Abstract
This article compares Fra Angelico’s devotional panels at San Marco with Jan van Eyck’s Arnolfini Portrait in order to distinguish two kinds of pictorial difficulty: difficulty that belongs to an object exceeding comprehension, and difficulty that belongs to the procedures by which meaning is authorised. It accepts Georges Didi-Huberman’s account of Angelico’s grounds, blotches, and colours as theologically and materially intelligent operations of figura and dissemblance. It argues, however, that Angelico’s difficulty deepens participation in a world whose ultimate meaning remains secured by Christian revelation. His unstable marks and dissemblant surfaces multiply interpretation while directing the viewer toward an authority the tradition already supplies. The Arnolfini Portrait operates differently. Its gestures, mirror, inscription, Christian signs, and material details invite competing accounts of the event without providing a procedure for settling them. Each interpretation reopens the question of what kind of scene is depicted and what would authorize an answer. The article develops the concept of “settlement” to distinguish the painting’s closed history as property from its open history as an object of interpretation. It concludes that Angelico makes the limits of depiction serve the security of faith, whereas Van Eyck makes the search for security part of the encounter with painting itself.
I. Two Kinds of Unsettling
A Dominican friar walks along the east corridor of the convent of San Marco in Florence some time in the 1440s. At the height of his eyes, beneath a painted sacra conversazione in which the Virgin and Child are enthroned among saints, he passes four rectangular panels covered in spattered, dripped and flecked pigment. From a few steps away they read as imitation marble, the kind of veneer that might clad an altar or the lower wall of a chapel. If he slows down, the resemblance begins to come apart. The flecks describe no particular stone; they are colour thrown at a wall. And if he knows his Scripture and its commentators, as a friar of San Marco was trained to, the marks open onto a field of associations: Christ as the rock from which water flowed in the desert, the stone cut without hands in Nebuchadnezzar's dream, the sealed tomb through which the risen body passed, the Word made flesh. The panels decline to be pictures of marble, and in declining they turn the friar's attention towards what no picture could show.
A visitor, perhaps a guest of the household and perhaps a later owner, stands before a small oak panel painted in Bruges in 1434. Everything in Jan van Eyck's Arnolfini Portrait is visible with a clarity that seems to promise full disclosure: the pile of the fur trim, the brass of the chandelier, the oranges on the chest beneath the window, the room reflected in the convex mirror on the back wall. The longer she looks, the less the looking settles. A man raises his right hand. A woman lays her right hand in his left. Two small figures appear in the mirror at the threshold of the room. Above the mirror, in a careful script, the painter has written that he was here. The visitor can describe every object and still be unable to say what kind of event she is looking at, who is attesting it, or what she herself is doing in front of it.
Both images reward sustained looking. Both exceed any message that could be paraphrased in a sentence, and both have been taken as evidence that early fifteenth-century painting was doing far more than learning to imitate appearances. These resemblances make the comparison worth pursuing, but they cannot decide it. What decides it is a narrower and harder question: what does the difficulty of an image do to the authority of the world it presents?
An image may resist exhaustive interpretation because its object exceeds human comprehension. The difficulty then belongs to the object, and the viewer falls short before something whose excess has already been named. An image may also resist settlement because the procedures that would establish its meaning have themselves become problematic. The difficulty then belongs to the mediation: the viewer cannot tell which considerations ought to govern interpretation, and every attempt to decide reopens the question of what would count as deciding. These are two different structures of difficulty, and much of the confusion in discussions of pictorial ambiguity comes from treating them as one.
Georges Didi-Huberman's Fra Angelico: Dissemblance and Figuration (1995) is the most sustained demonstration that Angelico cannot be understood as a backward or merely pious participant in the Renaissance conquest of appearances. Its achievement is to recover the theological and material intelligence of Angelico's grounds, blotches and colours, and to show that they work according to a medieval logic of figura which the modern notion of "figurative" painting conceals. This article accepts that achievement. It asks a further question, one which Didi-Huberman's book does not set out to answer: does that intelligence transform symbolic mediation, or does it help secure an inherited mediation under changing pictorial conditions?
The answer proposed here is that Angelico's pictorial difficulties deepen participation in a world whose ultimate meaning remains secured by Christian revelation, whereas Van Eyck's Arnolfini Portrait, read through Living Value Theory (LVT), makes the conditions of mediation themselves uncertain. To keep this argument honest, three propositions have to be held apart throughout. The first is that Angelico is more sophisticated than a narrowly mimetic art history allows. The second is that his sophistication unsettles conventional accounts of what painting is and does. The third is that his painting unsettles the symbolic world within which it operates. The first two are accepted here without reservation. The third does not follow from them, and Didi-Huberman nowhere asserts it; the question is whether his vocabulary of disturbance, equivocation and presence lends it more support than it can bear. The comparison concerns particular works and their operations. Angelico does not stand here for Italian painting, and Van Eyck does not stand for the North.
II. What Didi-Huberman Enables Us to See
Didi-Huberman begins with the procedure that art history usually takes for granted. Following the opening chapter of Panofsky's Studies in Iconology, the historian who enters the first cell at San Marco first identifies a motif, the representation of a natural object, and then a figure, the motif raised to the dignity of theme, allegory or story. In Angelico's Noli me tangere this procedure works without friction. Christ and Mary Magdalene are recognised by their attributes, the spade on Christ's shoulder recalls the moment when she mistook him for a gardener, and the gesture of the right hand translates the three words of the Gospel into a single movement. Once the storia has been identified, the historian believes the subject of the work has been grasped. The ground then becomes a container: the meadow indicates that the scene takes place outdoors, a palm tree reminds us that we are in Palestine, and the story could in principle unfold without either (Didi-Huberman 1995: 14–17). The figure carries meaning; the place merely holds it.
Didi-Huberman's first move is to restore the ground to the argument. For a Dominican formed in the school of Albertus Magnus, place was no neutral container. Following Aristotle's remark that place has a certain power, Albertus called it an "active principle of engendering", a site in which forms are produced rather than merely housed (Didi-Huberman 1995: 18). Angelico's practice, Didi-Huberman argues, stands at the level of this concept. The meadow of the Noli me tangere is laid down in two shades of green with great economy, almost frontally, and across it the painter deposits small, slightly raised spots of white and then of terra rossa.
The second move is to follow those spots. At first glance they are flowers, and in Peirce's terms they would be icons of flowers in a meadow. On closer inspection they have no calyx, no petals and no stamens. Angelico could paint a flower when he wished to; here he chose to lay down small circular heaps of red. Asked what they are painted like, the only rigorous answer is that they are painted exactly as Angelico paints Christ's stigmata, and their constellation across the green ground ends in the same mark on Christ's hand and foot (Didi-Huberman 1995: 19–20). Seven times in the fresco the red spots appear in groups of five, the number of the wounds. Aquinas had answered the objection that the risen body's scars were a defect by granting them a "special beauty", so that the stigmata could be read as the flowers of Christ's body, or the risen Christ could be seen sowing his wounds in the garden of the earthly world (Didi-Huberman 1995: 20–21). The single pigment then travels across the whole surface. The same red marks sin in the Magdalene's robe, suffering in Christ's wounds, compassion where it returns to her, Passion and Resurrection in the scattered flowers, and the boundary of flesh wherever a red stroke outlines a body (Didi-Huberman 1995: 22). Didi-Huberman names such marks signa translata, "transit signs", whose material existence sets up "a labile movement between signs of very different semiotic status", so that they prohibit "any univocal relation of attribution" (1995: 21). Their value lies in "displacement, movement, and association rather than definition, identification, or predication" (1995: 21).
The third move explains why such displacement could be valued rather than tolerated. The Dionysian tradition, transmitted through Erigena's translation and the commentaries of Albertus and Aquinas, distinguished images fashioned in the likeness of their object from images that push fiction to the limit of the implausible. A bearded king on a golden throne resembles divinity; an earthworm does not. Dionysius preferred the earthworm, because "resemblance fools us": before the crowned king we risk "believing God is adequately and univocally figurable" (Didi-Huberman 1995: 51–52). The incongruous figure frustrates that mistake. It purifies the viewer's expectations and draws the mind beyond the visible, since nothing is humbler than an earthworm and Christ in the Passion humbled himself further still (Didi-Huberman 1995: 53).
The fourth move is the most ambitious, and it is the one that most resists a merely referential reading. The marbled panels of the Madonna of the Shadows are painted at eye level, in proximity and frontality, without perspectival recession. Didi-Huberman attributes to them "a quality of presence" and reads them as an "effective sign", figure and mystery at once (1995: 87). A figura Christi, he writes, "did not merely evoke a meaning, 'Christ,' but also actualized that meaning in the everyday time frame of whoever invoked or convoked it" (1995: 88). The image regards the friar who regards it, much as Nicholas of Cusa would later take a painted face that follows the beholder as a model of God's vision (Didi-Huberman 1995: 88). The panels intervene between the readable storia above them and the anagogical virtue of dissemblance: "it elicits an act of prayer; it invites an instant of contemplation" (Didi-Huberman 1995: 90).
None of this can be reduced to the formula that unusual things symbolise God. Didi-Huberman's account concerns material operations, the temporal movement of signs across a surface and through Scripture, and the devotional efficacy of images in a convent whose library held seven copies of the Dionysian corpus (Didi-Huberman 1995: 51). It establishes convincingly that Angelico's painting does more than illustrate a known story. The question that remains is whether this excess constitutes a different symbolic mode, or an intensified operation within a theological order that it leaves secure.
III. Dissemblance within a Secured World
The first step towards an answer is to separate two kinds of uncertainty that Didi-Huberman's vocabulary tends to run together. A viewer may remain uncertain about the significance of a mark. The friar before the red spots may be unable to say whether they are flowers, wounds, a sowing, or all three. Christian revelation nevertheless supplies the horizon within which the mark becomes meaningful at all. Three questions can therefore be distinguished: which scriptural relations a detail activates, how an image conducts contemplation, and what ultimately authorises both the relations and the contemplation. The first two can remain open indefinitely while the third remains entirely secure. Uncertainty of interpretation is compatible with certainty of authorisation, and in Angelico's case the second is what makes the first productive.
The second step is to see why inadequacy, in this setting, reinforces authority. Dissemblance was never a free-floating aesthetic category. It had a precise theological address, the regio dissimilitudinis, the region of unlikeness into which Augustine found himself fallen, far from God (Didi-Huberman 1995: 45–47). Every human being lives in that region after Adam; the image of God persists in the soul alone, while bodies and the visible world carry only a "nonspecific" resemblance, the trace or vestigium that an effect leaves of its cause (Didi-Huberman 1995: 48). From this Didi-Huberman draws the consequence that devout painting "does not proceed by means of the image but by means of the vestige" (1995: 49). The painter of an istoria, even of Christ, paints "knowing full well it is not that", and yet "this knowledge of imperfection lifts me above it" (Didi-Huberman 1995: 49). The figure's failure counts as evidence of the object's transcendence. Far from weakening faith, the difficulty sustains it, by preventing a finite appearance from being mistaken for divine adequacy. Didi-Huberman states the result with precision: the Dionysian operation "allows us both to preserve the mystery's transcendence and to put constantly into play the immanence of the visuality of figures" (1995: 54). Transcendence is preserved, and the play of visuality is what preserves it.
The third step distinguishes inexhaustible mystery from recursive indeterminacy. Christian mystery need not be exhaustible in order to orient an encounter securely. Medieval exegesis, as Didi-Huberman describes it, created "an infinite world of relations", in which every particle of sacred text could enter into a new correspondence with every other, yet these relations swirl "endlessly around this central kernel", the mystery of the Incarnation (1995: 7). The threefold function of the figure that organises the first half of his book, memory of the mystery, prefiguration and presence, is a structure for multiplying meanings in a determinate direction (Didi-Huberman 1995: 60). The Annunciation is his clearest example. Its "temporal circle" joins the present instant of the Incarnation, the past origin of all flesh and the future death on the cross, folded together by the coincidence of 25 March as the feast of both conception and crucifixion (Didi-Huberman 1995: 76–77). The temporalities are interwoven, but every thread runs towards the same end. An infinite field of associations can remain directed towards a determinate religious goal, so multiplicity alone does not show that the framework of interpretation has become unstable. Didi-Huberman himself supplies the formula. Of the dogmas of virgin motherhood and the Word made flesh he writes that "dogmas that are paradoxes cannot be firmly established; their perpetual vacillations can only be energized and directed" (1995: 8). The vacillation is real, and it is directed.
The fourth step proposes a historical interpretation of what this direction achieved. In the decades when Angelico was working, painting was acquiring an unprecedented capacity to make appearances persuasive: perspectival construction, modelled bodies, coherent light, the istoria that Alberti placed at the summit of the art. Such a capacity carries a theological risk, since the more fully painting can show reality, the more plausible it becomes that what it shows is sufficient. Angelico's dissemblance can be read as a way of keeping theological mediation intact under these conditions. It demonstrates the image's material inadequacy precisely where visible likeness might otherwise appear to suffice. Didi-Huberman's reading of the Madonna of the Shadows already contains this structure: the upper register offers the semblance of divine motherhood, while the marbled panels below act as "a kind of retraction", a coloured blurring that says "no, it is not like that", and directs the gaze anagogically towards grace (1995: 56). Semblance is offered and then qualified, and the qualification is authorised by the same tradition that authorised the semblance. This is offered as an interpretation of what the works accomplish within their setting. It does not attribute to Angelico a conscious programme for defending theology against realism, and nothing in the argument requires one.
The fifth step guards against a misreading. To say that Angelico secures inherited mediation is not to cast him as a medieval remainder resisting innovation. Didi-Huberman insists that there is no reason to oppose, in his case, an avant-garde pictorial humanism to a backward-looking scholasticism (1995: 17), and that Angelico was "modern" in many ways while "constantly displacing his own stylistic innovations by conceptualizing them in terms of Dante or Albertus Magnus" (1995: 10). His Annunciations at Cortona and in the Prado employ architectural space and recession with full competence, and then set the event a few metres from the Garden of Eden, from which Adam and Eve are being expelled, producing a space and temporality that are aberrant from a historical point of view (Didi-Huberman 1995: 113). The contemporary resource is used, and it is redirected towards an inherited purpose. What looks like conservatism from the vantage of a history of styles is, from the vantage of mediation, the active reuse of new means for old ends.
Angelico's world can accommodate profound pictorial difficulty because that difficulty is itself theologically authorised. His forms unsettle the adequacy of depiction while strengthening the orientation of devotion.
IV. Symbolic Modes and the Effects of Images
To make the comparison exact, the analysis needs a vocabulary that does not rank images by how complicated they are. LVT distinguishes three symbolic modes. In coordination mode, a symbol participates directly in the situation it names. In reference-to-absence mode, it makes something not presently available matter to current coordination; memory, narrative, description and most ordinary discourse work this way. In re-presentation mode, it claims to make present again what is absent, lost, imagined, or never directly encountered; ritual, art and theory are its characteristic sites. A single symbolic act may operate in more than one mode at once, and the point of the distinction is to prevent distinct practical operations from collapsing into a generic category called meaning. The modes do not classify images as simple or difficult, realistic or abstract, weak or emotionally powerful. They describe which relations constitute an image's operation.
Used carefully, the distinction blocks three category mistakes. The first is to infer from an image's capacity to prompt conduct that it operates in coordination mode. Many images move bodies, from devotional panels to road signs, without participating in the situation they name in the way that a greeting participates in a meeting. The second is to infer from an image's capacity to evoke an absent object intensely that it operates in re-presentation mode. Intense evocation is often reference-to-absence at full strength, and the intensity may come precisely from the felt distance between the sign and what it signals. The third is to infer from an image's multiple or unstable referents that it has ceased to operate through reference-to-absence. A symbol can point in five directions at once and still be pointing. In each case the right question concerns which relations constitute the work's operation and what evidence establishes their presence.
These distinctions change how Angelico's devotional images should be described. Didi-Huberman shows that many surfaces at San Marco practise what he calls injunction. Above the door, Saint Peter Martyr, bleeding, lays a finger on his lips and enjoins silence; some frescoes integrate inscriptions of address (ave, salve) or admonition (cave); the nonbiblical figures at the edges of the scenes assume the gestures of prayer and offer them back to the friar as a mirror (Didi-Huberman 1995: 90; see also Hood 1993). These are images that organise conduct. Whether they operate in coordination mode is a further question. The friar's silence is coordinated by the Rule and by the common life of the house, and the image takes part in that coordination by depicting a saint whose authority over the corridor is presupposed rather than produced by the picture. The image prompts conduct by referring it to an authority not present in the corridor.
The same care applies to presence. Didi-Huberman's praesentia is explicitly sacramental. Presence and representation, he notes, "are exact opposites", unless representation is understood in a sacramental sense that presupposes mystery (1995: 87), and the marbled panels play towards the devout image "the role... that the altar plays in relation to the devout man in the real liturgical space" (1995: 91). The claim to make something present is real, but it borrows its warrant from the altar and the Eucharist, which the panels figure without being. And the dissemblant figure, by its own logic, declares the inadequacy of whatever it shows: it frees meaning "in view of something" that "always takes its distance or approaches without ever taking on substance" (Didi-Huberman 1995: 89). This is reference-to-absence at its highest intensity, an absence made to matter precisely by being marked as beyond figuration. To say so takes nothing away from the devotional force of the panels. It locates that force in a relation to an absent God whose absence is already theologically defined.
LVT's five recursivity levels sharpen the point. At L1 coordination proceeds without becoming separately salient; at L2 something becomes questionable, felt as misalignment before it can be stated; at L3 it becomes articulable within a particular interaction; at L4 such articulations are stabilised into portable forms that travel beyond their occasion; at L5 one can ask why a question is askable at all. The friar who slows down before the marbled panels experiences an L2 disturbance, since something about the marble is off. His exegetical associations articulate that disturbance at L3. Every articulation, however, draws on L4 stabilisations he did not make and is not invited to remake: the four senses of Scripture, the Dionysian doctrine of images, the liturgy of the altar. These stabilisations are the condition of the difficulty's productivity. In LVT's terms, Angelico's figures have very high intersymbolic dependency, since much of their intelligibility depends on access to external symbolic systems, and the systems on which they depend are authoritative.
The comparative criterion can now be formulated:
When interpretation encounters difficulty, what can secure orientation, and does the work make that securing authority itself an object of recursive scrutiny?
The question connects the symbolic modes to the demands that images place on viewers without turning the modes into a scale of increasing radicality. It also explains why the Arnolfini Portrait needs to be approached differently, as a different organization of the encounter rather than a more complicated symbolic code.
V. Van Eyck: Visibility without an Authoritative Account
The promise of visible evidence
The Arnolfini Portrait begins by offering more evidence than any viewer could need. Fur, wool, brass, glass, wood, wax, fruit and flesh are each differentiated with a precision that oil painting had not achieved before and that has rarely been matched since. The room is lit with a clarity that seems to leave nothing in shadow. Such precision encourages a specific expectation: that closer inspection will clarify the scene, and that whatever is needed for understanding must be present somewhere on the panel (Ecks 2026: 5). In LVT's vocabulary, the painting has extraordinary mesocosmic fidelity and habitability. A viewer registers a coherent, inhabitable interior long before any of its distortions, such as the oversized chandelier or the shortened bed, become noticeable.
The difficulty is that clearer visibility does not supply a procedure for deciding what the scene is. It would be an overstatement to say that everything in the painting is equally salient or that it yields no intelligibility. A great deal is locally intelligible: this is a man, this is a woman, this is a bed, a dog, a pair of pattens, a string of beads. The stronger and more defensible claim is that local intelligibility does not add up to a conclusive account of the whole. Detail multiplies rather than resolves the possible readings (Ecks 2026: 5), because each newly noticed object raises the question of which framework should organise the others.
Gesture and the uncertainty of the event
The two central gestures show the problem most clearly. The man's raised right hand resembles an oath, a blessing, a greeting or an attestation. The woman's right hand rests in his left, in a joining that suggests betrothal, marriage, intimacy or ritual display (Ecks 2026: 3, 7–8). Each gesture encourages a set of expectations, of witnessing, agreement or affection, without compelling any one of them. The scholarly record bears this out. Panofsky (1934) read a marriage solemnized before witnesses, Hall (1994) an engagement, Carroll (1993) a grant of legal authority within a marriage, Koster (2003) a memorial for a wife who had died before 1434, and Harbison (1990) a performance of sexuality and social standing.
The important uncertainty is not which of these readings is correct, but which framework should organise interpretation in the first place. A gesture cannot be decoded securely until the event is identified, and the event can only be inferred partly from the gesture. Interpretation turns in a circle that the painting does nothing to break. The bedroom setting deepens the circle, since it induces the viewer to look for a formal act in a space normally separated from the spaces where such acts were performed (Ecks 2026: 4). The viewer must tolerate "not just uncertainty about what is happening, but uncertainty about what kind of thing is happening" (Ecks 2026: 12).
Mirror, inscription, and the viewer's role
The convex mirror and the inscription above it sit at the centre of the back wall, one above the other, and they are best analysed together. The mirror extends the depicted field beyond the space directly shown, revealing the backs of the couple and, at the threshold, two small figures entering the room. The reflected figures introduce questions of presence and witnessing that the main scene alone would not raise. The inscription, Johannes de eyck fuit hic 1434, asserts the painter's presence in a hand that recalls the attestations of legal documents, yet it does not fix the capacity in which he was present: witness, notary, guest, maker, or none of these (Ecks 2026: 8–9).
The recursion between these devices is what matters. Evidence about who is present changes the interpretation of the event; if one of the reflected figures is the painter and his presence is a legal attestation, the scene becomes a contract. The proposed event in turn changes what presence and witnessing would mean; if the scene is a memorial, the painter's having been there becomes a claim about the past rather than a signature on a present act. Neither consideration can be settled first.
The viewer is drawn into this recursion. The reflected figures stand roughly where a viewer of the panel stands, so that the painting invites the viewer to relate her own position to the threshold it depicts. "The painting assigns us a position without assigning us a role" (Ecks 2026: 4). Some care is needed here. The claim is that the painting invites identification with the threshold, and that this invitation makes the viewer's role part of what is in question. It is not that every actual viewer occupies one geometrically or socially fixed position, which neither the optics of the mirror nor the history of the painting's reception would support.
Cross-register revision
LVT's five mediations, multisensorial embodiment, multispecies being-with, multiversal dwelling, multimateriality and multisymbolization, describe how the viewer's attention moves across the painting, and how each move revises the others. One concrete sequence makes the point. The gesture of the raised hand, read through embodiment, suggests a formal act. The room, a fully multimaterial environment of bed, chandelier, carpet and chest, complicates the expected setting of that act, and the single candle burning in daylight adds a further anomaly. The mirror introduces witnessing, which belongs to being-with, and the small dog at the couple's feet extends that being-with across species, while also inviting a symbolic reading as fidelity. The inscription, a deposit of multisymbolization, then complicates the identity of the witness. The material objects (oranges, beads, the carved figure on the chair, the discarded pattens) invite symbolic interpretation as fertility, piety, wealth or ritual removal of shoes. Those interpretations return the viewer to the unresolved event. Even dwelling, the non-human-made conditions of light, season and weather glimpsed through the window, is unsettled by illumination that seems to come from more than one direction. Each proposed explanation changes the relevance of the others, so that the viewer "circulates through the painting without ever arriving" (Ecks 2026: 8; see also 11–12).
This is where the LVT criterion becomes decisive. In Angelico, the L3 articulations prompted by difficulty return to stable L4 forms. In the Arnolfini Portrait, each L3 articulation ("this is a betrothal", "this is a memorial") fails to stabilize into an L4 account that the painting would confirm, and the failure keeps reopening the question of what could confirm any account. The painting does not perform L5 inquiry, but it repeatedly places viewers at the point where that inquiry becomes available: why are we asking what kind of transaction this is, and what would entitle us to an answer?
Christian signs without guaranteed pictorial settlement
The Arnolfini Portrait is full of Christian signs. Ten small roundels around the mirror show scenes of the Passion; the beads by the window are for prayer; the single candle has been read as the presence of Christ. The comparison with Angelico requires precision here. The argument should not be that Christian meaning disappears from the painting, nor that the small scale of the Passion scenes proves their irrelevance. It has been argued that the Passion, "the ultimate divine transaction", is miniaturized and displaced to the edge of a merchant's mirror (Ecks 2026: 9, 14). That observation is best stated as a question about function. Do these scenes conclusively establish the nature of the central event, or the viewer's relation to it? In Angelico's Noli me tangere, the red spots return the viewer, however circuitously, to the wounds and the Resurrection, which organise the whole field. In Van Eyck, the Passion roundels frame the mirror that reflects the unidentified witnesses, but they do not tell the viewer whether the scene below them is a sacrament, a contract, a commemoration or a domestic moment. The Christian signs are present and potent, and they do not settle the painting.
Van Eyck's radicality, in LVT's reading, lies in this: viewers are made repeatedly to revise the conditions under which they expect painting to become intelligible. Where Angelico's difficulty is answered by an authority it presupposes, Van Eyck's difficulty extends to the question of authority itself.
VI. Money, Ownership, and the Account That Remains Open
The distinctive contribution of an economic reading of the Arnolfini Portrait lies in a concept that can now be generalized: settlement (Ecks 2026: 9–11). Settlement is the closing of an account. A payment is made, a debt discharged, a title transferred, and nothing further is owed. The concept is useful because it distinguishes two things that ordinary talk about art tends to merge: settling ownership or payment, and settling interpretation. An object may circulate successfully as property while remaining resistant to a definitive account of its significance.
The Arnolfini Portrait is an object of exactly this kind. As property it has behaved impeccably. From the inventories of Margaret of Austria in the early sixteenth century to its purchase by the National Gallery in London in 1842, it has been owned, valued, gifted, catalogued and displayed (Campbell 1998). Every change of hands has closed. As an object of interpretation, by contrast, it has never closed, and six centuries of readings have accumulated without any one of them discharging the others (Ecks 2026: 9–10).
The mirror and the inscription operate on both sides of this distinction at once. The inscription is a maker's mark. It authenticates authorship and so supports attribution, provenance and value; it is the one element in the painting that functions "exactly as money requires" (Ecks 2026: 9). The same line of script declares a presence whose capacity cannot be determined and introduces a temporal puzzle, since it records in the past tense an event that the scene seems to present as happening now. The mirror, likewise, is the supreme demonstration of technical accomplishment, the passage that justifies the price; and it is also where the unidentified witnesses, the viewer's implied position and the Passion roundels converge. Authorship and virtuosity support possession. They do not confer interpretive mastery. The work's material availability for ownership leaves the demands it places on looking exactly where they were.
For LVT, the significance of this distinction is that the encounter with the painting exceeds the transaction through which the painting becomes someone's property. Ownership cannot discharge the demands the work places on looking and interpretation. Every owner acquires the panel; no owner acquires the settlement of its meaning. There is a further asymmetry. The painting is saturated with the forms of attestation, with gesture, witnesses and a signed declaration of presence, yet the one exchange that certainly took place, the payment that turned the painter's labour into the household's possession, has left no trace in the image or in the archive (Ecks 2026: 11). The transaction that made the painting closed so completely that it vanished, while the scene it produced keeps its own account open. The painting is "what the transaction produced and what the transaction could not absorb" (Ecks 2026: 11).
The contrast with Angelico must be stated carefully, because it can easily be misread as an opposition between commercial corruption and religious purity. San Marco was itself rebuilt with Medici money, and devotional images throughout Europe were commissioned, paid for and valued. The relevant difference concerns the available sources of orientation. The friar before the marbled panels is oriented by a liturgical and exegetical order that the image figures and on which it depends. The viewer before the Arnolfini Portrait stands in a room full of the products of completed transactions, and before a scene whose apparatus of witnessing and attestation borrows the forms of settlement without completing one.
The historical claims on which this reading rests need to be kept proportionate. Nothing in the argument requires that money was newly invented in fifteenth-century Bruges, that every monetary transaction eliminated all relational residue, or that the original commission of the painting is documented, which it is not. What the argument does require is a setting in which practices of accounting, credit and valuation had become unusually dense. Bruges was the principal commercial and financial centre of north-western Europe, a city of money-changers' deposit banks and of Italian merchant-bankers who drew and settled bills of exchange on Florence, Lucca, Genoa and Venice (de Roover 1948; Murray 2005; Spufford 1988). The Arnolfini, a Lucchese family active in the luxury trade and in finance for the Burgundian court, belonged to this world (Campbell 1998), and Seidel (1993) has emphasised how closely the painting belongs to that commercial context.
It is also necessary to distinguish the painting's structural analogy with settlement from any claim about Van Eyck's conscious economic critique. The argument identifies a relation between what the painting does and what the practices around it were built to do. Whether Van Eyck intended that relation cannot be known and does not need to be. The analogy holds at the level of the work's operations, and its usefulness lies in showing that the question of who owns an image and the question of what authorises an account of it can come apart completely.
VII. Objections That Strengthen the Distinction
A comparison of this kind risks hardening into an overdrawn opposition. Five objections show where the distinction needs qualification, and in each case the qualification makes it sharper.
The first objection is that Didi-Huberman already theorises an unsettled viewer. His friar is regarded by the image as much as he regards it; the figure works at the conversion of the soul; the encounter changes the one who undergoes it (Didi-Huberman 1995: 88–90). This should be conceded in full. Angelico's viewer is unsettled in a strong sense, and nothing in the present argument depends on treating devotional looking as passive or pre-scripted. The question concerns the destination and authorisation of the transformation. Conversion, in the sense Didi-Huberman intends, has a direction specified in advance: from the visible towards the invisible, from the region of unlikeness towards the image of God. Didi-Huberman's own term for the figure's work on the soul is tropology, the sermo conversus through which Christ "is born daily" in the devout soul (1995: 88), and the very phrase names a conversion whose end is already known. A viewer can be profoundly changed within a religious framework whose authority remains secure, and in Angelico the security of the framework is what gives the change its meaning.
The second objection is that Van Eyck's viewers also inhabited Christian worlds. This is true, and it too should be conceded fully. The Arnolfini household, the painter and every fifteenth-century viewer of the panel lived within Christian institutions, prayed, confessed and were buried by the Church. The argument concerns the painting's provision of interpretive authority, not the beliefs of its viewers. LVT does not need to portray Van Eyck as secular or unbelieving in order to identify a different demand. A Christian viewer can stand before a painting that supplies no decisive Christian reading of its own central event, and the experience of doing so is part of what the painting makes available. Van Eyck's own work confirms that the difference lies in the operations of a particular painting rather than in the painter's faith. Two years before the Arnolfini Portrait, the Ghent Altarpiece had been installed in the church of Saint John (now Saint Bavo) in Ghent, and its Adoration of the Lamb places the viewer before an altar on which sacramental mediation is pictured as fully secure. The same hand that could paint the assured centre of the liturgy could also paint a room in which no centre holds.
The third objection is that the Arnolfini scene seems unresolved only because historical knowledge has been lost. Had the commission survived, it might name the couple, the occasion and the purpose, and the indeterminacy would dissolve. Disagreement among later scholars is indeed insufficient proof of designed inexhaustibility; scholars disagree about many things that were perfectly clear to contemporaries. The argument must therefore rest on observable relations among pictorial features, such as the circularity between gesture and event, or between the identity of the witness and the nature of the act, rather than on the record of disagreement. Additional evidence might well clarify identities or occasion. It would not thereby exhaust the painting's operations, since those operations are carried by the relations among its features, which no document could rearrange. The history of the identification makes the point. Archival work showed that Giovanni di Arrigo Arnolfini did not marry until long after 1434, and that the wife of his cousin Giovanni di Nicolao was dead by 1433 (Campbell 1998; Koster 2003). These findings changed which readings were available, moving scholarship from a wedding towards commemoration. They did nothing to resolve the circularity between gesture and event, which simply reappeared within the new reading: if the scene commemorates a dead wife, the living immediacy of her presence, the raised hand and the witnesses in the mirror all have to be reinterpreted again.
The fourth objection is that the comparison turns sophistication into a competition, with Van Eyck awarded the prize. The article assesses radicality along one specified dimension only: the reorganization of mediation and the recursive demands an image places on its viewers. Along other dimensions, such as theological depth, the intelligence of colour, or the integration of image and liturgy, Angelico's achievement is at least as great, and Didi-Huberman's book is the proof. Angelico's accomplishment does not depend on winning a comparison it was never designed to enter. Nor does the dimension chosen imply a hierarchy of value. An image that secures orientation for a community at prayer performs a function no less demanding than an image that withholds it, and in many settings a more necessary one.
The fifth objection asks whether the argument implies that Van Eyck prefigures Reformation iconoclasm. If the Arnolfini Portrait makes the authorisation of mediation uncertain, it may seem to anticipate the moment, a century later, when the mediating power of images was rejected outright. A restricted hypothesis is defensible: exposing the difficulty of authorising mediation may help make mediation available for scrutiny. Nothing more can be inferred from the painting alone. A causal path from Van Eyck to iconoclasm would require separate evidence concerning reception, institutions, theology and transmission, and the present argument offers none.
VIII. Conclusion: The Point at Which Mediation Becomes a Problem
This article has accepted what Didi-Huberman establishes. Angelico's grounds and blotches are intellectually and devotionally consequential. The red spots of the Noli me tangere move between flower and wound; the marbled panels of the Madonna of the Shadows turn the friar from semblance towards mystery; colour itself carries the memory of the Incarnation across the walls of San Marco. Didi-Huberman makes visible operations that a narrowly mimetic account of Renaissance painting misses entirely, and his reading should change how any historian approaches Angelico.
What does not follow is equally important. Exceeding mimetic depiction does not establish a transformation of symbolic mode, and interpretive richness does not establish that the authority securing an encounter has become uncertain. The red spots are unstable in reference and stable in direction. The dissemblant figure declares its own inadequacy, and that declaration is authorised by the tradition that defines the divine as beyond figuration. Angelico's images can be extraordinarily difficult without becoming difficult about what makes them meaningful.
The two encounters with which the article began can now be described more exactly. The friar who slows down before the marbled panels is unsettled, and the unsettling is the beginning of a passage whose route and destination the convent has prepared for him: from colour to stone, from stone to Christ, from the corridor to prayer. The visitor before the Arnolfini Portrait is unsettled too, but no route has been prepared. Each path she takes through the painting, from gesture to mirror, from mirror to inscription, from inscription to the objects in the room, brings her back to the question of what kind of scene she is in, and to the further question of what would entitle her to say.
The final comparison can therefore be stated now. Angelico's dissemblance directs the viewer through the inadequacy of appearance towards an authorised divine mystery. The difficulty is real, and it is answered in advance by the mystery it serves. Van Eyck's recursive demands make viewers reconsider what could authorize an account of the scene at all, including their own role in producing that account. The difficulty is real, and nothing in the painting answers it.
LVT's contribution to this comparison lies in distinguishing kinds of relation and kinds of demand that broad terms such as ambiguity, presence and figuration tend to conceal. Presence may be sacramentally borrowed rather than pictorially produced. Ambiguity may concern an object's excess or the procedures for settling an account. Figuration may multiply meanings around a fixed kernel or keep reopening the question of which kernel there is. The same distinctions caution against assigning the whole Arnolfini Portrait to a single symbolic mode simply because it is radical. Its significance lies in how it reorganizes the relations among coordination, reference to absence and re-presentation, and in how it exposes the work required to sustain orientation when no pictorial element can secure it on the viewer's behalf.
Angelico makes the limits of depiction serve the security of faith. Van Eyck makes the search for security part of the encounter with painting itself.
Selected References
Campbell, Lorne. 1998. The Fifteenth Century Netherlandish Schools. National Gallery Catalogues. London: National Gallery Publications.
Carroll, Margaret D. 1993. "In the Name of God and Profit: Jan van Eyck's Arnolfini Portrait." Representations 44: 96–132.
de Roover, Raymond. 1948. Money, Banking and Credit in Mediaeval Bruges: Italian Merchant-Bankers, Lombards and Money-Changers. Cambridge, MA: Mediaeval Academy of America.
Didi-Huberman, Georges. 1995 [1990]. Fra Angelico: Dissemblance and Figuration. Translated by Jane Marie Todd. Chicago: University of Chicago Press.
Ecks, Stefan. 2026. "Unsettling the Accounts: The Origins of the Money Economy in the Arnolfini Portrait." Living Value Theory, livingvaluetheory.org.
Hall, Edwin. 1994. The Arnolfini Betrothal: Medieval Marriage and the Enigma of Van Eyck's Double Portrait. Berkeley: University of California Press.
Harbison, Craig. 1990. "Sexuality and Social Standing in Jan van Eyck's Arnolfini Double Portrait." Renaissance Quarterly 43 (2): 249–291.
Hood, William. 1993. Fra Angelico at San Marco. New Haven: Yale University Press.
Koster, Margaret L. 2003. "New Light on the Arnolfini Portrait." Apollo 158 (499): 3–14.
Murray, James M. 2005. Bruges, Cradle of Capitalism, 1280–1390. Cambridge: Cambridge University Press.
Panofsky, Erwin. 1934. "Jan van Eyck's Arnolfini Portrait." Burlington Magazine for Connoisseurs 64 (372): 117–127.
Seidel, Linda. 1993. Jan van Eyck's Arnolfini Portrait: Stories of an Icon. Cambridge: Cambridge University Press.
Spufford, Peter. 1988. Money and Its Use in Medieval Europe. Cambridge: Cambridge University Press.