Jan Van Eyck Was Here: Inhabitability and the Arnolfini Portrait

Abstract

Jan van Eyck’s inscription above the mirror in his portrait of 1434, Johannes de eyck fuit hic, Jan van Eyck was here, makes an extraordinary claim: that the painter shared bodily continuity with the room he painted. This article asks what a painting must achieve before such a claim becomes even minimally intelligible, and argues that the answer is inhabitability rather than resemblance, iconographic completeness, or optical accuracy. Read against Robert Campin’s near-contemporary Mérode Altarpiece, where an intensely realized domestic interior still exists to host the Incarnation, the Arnolfini Portrait marks the point at which an ordinary domestic mesocosm becomes sufficient on its own terms, needing no sacred event to justify its representation. The painting’s famous dog, oranges, and mirror are not, on this account, symbols awaiting decoding. They are the visible surface of an achievement that lies elsewhere, in the coherence of a world someone could, in principle, have stood inside. Before any viewer asks what this world means, Van Eyck has already made the more fundamental claim that it is somewhere one could have been.

I. Fuit Hic

Above the convex mirror at the centre of Jan van Eyck’s portrait of Giovanni Arnolfini and his wife, in an ornate script that draws attention to itself as writing rather than as mere information, is a sentence: Johannes de eyck fuit hic. 1434. Jan van Eyck was here.

The sentence is stranger than it first appears. Van Eyck does not write I painted this. He does not write I made this, or even the simple self-assertion a signature would supply on its own. He claims to have been somewhere. Fuit hic asserts a location, a presence, a body once positioned inside a space, in a tense that has already closed by the time anyone reads it.

This article asks what a painting has to achieve before such a claim becomes even minimally intelligible. Not a claim about art, and not a claim about representation. A claim about place. Before a painter can meaningfully say I was here of a room he painted rather than merely depicted, that room has to be the kind of thing bodily presence could occupy. It has to be, in a sense this article will spend its length making precise, inhabitable.

The apparently trivial detail through which this becomes visible is a small dog standing between the two figures. Van Eyck has given it individuated hairs, a coat that catches light differently along its back and flanks, paws that visibly bear weight on the floorboards. It stands exactly where a household dog would stand, present without announcing its presence. Viewers have found it very difficult to leave alone. The most familiar reading assigns it a meaning: dogs signify fidelity, a commonplace of medieval symbolism, and once that possibility is admitted the dog is drawn into a wider interpretive programme in which the couple’s joined hands become a marriage vow, the burning candle becomes a divine witness, the discarded shoes become a sign of holy ground, and the mirror figures become legal witnesses to whatever the painting is thought to record.

But the dog and the inscription are, on the argument of this article, the same problem approached from two directions. The inscription asks what kind of world a body could once have stood inside. The dog asks what kind of thing needs to be true of a painted object before it can simply be that object, without owing a further debt to symbolic meaning. Both questions converge on the same answer, and the rest of this article is an attempt to state that answer with enough precision to be worth having.

II. The Mérode Altarpiece and the Almost-Inhabitable Room

Van Eyck did not invent the inhabitable domestic interior, and the claim this article makes would be considerably weaker if it depended on his having done so. The crucial comparison is Robert Campin’s Mérode Altarpiece, painted in Tournai in the 1420s, a full decade or more before the Arnolfini Portrait and among the most extraordinarily realized domestic interiors in early Netherlandish painting. A room with a bench, a table, a fireplace, a laver, a towel, windows with round bullseye panes, a book left open by a breeze, is rendered with a specificity that anticipates almost everything Van Eyck would later do with fur and floorboards.

And yet the room exists, finally, to host something else. The subject of the central panel is the Annunciation, the moment at which Gabriel arrives to tell Mary she will bear the Christ child, and every mesocosmic achievement in the painting is placed in service of that event. The extinguished candle marks the moment of the Incarnation. The lilies mark Mary’s purity. The laver marks ritual cleansing. And in the upper left, entering through the round window on seven rays of golden light, a tiny naked Christ child carries a small cross toward Mary, the Holy Spirit’s usual dove replaced by a miniature incarnate body already bearing the instrument of its own eventual death.

This is worth pausing on, because it is the exact moment at which Campin’s room stops being merely a room. However convincingly rendered, however full of towels and benches and the ordinary furniture of a Flemish household, the space cannot sustain its own mesocosmic coherence once a naked infant begins flying through the window on visible rays of light. The miraculous does not merely visit this room. It claims it, overriding whatever ordinary physical continuity the rest of the painting had so carefully built. Mary and Gabriel are, for all their contemporary dress and their contemporary furniture, bodies belonging to sacred representation first and domestic inhabitants only second. Their proportions can be adjusted for compositional or symbolic reasons no ordinary body would tolerate. Their draperies can fall in patterns governed by devotional convention rather than by gravity. The room, in short, remains a stage rather than becoming, in its own right, a world.

Campin comes extraordinarily close to the achievement this article is tracking, closer than almost any painter before Van Eyck, which is exactly what makes the comparison useful rather than merely illustrative. He demonstrates that a painter working a full generation earlier already possessed the technical and observational capacity to build a domestic interior a viewer could nearly believe in. What he does not do, and could not have done without abandoning his subject altogether, is let that interior stand on its own mesocosmic feet, sufficient without a miracle to justify the labour spent on it.

This gives the historical claim its proper shape, and it is worth stating with some care, because it is easy to overstate and the overstated version is considerably less interesting than the accurate one. Van Eyck did not invent the inhabitable domestic interior. He helped emancipate it from the sacred event that had previously been required to justify its existence. The Annunciation makes sacred history domestic. The Arnolfini Portrait makes domesticity sufficient. Between those two sentences lies most of what this article has to say.

III. What Inhabitability Means

Inhabitability needs to be defined precisely enough to carry the weight this article puts on it, because it is easy to let the word collapse into something already familiar, realism, naturalism, optical accuracy, correct perspective, and each of those substitutions would misdescribe what is actually being claimed.

A room is inhabitable, in the sense intended here, insofar as its five mediations cohere sufficiently to afford possible continuations. A viewer standing before the Arnolfini Portrait can, in imagination, cross the floor. Stroke the dog. Remove the pattens lying discarded near the bed. Sit on the bench, or on the edge of the bed itself. Run a hand along the fur trim of the wife’s gown. Open the window and feel whatever air comes through it. Eat one of the oranges resting on the chest beneath the window. Turn and greet the two figures visible in the mirror, entering from a doorway the viewer’s own position implies. None of these actions can actually be performed. The floor is paint. The dog cannot be stroked. The orange has no taste. What the painting supplies is not the action itself but the affordance of the action, the sense that the action would make sense here, would meet no resistance from the world’s own construction, would continue smoothly into whatever came next. This is what distinguishes an inhabitable painted world from a merely well-observed one. A still life can render fruit so precisely that a viewer registers its weight and texture without for a moment believing there is a room around it to eat the fruit in. The Arnolfini Portrait supplies the room as well.

Notice, now, how little this definition has to do with geometric correctness, and this is the paradox on which the rest of the analysis depends. The room Van Eyck paints is not a room that could exist exactly as depicted. The chandelier does not sit comfortably within the space assigned to it. The bed is oddly short for the figures who would presumably use it. The convex mirror is almost certainly larger than any mirror of its kind that could have been manufactured in the period. There is nowhere in the room for a fireplace, in a climate where one would be a practical necessity rather than a luxury. The perspective does not resolve into a single consistent geometric system if traced through the floorboards, the window recess, and the bed canopy together. Even the mirror, which promises an exact optical record of what stands before it, does not reproduce the foreground scene with anything like the fidelity a real convex mirror would.

None of this registers, on a first encounter with the painting, as a problem. It has to be sought out, measured, and argued for by scholars equipped with photographs, tracings, and a willingness to disbelieve their own initial impression. The room feels correct long before anyone has checked whether it could be built, and this tells us what inhabitability actually requires and what it does not. It is not photographic accuracy. It is not geometric consistency. It is coherence across a world’s mediations sufficient to make that world recursively available as a place one could, in imagination, walk into, sit down in, and continue living in without the whole arrangement collapsing under scrutiny. Van Eyck can misjudge the chandelier’s fit to its recess, shorten the bed, enlarge the mirror, and abandon a single consistent perspective system, while never once losing the sense that a viewer is looking into an actual room rather than at a diagram of one. He can even improve the sense of inhabitability by doing so, since a room built to strict geometric consistency can feel more like an architect’s plan than like a place where furniture accumulates the way furniture actually accumulates in a lived space, at odds with one another, never quite squared away.

The room is not perfect. It is inhabitable. Those are different claims, and this distinction is one of the article’s most theoretically important results, because it means mesocosmic fit and geometric fit are simply not the same achievement, whatever the history of art criticism has sometimes assumed.

IV. Five Mediations, and What Painting Cannot Do

What actually constructs this inhabitability can be traced through the five mediations that constitute the human mesocosm: multisensorial embodiment, being-with, multimaterial forming, multiversal dwelling, and multisymbolism. Each does real work in the room, and it is worth following each rather than merely naming it, since the analysis gets considerably stronger the further it is pushed.

Multisensorial embodiment is articulated through fur, skin, hair, and polished surfaces rendered so that they evoke touch and not merely sight. The dog’s coat is built from strokes that suggest softness rather than simply depicting brown pigment in a dog-like arrangement. The floorboards evoke hardness. The beads hanging by the mirror evoke a small cool weight. The wife’s heavy gown evokes a real, dragging mass, distinct from the crisper resistance the fur trim would offer a hand. Being-with is articulated through the couple’s specific bodily arrangement, hands joined, bodies angled toward each other while remaining available to the viewer, and through the two further figures visible in the mirror, who extend the room’s sociality beyond the two principal figures without the painting needing to depict them directly. The dog itself participates in being-with rather than remaining mere pictorial furniture, since a dog standing exactly there, between two particular people, is doing something a dog painted anywhere else in the room would not be doing. Multimaterial forming is articulated through the chandelier, the shoes, the fruit, the bed, and the window, all rendered not simply as shapes but as things that afford particular actions: shoes that can be removed, as the discarded pattens attest, fruit that can be eaten, a chandelier that can be lit, polished, and bumped into by someone tall enough to reach it. Multiversal dwelling is articulated through the room’s status as somebody’s room rather than a generic interior, furnishings and possessions establishing a household with a particular, situated life, socially placed as much as spatially placed. And multisymbolism runs through all of it, in the writing on the wall, the tiny Passion scenes framing the mirror, and the clothing that carries its own social distinctions, so that some of what the painting depicts is already symbolic before the painting gets to it.

But a further distinction has to be introduced here, because everything said so far risks suggesting that the painting reproduces the mesocosm it depicts, and this is precisely what it cannot do. The human mesocosm, in ordinary lived experience, is constituted through continuous recursive loops across these five mediations: a body sees a chair, approaches it, sits on it, feels its resistance, occupies the room through it, perhaps recognizes it as somebody’s chair, and the loop continues, endlessly revisable, endlessly responsive to what happens next. An artwork cannot reproduce these loops. A viewer standing before the Arnolfini Portrait cannot actually stroke the dog, cannot discover whether its coat feels as soft as it looks, cannot walk behind the couple to inspect their clothing from another angle, cannot open the window, eat the orange, or ask the couple what they are doing. Nothing recursive actually happens between viewer and painted world. What happens instead is that the painting articulates enough mediational cues, cues drawn from embodiment, being-with, materiality, dwelling, and symbol, for a viewer to recursively reconstruct an inhabitable world from them, doing the actual recursive work herself, in imagination, using resources the painting has merely supplied.

This licenses a proposition worth stating on its own. A painting can simulate mesocosmic coherence without reproducing mesocosmic recursivity. It does not need to give a viewer an actual room to loop through. It needs only to supply cues rich enough that the viewer’s own embodied, social, material, and symbolic competence can complete the loop unaided, filling in resistance a floor would offer, warmth a hand would find, weight a fabric would carry, without ever noticing that any filling in was required. This is one of art’s genuine freedoms, and it is the freedom Campin’s room, for all its accomplishment, could not fully exploit while a body was still flying through its window.

V. Four Kinds of Symbol

The word symbol has been doing too much work in accounts of paintings like this one, covering relations that are not, on inspection, the same relation at all. At least four need to be kept separate.

The first is referential symbolization, the operation ordinary language performs. The spoken or written word dog makes something absent recursively available. There need be no dog in the room for someone to say the dog is outside. This kind of symbolization substitutes for presence, and its success depends on sufficient closure: dog cannot ordinarily mean chandelier, Tuesday, or potato, whatever latitude context and metaphor allow at the margins.

The second is conventional symbolic surplus, the operation iconography identifies. The painted dog is already present within the depicted mesocosm, seen rather than merely referred to, and no substitution is needed to make it available. A further relation can nonetheless be opened from it, dog toward fidelity, adding rather than replacing: nothing about the dog’s presence is cancelled by the added relation, and the added relation is stabilized, where it is stabilized at all, by a shared repertoire external to the painting rather than by anything internal to the dog’s own depiction.

The third is indexical relation, in which one thing’s presence or position provides evidence about another thing by virtue of an actual causal or material connection, rather than by convention or resemblance. Smoke indexes fire. A trail indexes passage.

The fourth is mesocosmic constraint, in which a depicted configuration makes some further state of affairs implausible or unavoidable simply by virtue of what the configuration, taken as a whole, would actually require. This is not a symbolic relation in any of the preceding senses. Nothing stands for anything else. The configuration itself does the constraining.

The dove in Van Eyck’s own Annunciation paintings shows why the distinction matters, and it makes a useful bridge back to Campin’s altarpiece before the argument returns to Arnolfini. The dove is iconic insofar as it resembles a dove. It is conventionally symbolic insofar as the relation between dove and Holy Spirit depends on an established Christian repertoire, the same repertoire Campin’s flying Christ child draws on and revises. And within the represented event it also carries indexical force, since its trajectory toward Mary along divine rays indicates, within the painting’s own narrative logic, the actual occurrence of divine intervention rather than merely alluding to it from outside. Three relations operate on the same painted bird at once, iconic, conventional, and indexical, and its mesocosmic presence as an ordinary bird that happened to fly into a room is, given this configuration, close to impossible. The dove cannot just be a dove.

Now return to the Arnolfini room, and notice how differently its objects behave. The oranges resting near the window are not obviously indexical of anything beyond their own presence, are not caught up in any conventional Christian repertoire the way the dove is, and are not obviously referential in the way the word orange would be. Whatever significance they carry, and the article returns to this shortly, arises through mesocosmic constraint rather than through symbolization in any of the first three senses. Expensive fur and imported oranges need not symbolize wealth at all. They can simply make certain further facts about the household implausible, which is an entirely different, and considerably more precisely constrained, kind of achievement.

VI. The Dog Can Just Be a Dog

Set the dove and the dog beside one another and the historical transformation this article is tracking becomes visible without requiring any inflated claim about the birth of modernity to carry it. The dove cannot just be a dove. The Arnolfini dog can just be a dog. That single sentence is worth returning to more than once, because it compresses an entire argument that would otherwise require several paragraphs to restate.

There is, correspondingly, comparatively little room for disagreement about what the dog mesocosmically is. One might dispute its exact breed, or how common such dogs were in prosperous fifteenth-century households, but the space of reasonable disagreement is narrow. Nobody seriously proposes that the shape between the two figures is a cat, a rabbit, or an allegorical figure wearing a dog costume. It has canine proportions, a canine gait implied in its stance, a coat that behaves like fur rather than like anything else, and a scale that fits comfortably against the legs of the two people beside it. There is, in every useful sense, one dog.

Now ask what the dog symbolizes, in the second, conventional sense distinguished above, and the space of reasonable disagreement expands immediately and does not stop expanding. Fidelity. Marriage. Domesticity. Wealth. Status. Vigilance. Companionship. Sexuality, on some readings. A specific, now unrecoverable biographical relationship, on others. Several of these simultaneously, on still others. Or nothing at all.

This gives the article its central constraint asymmetry, worth stating as a general proposition rather than leaving as an observation local to this one painting. Mesocosmic fit strongly constrains what something can be. Intersymbolic fit much more weakly constrains what something can mean. A dog cannot become just anything while continuing to fit as a dog, a chair cannot sustain just any relation to a body while continuing to function as a chair, a room cannot violate its own geometry indefinitely while continuing to feel inhabitable. Conventional symbolic interpretation, by contrast, is comparatively unconstrained, particularly once a shared repertoire has weakened or a painting has travelled far enough from its original viewers that competing repertoires can all seem equally plausible. The dog remains stubbornly singular. What the dog might mean multiplies around it without limit, and none of that multiplication threatens the dog’s own mesocosmic stability in the least.

Call this the painting’s symbolic halo. Civilization surrounds relatively stable mesocosmic things with proliferating symbolic doubles, so that around the dog there accumulates not only fidelity but marriage, domesticity, status, a century of iconographic scholarship, criticism of that scholarship, criticism of the criticism, and now an argument about why scholars felt the need to turn the dog into fidelity in the first place. The dog does not proliferate. Its halo does.

VII. What the Oranges Are Not Symbols Of

It has become customary to say that the oranges resting near the window symbolize wealth, and this article has been building toward the claim that the sentence, however common, describes the wrong relation.

Run a brief thought experiment. Twenty oranges sitting in an orange grove tell a viewer almost nothing about wealth, because oranges growing where oranges grow are exactly what the local mesocosm would produce regardless of who owns them. The same oranges, resting on a chest beneath a window in fifteenth-century Bruges, a city built on long-distance trade rather than citrus cultivation, tell a viewer a great deal, because their presence there is not what the local mesocosm would produce on its own. Someone had to import them, at real cost, across real distance, through a real commercial network, and that fact is not added to the oranges by a symbolic convention. It is built into what their presence in this particular room, at this particular moment, actually requires.

This is mesocosmic constraint operating exactly as the previous section defined it, and it needs to be kept sharply distinct from conventional symbolic association, because the two support very different kinds of inference. Dog toward fidelity is a symbolic association: the dog is fully present and fully itself before any such relation is added, the relation is not required by anything the dog’s presence alone would entail, and a viewer could reject the association entirely without the dog becoming somehow incoherent. Expensive imported goods toward an affluent household is a mesocosmically constrained inference: the goods do not stand for wealth the way a word stands for its referent, they make certain further states of affairs, principally poverty, implausible given everything else the room’s configuration already establishes. Fur, imported fruit, elaborate clothing, fine furnishings, a substantial domestic space, and the capacity to commission one of Europe’s most technically extraordinary painters simply do not fit, taken together, with a modest household. No hidden iconographic code is required to reach that conclusion. The conjunction does the work on its own.

This is why the fur need not symbolize wealth. It cost money, and its cost is legible in the room without any further symbolic apparatus. The imported oranges need not symbolize wealth. They cost money, and their presence, given where the room is, already constrains what kind of household this can be. Commissioning Van Eyck himself cost money, considerably more than the fur or the fruit, and the painting’s own existence is, by the same logic, evidence of the same fact. This does not mean wealth is unimportant to the painting. It means wealth enters the painting mesocosmically, through the constrained inference a whole configuration supports, rather than symbolically, through a convention any one object might or might not be carrying.

VIII. When a Dog Can Afford to Be Just a Dog

A further question has been waiting throughout this argument, and it deserves a section of its own rather than a passing mention, because the answer turns out to be as historically consequential as anything about mediations or mirrors. Why did a domestic room, populated by nothing more dramatic than a merchant, his wife, and their dog, become worth this much pictorial labour in the first place.

Paintings are costly condensations of skilled labour. They require materials, training, workshop organization, patronage, and time, and painting of the technical sophistication Van Eyck practised is especially costly. For most of the preceding centuries of European art, that expenditure was justified by purposes exceeding the depicted mesocosm itself. Altarpieces, devotional panels, and narrative cycles of sacred history participated in worship, salvation, and ecclesiastical authority, and it was this transcendental economy, not the pictorial interest of the depicted objects themselves, that justified pouring extraordinary labour into their production. Campin’s Christ child had to fly through a window because the labour surrounding him needed a justification that ordinary domestic life, however beautifully rendered, could not on its own supply.

The Arnolfini Portrait belongs to a different economy, sustained by mercantile wealth, long-distance trade, and an expanding class of bourgeois patrons who could pay for representational labour without needing to route the payment through the church. A wealthy merchant could commission a painter to render the reflection of light across polished wood, the individual hairs of a household dog, the dirt accumulated on a pair of discarded shoes, because the merchant could pay for it, and because the room and its contents were, in the most literal sense, his to have painted. This is the profound novelty, and it is easy to understate it by saying merely that merchants became new subjects for portraiture. The deeper transformation is that mercantile wealth made it possible to expend extraordinary artistic labour on representing things that no longer required transcendental justification to earn that expenditure. Mesocosmic specificity had itself become valuable enough to support the cost.

Money belongs to this transformation as a symbolic system in its own right, and it operates with strikingly different properties from either of the other two symbolic economies this article has been tracing. Christian iconography stabilizes symbolic surplus through inherited, shared relations, the dove that cannot just be a dove, the candle that marks the Incarnation. Artistic symbolism of the conventional kind, dog toward fidelity, can open surplus, generating readings a shared repertoire only loosely constrains. Money does neither. A price is designed to suppress rather than generate interpretive proliferation. If something costs a given sum, that figure is not ordinarily an invitation to ask what deeper reality it represents. Its efficacy depends on closure, not surplus, and this is precisely what allows money to make heterogeneous things, fur, fruit, labour, paintings, commensurable with one another without requiring any of them to be hermeneutically decoded first.

The Arnolfini Portrait therefore sits at the meeting point of two overlapping symbolic economies operating by opposite logics. One produces value by relating mesocosmic things to transcendental meanings, and it survives in the painting’s Passion scenes and its Christian inheritance more broadly. The other produces value by making mesocosmic things commensurable through price, indifferent to transcendence altogether, and it survives in the sheer, unapologetic material specificity with which everything in the room has been rendered. Neither economy replaces the other within the painting. Their coexistence is precisely what the room records, and it is why a dog can, for the first time this completely, afford to be just a dog.

IX. The Mirror between Two Economies

This coexistence has a precise spatial expression within the painting, and it is worth tracing before turning to the inscription that gives this article its title.

The foreground of the Arnolfini Portrait is overwhelmingly domestic and secular. Two contemporary people, a dog, clothes, fruit, furniture, shoes, a bed, a chandelier. Christian sacred history has not vanished from the painting, and the stronger claim is precisely that it has not. It has receded, concentrated now in ten tiny scenes from the Passion of Christ framing the convex mirror at the room’s centre. A Christian symbolic order and an increasingly autonomous domestic mesocosm occupy the same panel without either one displacing the other.

The mirror sits exactly at their meeting point, and this is what makes it more than an especially impressive piece of optical bravura. Around its frame is the Passion, the inherited symbolic economy in its most concentrated form, judged against scripture, doctrine, and centuries of accumulated devotional practice. Inside it is the contemporary domestic room, repeated in miniature, along with two additional figures entering from a doorway behind the implied viewer. The mirror is therefore the hinge between the painting’s two economies, holding the older, transcendentally justified order at its perimeter while containing, at its centre, the newer, mesocosmically sufficient world the rest of the painting has been building. Nothing in the painting insists that a viewer choose between them. The room, and the century of transformation it sits within, could evidently hold both.

X. Jan van Eyck Was Here

Return, now, to the inscription this article opened with, and reconceive it entirely in light of everything the intervening sections have established.

It would be easy, and it has been common, to read Johannes de eyck fuit hic primarily as an act of artistic self-assertion, a painter proudly signing an unusually ambitious work, or as a claim to legal witnessing, the artist attesting, in the manner of a notary, to some ceremony the painting records. Both readings assume the sentence is fundamentally about status, about Van Eyck asserting his own importance or his own reliability as a witness. Neither reading is wrong exactly, but both lead with the wrong term. They lead with Jan van Eyck. The sentence itself leads with a place.

Lead with inhabitability instead, and the inscription becomes a claim of an entirely different kind. Jan van Eyck was here does not merely assert that a particular painter existed and once stood somewhere. It asserts continuity between two mesocosms, the painter’s own lived, embodied, historical existence and the represented mesocosm the panel has constructed, a continuity that could only be claimed if the represented room were the kind of thing bodily presence could actually have occupied. The sentence, unpacked, says something closer to this. I walked this floor. I breathed this air. I stood before these people. I occupied this room. None of that could be said of a merely decorative or emblematic space. It can only be said of a room inhabitable enough that a real body might once have stood inside it, which is exactly the achievement the rest of this article has been tracing through fur, floorboards, chandeliers, and oranges.

The mirror deepens the claim rather than merely decorating it, and the two elements, read together, supply a structure worth setting out explicitly, because it is the article’s central analytical move. The inscription, in language, says I was here. The mirror, in reflected image, says someone was here, since it shows two figures entering the room from precisely the position the viewer, or the painter, would occupy standing before the panel. And the room itself, in its accumulated mesocosmic coherence, says here is somewhere someone could be, supplying the condition without which neither of the other two claims could mean anything at all. Language asserts presence. Optical reflection implies a witness. Inhabitability is what makes both assertions coherent rather than empty. Remove the room’s mesocosmic achievement and the inscription becomes a claim about nowhere in particular, and the mirror becomes a trick with no room behind it to reflect.

The contrast with biblical painting sharpens this further, and it is worth stating plainly because it marks a limit Van Eyck himself could not cross. He could not have written Jan van Eyck was here across one of his own paintings of the Annunciation, and not merely as a matter of decorum. He was not at the Annunciation. No one alive when he painted was. The event reaches every painter of it through symbolic transmission alone, through scripture, doctrine, and an inherited pictorial tradition, never through anything a body could claim continuity with. The Arnolfini room belongs, by contrast, to the same historical mesocosm Van Eyck himself inhabited, a room in a city he could have walked through, furnished with objects a merchant of his acquaintance could plausibly have owned. Only a room of this kind permits the inscription’s peculiar claim. Sacred history can be painted with total conviction and still remain, for the painter, somewhere he was never actually able to be.

One further detail rewards attention, and it is easy to pass over because it looks like a grammatical accident rather than a deliberate choice. Fuit is past tense. Not est, is. Was. The living encounter the inscription claims has already closed by the time anyone reads the words above the mirror. Whatever bodily presence once occupied that room has departed, along with everything else that made the moment recursive, responsive, alive to what happened next. What remains is a nonrecursive material object, a panel of oak and pigment, capable nonetheless of reopening something of that vanished presence for every subsequent viewer who stands where Van Eyck once claimed to have stood. I was here. I am no longer here. This painting makes the here recursively available again, not by restoring the original recursive encounter, which is gone beyond recovery, but by supplying enough mesocosmic coherence that a later viewer’s own recursive competence, her own capacity to imagine crossing a floor or stroking a dog, can complete what the painting itself can only ever propose.

XI. Panofsky’s Necessary Dog

This returns the argument to the discipline that has produced most of what is known and disputed about this painting.

Erwin Panofsky’s 1934 reading of the Arnolfini Portrait as an artistic marriage certificate remains fascinating not because its individual claims have since been challenged, though many have, but because the interpretive reflex behind it is itself worth examining. Confronted with an extraordinarily detailed domestic scene, the reflex asks, in effect, surely this cannot be all. Surely the dog is not merely a dog. Surely the shoes, the candle, the joined hands, and the figures in the mirror are not merely a household’s ordinary furniture and ordinary gestures. There must be a further level, and iconography supplies it: dog becomes fidelity, shoes become sacred ground, candle becomes divine presence, gesture becomes matrimonial vow, mirror figures become witnesses, inscription becomes legal attestation.

Iconographic interpretation, described this precisely, performs a specific operation. It converts a mesocosmically sufficient object, an object that already has an entirely adequate reason for being present, into an intersymbolically incomplete one, an object that cannot be considered fully understood until its further symbolic destination has been supplied. The dog cannot simply be a dog under this operation, because a dog that is simply a dog closes an iconographer’s work almost as soon as it begins.

Technical study conducted since Panofsky’s article has made his particular totalization increasingly difficult to sustain on its own terms. Infrared reflectography carried out at the National Gallery found no underdrawing at all beneath the dog. Van Eyck painted it directly onto the already finished interior rather than planning it as part of the design from the outset. The chandelier was added the same way, and several other details long treated as central to the painting’s supposed symbolic programme turn out, on the same evidence, to have been introduced at a late stage, without preparatory drawing beneath them. This does not prove the depicted objects carry no symbolic significance whatsoever. It weakens something more specific, the assumption Panofsky’s reading requires, that the painting was generated from one prior symbolic programme in which every mesocosmic detail already possessed its assigned intersymbolic destination before Van Eyck ever picked up a brush. Scholars since, Linda Seidel and Margaret Carroll prominent among them, have proposed rival readings, a legal instrument rather than a sacrament, a record of commercial rather than devotional standing, each internally consistent, which is itself evidence of how weakly the intersymbolic layer constrains interpretation once the mesocosmic layer has finished its own, considerably better constrained, work.

None of this means iconography is simply mistaken whenever it finds a symbolic relation, and the qualification matters enough to state carefully. The dog may indeed have evoked fidelity for contemporary viewers. Nothing in this article denies that possibility, and nothing in the painting rules it out. The mistake is narrower and more specific: treating symbolic completion as though it were necessary for pictorial completion, as though the painting could not be said to have succeeded until its objects had been assigned their proper symbolic destinations. The painting does not require fidelity in order to succeed. It had already succeeded, mesocosmically, before any symbolic destination was proposed for the dog at all.

This licenses a mischievous question worth keeping, provided it is understood correctly. What if Panofsky could not afford to let the dog be a dog. The question is not a claim about Panofsky’s private motives, and nothing here should be read as an accusation of bad faith on his part or anyone else’s. It is a structural claim about how a discipline organized around discovering hidden meaning tends to allocate its attention. Mesocosmic fit closes interpretation relatively quickly. This is a small domestic dog in an affluent household, and one can investigate its breed, its ownership, and the material culture surrounding it without generating an inexhaustible hermeneutic problem. Turn the dog into Fidelity, however, and the field expands considerably, drawing in Christian symbolism, marriage iconography, comparative examples from other paintings, and disputes among scholars about all of the above. Intersymbolic surplus generates scholarly surplus in a way mesocosmic identification simply does not, and a discipline whose expertise depends on discovering what an ordinary viewer cannot see has a structural incentive, whether or not any individual scholar experiences it as such, to prefer readings that generate more of that surplus over readings that close it down. This helps explain something otherwise puzzling: why disciplines built around interpretation so consistently privilege hidden symbolic meanings over the considerably harder to narrate, but no less real, achievement of a distributed mesocosmic accomplishment. The ordinary viewer can see the dog. The expert can tell you what the dog really means, and the painting becomes correspondingly deeper exactly as the expert becomes correspondingly more necessary to understanding it.

XII. From Bounded to Open Symbolic Surplus

The historical trajectory this argument opens onto deserves to be named, though briefly, since the article’s weight belongs with Van Eyck rather than with what came after him.

The more precise trajectory is not a simple story of sacred giving way to secular. It runs instead from constrained intersymbolic destination, in which nearly every mesocosmic detail owes its presence to an inherited symbolic programme, through increasing mesocosmic autonomy, in which a depicted object can stand on its own mesocosmic feet without forfeiting the option of also carrying symbolic weight, toward what might be called open symbolic surplus, in which a viewer’s question changes from which established relation does this object instantiate to the considerably more open does this object symbolize anything at all. Christianity does not need to disappear before the dog can become just a dog. The Passion scenes still frame the Arnolfini mirror. What has changed is that the room no longer needs them there in order to justify its own existence.

Later painting pushes this openness further still, and Van Gogh’s chairs make a useful, brief endpoint rather than a further case study this article has room to develop properly. A chair in one of his rooms can certainly be related across his body of work, connected to his biography, compared with Gauguin’s chair, and drawn into a network of symbolic association. But what does this chair symbolize no longer functions as the automatically privileged question the way it would for a medieval attribute. The chair can matter through colour, through the pressure of the brushwork, through the way it organizes the surrounding space, through its own material presence as an object someone once sat in and then left empty. The mesocosmic object has acquired, by this point, a pictorial autonomy the Arnolfini dog was only beginning to make available four centuries earlier.

XIII. Conclusion: Somewhere One Could Have Been

It is tempting to compress everything this article has argued into a single phrase, the world is the hero, and the phrase is not wrong exactly. But it borrows too much from the vocabulary of character and plot, and this painting has no plot and needs no hero to make its claim. The more precise way to put it is this. What is actually at stake in the Arnolfini Portrait, and in the interpretive habits it has generated for six centuries, is never any single object’s meaning. It is the sustained, coherent, inhabitable viability of an entire world, achieved across embodiment, sociality, materiality, dwelling, and symbol together, an achievement no single detail within the room could carry alone and none is required to.

Van Eyck painted a room that cannot exist exactly as depicted and that has felt, to nearly everyone who has ever stood before it, like a room they could walk into. He painted a couple whose joined hands may or may not record a specific legal or sacramental act, and a household whose ordinary possessions may or may not carry an inherited symbolic charge. He painted a mirror that behaves like evidence while remaining entirely under his own control, and signed the whole arrangement with a sentence that places its own maker inside the world he made. None of these achievements depends on the others resolving in any particular direction.

The dog may mean fidelity. It may have carried that association for contemporary viewers, and nothing here denies it. But the painting does not depend upon it. The dog already belongs, mesocosmically, before any symbolic destination is proposed for it. So do the oranges, the shoes, the bed, the chandelier, the fur, and the people. Their world does not await symbolic completion before becoming pictorially sufficient. Van Eyck’s achievement was to make that ordinary world inhabitable with an intensity that made an extraordinary new claim possible, a claim no painter of an Annunciation could have made about the room his own figures stood in. He did not merely sign the picture. He located himself inside its world. Johannes de eyck fuit hic. Jan van Eyck was here.

Before the viewer asks what this world means, Van Eyck has made a more fundamental claim: it is somewhere one could have been.