentially a genuine performative. If it no longer is, it is because the fragment of the tympanum with the fragment of the portal inscription on its lower edge is in a museum. The conditions for a performative are not happy, not only because the inscription is incomplete, but more importantly because it is not properly in place above the portal of a functioning church, which is (or was once thought to be), literally, Christ and the Gate of Heaven, and thereby endowed with their voice.

    It would exceed the limits of this paper to chart the breakdown of the Romanesque allegorization of the church. But I will suggest that a crucial turning point comes when Abbot Suger had an image of himself carved on the Last Judgment tympanum of St.-Denis and inscribed his own words in hexameter verses with disyllabic end rhyme on the lintel:

*Suscipe vota tui iudex districte Sugeri

Inter oves proprias fac me clementer haberi.40

(Accept, strict Judge, the prayers of your Suger; grant that I be preserved among your own sheep.)

The voice is still in the first person, but this is a mimesis of the prayer of the living Suger, not a performative utterance. It has the curious effect of transforming the figure of Christ on the tympanum above him into a representation as well. The material things of this world, which includes the art of the tympanum, are intended to lead the faithful toward the contemplation of the immaterial things of the next. On the gilded bronze doors beneath the tympanum, Suger added these verses:

*Nobile claret opus, sed opus quod nobile claret

Clarificet mentes, ut eant per lumina vera

Ad verum lumen, ubi Christus ianua vera.41

(Bright is the noble work; but, being nobly bright, the work

Should brighten the minds, so that they may travel, through the true lights,

To the True Light where Christ is the true door.)

Suger has reinterpreted the allegorization of the church by means of the pseudo-Dionysian doctrine of transcendental beauty. He had political reasons for doing so, but his reinterpretation also suggests that Bedan allegorical hermeneutics as conventionally understood no longer seemed fully adequate to the task of mediating between man and God. The portal of the church, properly decorated, might be able to lift the minds of worshipers to the "True Light," but it was there, not here, that Christ was the "true door." Henceforth, prosopopoeia in an inscription carved in stone could only be a mimesis of a speech act. It is perhaps no accident that the sculptural programs on the facades of the major Gothic cathedrals were not accompanied by verse inscriptions. The voices of the portals had fallen silent.42