<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Digital Romanticism on Medium]]></title>
        <description><![CDATA[Stories by Digital Romanticism on Medium]]></description>
        <link>https://medium.com/@digitalromanticism?source=rss-24573cc2a3b2------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*KQzWZ_5kE3gI3eMuCD8r1w.png</url>
            <title>Stories by Digital Romanticism on Medium</title>
            <link>https://medium.com/@digitalromanticism?source=rss-24573cc2a3b2------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 15:17:52 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@digitalromanticism/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Blind Architecture]]></title>
            <link>https://medium.com/@digitalromanticism/blind-architecture-78d48ce7b309?source=rss-24573cc2a3b2------2</link>
            <guid isPermaLink="false">https://medium.com/p/78d48ce7b309</guid>
            <category><![CDATA[embodied-cognition]]></category>
            <category><![CDATA[design-philosophy]]></category>
            <category><![CDATA[architecture]]></category>
            <category><![CDATA[architectural-theory]]></category>
            <category><![CDATA[digital-architecture]]></category>
            <dc:creator><![CDATA[Digital Romanticism]]></dc:creator>
            <pubDate>Wed, 28 Oct 2026 23:24:35 GMT</pubDate>
            <atom:updated>2026-10-28T23:30:43.258Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/918/1*c8ft5rcTUIeeOVPbb44Uhw.jpeg" /><figcaption>Chris Downey, architect. Photograph: Azure Magazine.</figcaption></figure><h3><strong>Architecture and Music</strong></h3><p>During a recent review, a non-architecture student presented a project exploring the relationship between music and architecture. We examined musical notation alongside architectural examples, comparing the arrangement of notes, rhythms, and intervals with the repetition of windows, columns, and structural bays. The analogy is familiar, yet compelling. Both disciplines operate through proportion, repetition, variation, and sequence. But as we discussed the drawings, something about the comparison remained unresolved.</p><p>A musician can read a score and imagine a composition without hearing it performed. Beethoven, despite his progressive hearing loss, continued composing some of his most remarkable works after becoming profoundly deaf. Musical composition could proceed through notation, memory, and an understanding of relationships without requiring immediate sensory confirmation. This raises similar questions for architecture, another art concerned with proportion and composition.</p><p>Architecture has its own notational systems — plans, sections, elevations, and diagrams — yet we remain remarkably dependent on seeing what we design. We produce drawings to communicate spaces and building elements, renderings to anticipate their appearance, and increasingly AI-generated images to explore formal possibilities. Even our most advanced computational tools tend to return us to the screen.</p><p>Architecture is perhaps one of the most embodied of the arts, yet its design culture remains overwhelmingly visual. We have developed sophisticated methods to represent buildings before they exist, but comparatively limited means to anticipate the bodily experience of inhabiting them. The problem is not cognition itself: movement, touch, hearing, balance, and orientation also involve cognition. Rather, it is the dominance of visual cognition in how architecture is conceived, communicated, and judged.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yqfTQUklerzE9zeEIve1NQ.png" /><figcaption>Ludwig van Beethoven, manuscript from Symphony №9(1822–1824). Musical notation allows composition to be conceived independently of its immediate performance.</figcaption></figure><h3><strong>Goethe’s essay on architecture</strong></h3><p>In 1795, Johann Wolfgang von Goethe proposed a remarkable thought experiment in his unfinished reflections on architecture, <em>Baukunst</em>. He questioned the assumption that architecture, as a fine art, should principally address the eye. Instead, he suggested that a well-proportioned building should also produce a pleasurable experience through bodily movement.</p><p>“One might think that architecture as a fine art works for the eye alone; yet it should work above all — and this receives the least attention — for the sense of the mechanical movement of the human body. We experience a pleasant sensation when we move in a dance according to certain laws; we ought to be able to arouse a similar sensation in someone whom we lead blindfolded through a well-built house.” <em>— Johann Wolfgang von Goethe, Baukunst (1795), translated from the German.</em></p><p>Goethe’s blindfolded visitor introduces an extraordinary proposition: architectural proportion should remain perceptible even when visual information is removed. He compared this experience to dancing according to a set of rules, suggesting that proportion could be encountered through bodily motion rather than merely observed.</p><p>We usually understand proportion through drawings. We compare the width of a room to its length, the diameter of a column to its height, or the intervals between openings on a façade. Yet architectural experience involves relationships that unfold through movement. The organization of a plan determines how we approach, enter, traverse, and eventually comprehend a building.</p><p>These experiences cannot be fully conveyed through a single drawing. They require the body to participate in the construction of architectural knowledge. Goethe’s analogy with dance is particularly revealing: choreography can be written as instructions or diagrams, but it only becomes an experience through movement. Perhaps architecture operates similarly. Its proportions are not merely relationships to observe, but conditions to perform.</p><h3><strong>Renaissance and Representation</strong></h3><p>The Renaissance established a powerful intellectual project through which architecture could be conceived, measured, and communicated before construction. From Alberti onward, drawing increasingly distinguished the architect’s intellectual work from the physical labor of building. Plans, sections, elevations, and the study of proportion enabled architecture to operate as an abstract discipline, governed by geometrical relationships that could be examined independently of their material realization.</p><p>This was an extraordinary achievement. It allowed architecture to become increasingly systematic and communicable through a shared visual language. It also helped establish the architect’s authority as the author of a design. Although the technologies changed, the underlying relationship remained remarkably consistent: architecture could be conceived through representations of buildings that had yet to exist.</p><p>Today, digital simulations can anticipate complex geometries, structural behavior, and environmental performance. Rendering engines produce nearly photographic images of imagined environments. Yet the experience of inhabiting those environments remains difficult to communicate. The issue is not that these media are unnecessary. They are indispensable instruments of architectural knowledge. The problem emerges when we mistake the knowledge they provide for architecture itself even virtual reality often emphasizes visual immersion over bodily movement and multisensory perception.</p><p>Our design culture reinforces this tendency. Architecture is taught, reviewed, awarded, and increasingly circulated through drawings and images. The ability to mentally reconstruct space from visual information becomes a professional skill, while the bodily understanding of architecture is frequently postponed until the building is occupied. The interior exposes how much of architecture is temporal and embodied, and how little of that experience survives in a static image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rcDZJ1XSAiaLwxC07SLeoA.png" /><figcaption>Leon Battista Alberti, façade of Santa Maria Novella, Florence (c. 1458–1470). Geometric analysis illustrating the proportional relationships underlying its Renaissance composition.</figcaption></figure><h3><strong>The blind architects</strong></h3><p>After the review, I began looking into whether there were practicing blind architects. Two remarkable figures emerged from this search.</p><p>American architect Chris Downey lost his sight in 2008 following surgery for a brain tumor. Rather than abandoning architectural practice, he developed alternative methods for reading, modifying, and communicating drawings through embossed plans and flexible wax sticks. Downey describes experiencing architecture through acoustic changes, material transitions, bodily orientation, and the feedback of his cane as he moves through space.</p><p>Portuguese architect Carlos Mourão Pereira, who became blind in 2006, similarly continued his practice while exploring multisensory architectural environments. His work investigates how touch, sound, movement, and orientation contribute to spatial understanding. Interestingly, Pereira encouraged Downey to continue practicing by invoking Beethoven’s ability to compose music despite losing his hearing.</p><p>Their practices challenge the assumption that architectural thinking must begin with a visual image. Neither architect works without cognition. On the contrary, tactile representations and bodily exploration require sophisticated memory, abstraction, and spatial reasoning. What changes is the sensory foundation through which architectural knowledge is developed.</p><p>The question is therefore not whether a blind person can practice architecture. Downey and Pereira have already demonstrated that possibility. The more uncomfortable question concerns our own disciplinary assumptions: why do architectural schools continue to treat visual drawings and renderings as the principal evidence of design intelligence?</p><h3><strong>Architecture as a score</strong></h3><p>This brings us back to the relationship between music and architecture.</p><p>A musical score is not a picture of sound. It is a system of instructions through which relationships can be communicated and performed. Similarly, architectural notation could be understood not simply as a representation of completed geometry, but as a means of specifying operations, sequences, and transformations.</p><p>What if we began by describing architectural movement rather than its visual outcome?</p><p>Computational scripting makes this possibility particularly interesting. Unlike conventional drawings, a script can encode the rules through which geometry is generated, modified, and organized. In this sense, computation offers an alternative to the representational tradition.</p><p>Yet much contemporary computational practice remains tied to visual judgment. Scripts generate geometry, geometry produces models, and models return to the screen for evaluation. Generative AI has intensified this tendency, making it possible to produce convincing architectural appearances almost instantly while leaving their spatial and bodily consequences largely unexplored.</p><h3><strong>Beyond the image</strong></h3><p>Beethoven continued composing through accumulated musical knowledge, memory, and structural understanding. Similarly, chef Grant Achatz continued developing dishes at Alinea after cancer treatment deprived him of his sense of taste, relying on experience, technique, and collaboration. These examples reveal how creative intelligence is shaped by embodied knowledge that can persist even when immediate sensory access changes.</p><p>Architecture might learn from this distinction. We often treat design as an intellectual process of visual prediction and abstract problem-solving that precedes physical realization. The body enters later, when the proposal is built, occupied, or tested.</p><p>But what if embodied experience were not the eventual outcome of design, but one of its starting points?</p><p>Goethe’s blindfolded visitor presents a demanding challenge. If architectural proportion can be meaningful through movement without sight, why should its conception remain so dependent on visual representation?</p><p>Perhaps the question is no longer whether a blind architect can design architecture. It is whether architecture itself has become blind to the experiences that its representations cannot capture.</p><p>And perhaps the next step for computational design is not another machine capable of producing images of buildings, but a notation through which architecture can be conceived, communicated, and performed before it is ever seen.</p><h3>Further Reading</h3><ul><li>Johann Wolfgang von Goethe, <a href="https://www.projekt-gutenberg.org/goethe/baukunst/baukunst.html"><em>Baukunst</em></a> (1795).</li><li>Juhani Pallasmaa, <em>The Eyes of the Skin: Architecture and the Senses</em> (1996).</li><li>Steen Eiler Rasmussen, <em>Experiencing Architecture</em> (1959).</li><li>Chris Downey, <a href="https://www.ted.com/talks/chris_downey_design_with_the_blind_in_mind"><em>Design with the Blind in Mind</em></a> (TED Talk, 2013).</li><li>Juan Calduch Cervera and Alberto Rubio Garrido, <a href="https://doi.org/10.1007/978-3-319-58856-8_95"><em>Baukunst: Goethe’s Notes for a Treatise on Architecture</em></a> (2018).</li></ul><h3>Notes</h3><p>Goethe’s discussion of the blindfolded visitor comes from his unfinished <em>Baukunst</em> (1795). The English quotation is translated from the original German.</p><p>This essay is part of <em>Digital Romanticism</em>, an ongoing series exploring architecture, morphology, technology, and computational design.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=78d48ce7b309" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Unfinished Novelty of Symmetry]]></title>
            <link>https://medium.com/@digitalromanticism/the-unfinished-novelty-of-symmetry-ad03803b83b5?source=rss-24573cc2a3b2------2</link>
            <guid isPermaLink="false">https://medium.com/p/ad03803b83b5</guid>
            <category><![CDATA[digital-architecture]]></category>
            <category><![CDATA[willambateson]]></category>
            <category><![CDATA[symmetry]]></category>
            <category><![CDATA[greg-lynn]]></category>
            <category><![CDATA[morphology]]></category>
            <dc:creator><![CDATA[Digital Romanticism]]></dc:creator>
            <pubDate>Thu, 01 Oct 2026 16:19:40 GMT</pubDate>
            <atom:updated>2026-10-01T16:20:55.915Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1011/1*AOZaBUF5NXp9zMny0UGyyQ.jpeg" /><figcaption>Assorted Coleoptera in the University of Texas Insect Collection</figcaption></figure><h3><strong>Greg Lynn, William Bateson, and a road not taken in digital architecture</strong></h3><p>By the end of the twentieth century, architecture had already spent several decades reconsidering the formal (un)certainties of modernism. Postmodernism symptomatically brought back architectural history, typology, figuration, composition, and conventional forms back into disciplinary debate. Symmetry, type, precedent, and variation were no longer simply residues of an obsolete classical tradition; they had become available again as architectural problems.</p><p>The arrival of digital modeling in the late 1980s and 1990s destabilized this renewed interest in form almost immediately. What Mario Carpo would later describe as the “first digital turn” introduced a radically different condition in which geometry could become increasingly malleable. Splines, animation, morphing, topology, and continuous deformation allowed architects to manipulate forms that were difficult to describe through the familiar vocabulary of fixed types and discrete architectural elements.</p><p>This new freedom created a theoretical problem as much as a technical one. Architects could suddenly produce unprecedented geometries, but the discipline still needed to explain why those geometries mattered. The architectural literature of the period consequently became preoccupied to justify folds, blobs, smoothness, complexity, continuity, and deformation through philosophy, biology, mathematics, and new models of space. The computer had made form increasingly fluid, while architectural theory struggled to establish a language adequate to that fluidity.</p><p>Greg Lynn was one of the figures most consistently preoccupied with this problem. His work did not simply advocate the computer as a new drawing machine. Within that context, his essay “The Renewed Novelty of Symmetry” remains unusually fresh. Rather than justifying digital form through a generalized rhetoric of fluidity, Lynn returned to a much older and stranger source: the nineteenth-century biologist William Bateson.</p><h3>Variation, mutation, and the strange return of symmetry</h3><p>Bateson’s biological studies appear at first to belong far outside architectural discourse. Materials for the Study of Variation, published in 1894, documented organisms exhibiting duplication, altered segmentation, unusual bodily correspondences, and other departures from expected morphology. These anomalies were valuable precisely because they exposed organizational relationships that remained difficult to see in normal specimens.</p><p>What interested Lynn was Bateson’s treatment of symmetry and differentiation. Symmetry, in this framework, does not always indicate greater organization or perfection. Under certain morphological conditions, increased symmetry can accompany a loss of differentiation. Parts that would normally develop distinct identities may instead become duplicated, mirrored, or generalized. Conversely, asymmetry can signal the acquisition of new information as initially related parts become progressively differentiated.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*04zrbmCwEVQApPPGri0X5g.jpeg" /><figcaption><em>William Bateson, cases of syndactylism from Materials for the Study of Variation (1894), Fig. 104. The anomalies show changes in duplication and differentiation that made symmetry a developmental problem rather than a fixed ideal.</em></figcaption></figure><p>This reverses one of architecture’s most familiar intuitions. Symmetry is usually associated with order, while asymmetry appears as its disturbance. Bateson allows the relation to be understood developmentally instead.</p><p>That distinction also separates two different concepts of variation.</p><p>A form can vary continuously by stretching, bending, scaling, twisting, or otherwise deforming while its underlying organization remains stable. D’Arcy Thompson’s famous transformation grids offer a powerful model of this kind of morphological continuity. Bateson draws attention to a different phenomenon: variation can also alter the organization of the body itself through duplication, discontinuity, asymmetry, or changed relations among parts.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/629/1*jWm3NSr0sF6squ2oSgjEyA.jpeg" /><figcaption><em>D’Arcy Wentworth Thompson, transformation grids comparing related fish forms in On Growth and Form (1917). Unlike Bateson’s emphasis on discontinuity and mutation, Thompson models variation through continuous geometric deformation.</em></figcaption></figure><p>One transforms geometry. The other can transform identity.</p><p>This distinction became particularly productive for digital architecture because the new computational tools of the 1990s were extraordinarily capable of the former. Digital geometry could easily be stretched, morphed, animated, and continuously transformed. Bateson suggested that the more radical question might concern not continuous deformation itself, but the moments when a transformation reorganizes the relations that constitute a form.</p><h3>Cardiff and the renewed novelty of symmetry</h3><p>Lynn explored this problem directly in his proposal for the Cardiff Bay Opera House competition in 1994.</p><p>The competition brief contained strong formal and programmatic expectations, including a symmetrical horseshoe-shaped performance hall and an important relationship to the Oval Basin of the site. Rather than simply accept symmetry as a fixed compositional condition or reject it in favor of an arbitrary irregular form, Lynn treated it as something capable of developmental differentiation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*CFWV4KQyfBqc1NrbMBmBhg.png" /><figcaption><em>Greg Lynn, Cardiff Bay Opera House, analytical and structural studies, 1994. The diagrams translate variation into changes of organization, structure, and relation rather than treating deformation as an isolated surface effect.</em></figcaption></figure><p>The resulting project organized a series of oval and curvilinear bodies whose relations could be inflected, displaced, differentiated, and transformed. Symmetry did not disappear; it became dynamic. Related parts could remain connected without remaining identical, allowing organizational difference to emerge from an initially shared condition.</p><p>This is what makes the Cardiff project more interesting than its status as an example of early digital form. The project does not use biological language merely to justify a biomorphic appearance. Bateson provides an actual conceptual mechanism through which architectural symmetry can be reconsidered.</p><p>For me, this remains one of the most intellectually ambitious moments of the first digital turn. At a time when architects were producing a considerable amount of theoretical writing to explain newly available digital forms, The Renewed Novelty of Symmetry offered something unusually precise. Rather than treating the computer itself as the source of novelty, Lynn used computation to test an existing intellectual problem and then allowed the architectural consequences to emerge from that encounter.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/876/1*fJ2Ldm5i4TFzHYrzjK78Vw.png" /><figcaption><em>Greg Lynn, Cardiff Bay Opera House competition proposal, 1994. The project treats symmetry as a mutable relation among differentiated oval and curvilinear bodies rather than as a fixed compositional order.</em></figcaption></figure><h3>The novelty that remained unfinished</h3><p>Yes. The section is conceptually right, but it currently stops short of fully explaining <strong>what remained unfinished</strong>. I’d make the Thompson/Bateson contrast explicit, then let AI sharpen rather than merely repeat the problem.</p><h3>The novelty that remained unfinished</h3><p>Digital architecture became extraordinarily successful at continuous variation. Through the late 1990s and 2000s, increasingly sophisticated modeling systems allowed architects to generate families of related forms, manipulate surfaces parametrically, maintain topological continuity, and control differences across large populations of components. Variation became one of computation’s defining architectural capabilities.</p><p>Yet the Batesonian side of Lynn’s argument remained comparatively underdeveloped. The discipline became increasingly adept at deforming forms without developing an equally precise account of when deformation becomes a change in organization. In this sense, digital architecture largely followed the trajectory suggested by D’Arcy Thompson: bodies could be stretched, bent, morphed, and continuously transformed while their underlying relational structure remained legible. Bateson posed a different problem. His mutations involved discontinuity, duplication, asymmetry, and changes in the relation of parts to the whole. They asked not simply how much a form could vary, but when variation produced a different organization.</p><p>The emergence of artificial intelligence makes this unfinished problem even more visible. Carpo’s second digital turn describes a computational culture increasingly structured by data, machine learning, prediction, and automation. Today the production of variation itself is becoming almost trivial. Generative models can produce hundreds of buildings, images, arrangements, and formal alternatives in the time it once took an architect to construct a single digital model.</p><p>But greater generative capacity does not automatically produce a more sophisticated theory of variation. AI can multiply difference at extraordinary speed, yet it does not necessarily distinguish between a stylistic modification, a continuous deformation, a broken symmetry, a duplicated part, or a genuine reorganization of the whole. In other words, the quantity of variation has increased dramatically while the disciplinary problem of how to qualify difference remains largely intact.</p><p>This may be why returning to Bateson and Lynn now feels unexpectedly productive. The first digital turn demonstrated that computation could become more than an instrument of representation; at its most ambitious, it could function as a theoretical apparatus through which established architectural concepts were reopened and tested.</p><p>The second digital turn has given architecture vastly more computational power, but it has not automatically advanced those questions. If anything, the abundance of generative variation makes them more urgent.</p><p>The challenge today is no longer to prove that computers can produce formal novelty. They clearly can.</p><p>The unfinished question is how to distinguish novelty from variation, mutation from deformation, and meaningful organizational difference from the endless proliferation of alternatives.</p><p>Lynn called for a renewed novelty of symmetry in the 1990s. Three decades later, that novelty still feels unresolved.</p><h3>Further Reading</h3><p>· William Bateson, Materials for the Study of Variation</p><p>· Greg Lynn, “The Renewed Novelty of Symmetry”</p><p>· Greg Lynn, Animate Form</p><p>· D’Arcy Wentworth Thompson, On Growth and Form</p><p>· Mario Carpo, The First Digital Turn</p><p>· Mario Carpo, The Second Digital Turn</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ad03803b83b5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Before Organic Form Became a Style]]></title>
            <link>https://medium.com/@digitalromanticism/before-organic-form-became-a-style-4d7d7d320274?source=rss-24573cc2a3b2------2</link>
            <guid isPermaLink="false">https://medium.com/p/4d7d7d320274</guid>
            <category><![CDATA[architectural-history]]></category>
            <category><![CDATA[organic-architecture]]></category>
            <category><![CDATA[architecture]]></category>
            <category><![CDATA[morphology]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Digital Romanticism]]></dc:creator>
            <pubDate>Tue, 08 Sep 2026 16:58:49 GMT</pubDate>
            <atom:updated>2026-09-08T17:00:39.635Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jEQlczixTFkRC-iHjtxP_Q.jpeg" /><figcaption>Detail of Rudolf Steiner’s Second Goetheanum, Dornach, Switzerland.</figcaption></figure><h3>The other history of organic architecture</h3><p>Long before the computer made irregular geometry easy to model, architects were already trying to escape the box.</p><p>The twentieth century is filled with buildings described as organic, expressive, plastic, continuous, bodily, or natural, yet these terms rarely meant exactly the same thing. For Louis Sullivan, organic unity involved the relation of structure, function, and ornament. Frank Lloyd Wright extended the idea toward the integration of building, material, interior, landscape, and site. Antoni Gaudí found architectural possibilities in natural geometry, structural equilibrium, craft, and growth. Elsewhere, the organic came to describe buildings that seemed vegetal, geological, anatomical, or simply resistant to regular composition.</p><p>Organic architecture, in other words, was never a style in any simple sense. It was a persistent attempt to rethink how architectural form might acquire unity without relying entirely on standardized parts or predetermined geometries.</p><p>By the time digital tools entered architectural discourse, non-standard form was already associated with a longer history of experiments in organic geometry. The curving surfaces and continuous bodies that emerged so forcefully in the 1990s appeared inseparable from splines, morphing, topology, animation, and later parametric modeling. Yet many of the formal ambitions associated with digital architecture had been explored for decades through drawing, sculpture, craft, material experimentation, and highly individualized construction.</p><p>The digital did not invent non-standard form.</p><p>It changed the conditions under which it could be produced.</p><h3>Steiner and the architecture of metamorphosis</h3><p>Rudolf Steiner offers one of the most unusual episodes in this earlier history.</p><p>Before developing Anthroposophy, Steiner worked extensively with Goethe’s scientific writings and became deeply involved with his theories of morphology. Goethe’s <em>Metamorphosis of Plants</em> described living form through transformation rather than fixed appearance. Expansion, contraction, polarity, differentiation, and serial development allowed individual organs to be understood as moments within a changing whole.</p><p>Steiner carried this morphological thinking into a far more speculative philosophical territory. Goethe’s disciplined engagement with observable nature became absorbed into Anthroposophy, an expansive spiritual worldview concerned with consciousness, education, medicine, art, agriculture, and human development. This connection requires a critical reading, as Goethean morphology does not naturally culminate in Anthroposophy, and Steiner’s metaphysical interpretations exceed the methodological restraint of Goethe’s science.</p><p>Yet his architectural translation of metamorphosis remains fascinating.</p><p>The First Goetheanum, begun in 1913, did not merely copy plants or animals. Its architectural elements were intended to differentiate through sequence. Columns and capitals changed progressively rather than repeating one order, while its two unequal domes formed a related but polarized pair. Ornament, structure, and interior articulation participated in an overall sculptural development in which one condition seemed to transform into another.</p><p>After the First Goetheanum was destroyed by fire, the Second Goetheanum translated this ambition into reinforced concrete. Steiner developed a continuous sculptural body in which wall, opening, roof, edge, and mass appear to merge, and whose geometry resists easy decomposition into familiar architectural elements.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*MRqjeC6uy9xKutCNJPbCKA.jpeg" /><figcaption>Rudolf Steiner with a model of the First Goetheanum, 1914. The project translated Goethean ideas of metamorphosis into a continuous sculptural language.</figcaption></figure><p>Steiner’s architecture therefore exposes an important ambiguity. Its conceptual origin lies in morphology, yet metamorphosis largely arrives architecturally as <strong>formal movement</strong>. The building looks as though its parts are transforming, but its tectonic and structural system is not itself generated by a morphological process.</p><h3>An organic architecture before computation</h3><p>Steiner was not alone in trying to move beyond conventional architectural form. Erich Mendelsohn’s Einstein Tower, Hans Poelzig’s cavernous interiors, Hermann Finsterlin’s visionary biomorphic projects, and later Frederick Kiesler’s Endless House all treated architecture as something closer to a continuous body than a composition of discrete elements.</p><p>These architects emerged from different philosophies and should be understood in their own terms. What they reveal collectively is a broader early twentieth-century dissatisfaction with standard form.</p><p>The same tendency continued in a more explicitly Anthroposophical architectural culture after Steiner. Erik Asmussen, Anton Alberts, Max van Huut, Imre Makovecz, Joachim Eble, Gregory Burgess, and others developed very different versions of an architecture based on differentiation, bodily analogy, craft, symbolism, landscape, and irregular geometry. These practices often remained peripheral to conventional histories of twentieth-century architecture, where they could appear as regional curiosities, spiritual experiments, or idiosyncratic alternatives to mainstream modernism.</p><p>Seen together, however, they form a surprisingly persistent architectural territory.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*86LRKjKJn0tOPTJyNvgKAA.png" /><figcaption>Book spread showing Imre Makovecz’s Lutheran Church in Siófok (1987) and Reduta Cultural Center in Szigetvár (1988), Hungary. From Anna Sokolina, ed., <em>Architecture and Anthroposophy</em>, pp. 248–249.</figcaption></figure><p>Their work demonstrates that architecture did not need digital technology to become formally non-standard. Roofs could resemble wings, columns could become tree-like bodies, walls could bend, plans could resist orthogonal organization, and architectural elements could transform continuously through craft and sculptural manipulation.</p><p>But this is also where organic architecture encounters a recurring problem: the translation between nature and architecture remains fundamentally ambiguous. Nature does not provide a single, clearly defined set of architectural rules. Its forms and processes must be selected, interpreted, and translated by the architect, and that translation is never neutral.</p><p>Once architectural elements become recognizable signs of nature, the organic itself risks becoming another formal vocabulary. Without a clearly defined relationship between natural process and architectural formation, organic architecture can therefore slide into subjective formalism: nature becomes a source of visual permission rather than a constraint on how form is generated.</p><p>The question shifts from whether architecture <em>looks alive</em> to how its form actually comes into being, and what, if anything, governs the translation between natural processes and architectural form.</p><h3>Frei Otto and the turn toward formation</h3><p>A striking perspective on organic architecture can be found in Frei Otto’s body of work. His soap films, hanging chains, tensile structures, and lightweight experiments do not begin by drawing a desired organic shape. Instead, they establish boundaries, forces, materials, and constraints from which geometry emerges through physical equilibrium.</p><p>The final form is therefore inseparable from the procedure that produced it.</p><p>This distinction is close to what Detlef Mertins later described through <strong>bioconstructivism</strong>, which shifts attention from imitating natural forms toward investigating the constructive and formative processes found in nature.</p><p>This gives us two very different organic approaches.</p><p>Steiner attempts to express Goethe’s idea of metamorphosis through a sequence of sculptural forms. His buildings suggest that one architectural element can develop into another through gradual variation, but the process remains primarily formal and representational: the building gives visible shape to the idea of transformation.</p><p>Otto, by contrast, allows architectural geometry to result from a formative process. In his experiments, the final shape is not designed first and then made to look organic. It emerges from the interaction of forces, materials, boundaries, and constraints.</p><p>With Steiner, metamorphosis is expressed through the appearance and sequence of architectural forms. With Otto, formation becomes the operative mechanism through which form is produced. The organic therefore moves away from the representation of movement and toward the actual generation of geometry through forces and relations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EnpDPLZdLa7r290PobsWoA.png" /><figcaption>Frei Otto, hanging-chain model for the Mannheim Multihalle, 1973. The final roof geometry was determined through physical form-finding and later measured photogrammetrically for engineering development. Image: Deutsches Architekturmuseum / Uwe Dettmar, CC BY-NC-SA 4.0.</figcaption></figure><h3>When the organic became digital</h3><p>By the 1990s, digital modeling transformed the status of non-standard geometry.</p><p>Spline surfaces, topology, animation, morphing, and parametric deformation allowed architects to manipulate continuous and irregular forms directly in three-dimensional space. Geometries that had previously demanded extraordinary drawing skill, physical models, sculptural intuition, or complex craft could increasingly be controlled, revised, multiplied, and eventually fabricated through digital means.</p><p>The computer did not make architecture organic.</p><p>It made <strong>non-standard form manageable</strong>.</p><p>Once almost any curvature could be modeled, irregularity ceased to indicate a particular theory of nature, morphology, or material process. Sculptural form could be produced for reasons ranging from structure and program to branding, formal experimentation, or visual effect. The digital turn liberated architecture from many geometric constraints, but it also separated non-standard form from the intellectual frameworks that had previously given organic architecture much of its meaning.</p><p>And today, generative AI pushes that transformation one step further.</p><p>Ask an image model for “organic architecture,” “biomorphic architecture,” “parametric architecture,” or simply “futuristic architecture,” and a remarkably familiar family of images often appears: flowing shells, white curving surfaces, tapered openings, continuous ribbons, and compositions that drift toward a generalized version of Zaha Hadid Architects.</p><p>This is not a criticism of Zaha Hadid’s work. It is evidence of how thoroughly a particular language of non-standard architecture has become visually codified.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TaTNHgdROgeRTWPiEg0aoA.jpeg" /><figcaption><strong>AI-generated design studies by Zaha Hadid Architects using DALL·E 2 and Midjourney.</strong> The office has incorporated generative image models into early-stage design exploration, creating a feedback loop between an established architectural language and machine-generated imagery. Source: <em>Dezeen</em>, 2023.</figcaption></figure><p>Architecture spent more than a century exploring nature, growth, movement, embodiment, morphology, and process as alternatives to predetermined form. Surprisingly, AI has learned much of this history primarily as <strong>appearance</strong>. Generative models can now compress a century of organic and non-standard architectural imagery into a visual category.</p><p><strong>Organic architecture has become a prompt.</strong></p><h3>After the organic image</h3><p>Steiner, the Anthroposophical architects, Kiesler, Mendelsohn, Otto, and later digital practitioners do not constitute one continuous organic tradition. Their architectures emerge from radically different intellectual, technological, and material conditions.</p><p>What connects them is a recurring attempt to loosen architecture from predetermined form. Sometimes that freedom is expressed through sculptural movement. Sometimes it is tied to landscape or bodily symbolism. Sometimes it emerges through physical forces and material behavior. Digital technology later made non-standard geometry increasingly independent of these commitments, while AI now makes its visual appearance almost trivial to reproduce.</p><p>Organic architecture can describe <strong>appearance</strong>, when a building resembles a plant, body, landscape, or fluid geometry. It can also describe <strong>organization</strong>, when differentiated parts belong to a coherent relational whole. But it may be most useful to reserve the term for <strong>formation</strong>, when geometry emerges through transformation, force, material behavior, growth, constraint, or other developmental processes.</p><p>This distinction brings the argument back, unexpectedly, to Goethe. Morphology was never primarily a search for shapes that looked natural. Its subject was transformation: how forms differentiate while remaining related through a process of becoming. In that sense, the organic may be less a property of architectural appearance than a way of understanding how form acquires coherence through change.</p><p>The question for the age of AI is therefore not whether machines can generate more convincing organic architecture. They clearly can. The more interesting question is whether computation can participate in processes of differentiation, adaptation, constraint, material feedback, and transformation rather than merely synthesizing images that resemble their results.</p><p>After organic form has become a style, the challenge may be to recover the processes that made the idea of the organic meaningful in the first place.</p><h3>Further Reading</h3><p>Caroline van Eck, <em>Organicism in Nineteenth-Century Architecture</em></p><p>Rudolf Steiner, <em>Ways to a New Style in Architecture</em></p><p>David J. Adams, “Rudolf Steiner’s First Goetheanum as an Illustration of Organic Functionalism”</p><p>Anna Sokolina, <em>Architecture and Anthroposophy</em></p><p>Frei Otto and Bodo Rasch, <em>Finding Form</em></p><p>Detlef Mertins, “Bioconstructivisms”</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4d7d7d320274" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Vitruvian Turn]]></title>
            <link>https://medium.com/@digitalromanticism/the-vitruvian-turn-55bfb4223992?source=rss-24573cc2a3b2------2</link>
            <guid isPermaLink="false">https://medium.com/p/55bfb4223992</guid>
            <category><![CDATA[architecture]]></category>
            <category><![CDATA[architectural-history]]></category>
            <category><![CDATA[computational-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[digital-culture]]></category>
            <dc:creator><![CDATA[Digital Romanticism]]></dc:creator>
            <pubDate>Wed, 26 Aug 2026 21:03:29 GMT</pubDate>
            <atom:updated>2026-08-26T21:06:31.978Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*si_H2SZ5DHrtIH5AaB6WmQ.jpeg" /><figcaption><em>Aby Warburg, Bilderatlas Mnemosyne — The Original, 2020. (Photo: Silke Briel / HKW.</em></figcaption></figure><h3><strong>Why the Image Is the Wrong Starting Point for Architectural AI</strong></h3><p>Almost a century before the arrival of generative AI, Aby Warburg covered a series of large black panels with images. His unfinished <em>Bilderatlas Mnemosyne</em> assembled photographs of paintings, sculptures, diagrams, newspaper clippings, advertisements, maps, and other fragments into changing visual constellations. By the time of his death in 1929, the final version contained 971 images distributed across 63 panels. Yet the Atlas was never intended as a stable catalogue. Warburg continually moved the images, changing their relationships in search of historical affinities that might become visible through juxtaposition.</p><p>There is something unexpectedly contemporary about this method. Warburg used images to think across historical distance. Forms migrated between antiquity and the Renaissance, between ritual and art, between gesture and representation. Meaning emerged not exclusively from individual objects but from their placement within a larger visual field.</p><p>Much of contemporary artificial intelligence operates within a vastly enlarged version of this condition. Generative models learn from enormous collections of images, constructing statistical relationships among appearances, concepts, styles, and visual structures. The scale is radically different, but the attraction is familiar: perhaps knowledge can emerge through the proximity of images.</p><p>For architecture, this possibility has been intoxicating.</p><h3>The architecture of the screen</h3><p>The rapid arrival of AI in architectural education has produced an interesting physical consequence. AI research spaces are increasingly built around enormous screens.</p><p>At the University of Texas at Arlington, the recently established Performative AI Lab combines machine intelligence with a curved high-resolution LED visualization wall, including a 25-foot AI-enabled display used for generated images, videos, and design feedback. Florida Atlantic University’s Creative AI Lab similarly centers an immersive 180-degree floor-to-ceiling screen alongside scanning, XR, and computational equipment.</p><p>These laboratories are doing much more than displaying pictures, of course. They connect visualization with mixed reality, simulation, and data. Yet the prominence of the screen is revealing.</p><p>The first great encounter between architecture and generative AI has been overwhelmingly visual. We prompt an image, evaluate an image, modify an image, project an image, and generate another image. Diffusion models are remarkably well suited to this culture because architecture already possesses an intense visual economy of renderings, photographs, drawings, references, competitions, portfolios, and precedents. AI arrived at precisely the point where architecture was already fluent.</p><p>In this sense, contemporary visual AI has what might be called a phenomenal bias. I use that term carefully. Generative images are not phenomenology in the philosophical sense of embodied experience, perception, and consciousness. They operate at a much thinner level: the synthesis of convincing appearances. Yet their extraordinary success makes the visible surface increasingly easy to confuse with the architectural object itself.</p><p>Perhaps the giant screen is therefore more than equipment. It is an emblem of the current computational moment: architecture looks at AI, and AI looks back through images.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WVFjTWNat5XEOO30GQrZ_w.avif" /><figcaption><em>Florida Atlantic University’s Creative AI Lab. Image courtesy: Florida Atlantic University.</em></figcaption></figure><h3>A Vitruvian interruption</h3><p>There is a strange historical counterpoint to this abundance of images. The foundational architectural treatise of antiquity reached later architects primarily through words.</p><p>Vitruvius wrote <em>De architectura</em> in the first century BCE, describing architecture through ideas of order, arrangement, proportion, construction, geometry, materials, machines, and representation. The status of the original illustrations remains uncertain because they have not survived intact. More remarkably, the first printed edition of Vitruvius, produced in Rome around 1486, contained no illustrations at all. The first extensively illustrated printed edition did not appear until Fra Giovanni Giocondo’s Venetian publication of 1511.</p><p>The point is not that architecture was somehow image-free. Vitruvius discusses plan, elevation, perspective, geometry, and drawing explicitly, and architecture has always depended on visual representation. What is striking is that architectural knowledge could travel without pictorial completeness because a building could be communicated through relations, dimensions, proportions, and geometric procedures. Text did not need to resemble architecture visually in order to participate in its production; it could describe a system sufficiently well for architectural order to be reconstructed.</p><p>This becomes particularly interesting when placed beside generative AI. Our most advanced computational systems can produce thousands of persuasive architectural images, yet they often struggle to return the explicit relations through which those images could become architectural objects.</p><p>This is where I would propose a <strong>Vitruvian turn</strong>: not a return to classical architecture, nor a rejection of images, but a return to architecture as description, to rule, relation, geometry, sequence, and construction.</p><h3>Gothic Diffusion</h3><p>I encountered this problem directly through a project I developed called <strong>Gothic Diffusion</strong>.</p><p>The experiment began with a curated collection of Gothic architectural material: sections, elevations, rose windows, tracery studies, vaults, ornamental fragments, and plans. These images were used to fine-tune an SDXL model through LoRA training, while ControlNet provided additional guidance through line, depth, and edge conditions. The workflow was orchestrated through ComfyUI and later extended toward multi-view generation, approximate image-to-mesh reconstruction, and resin printing.</p><p>The results were often surprisingly convincing. Rose windows proved particularly revealing. A matrix of generated examples immediately displayed the visual grammar the model had learned: radial symmetry, concentric organization, tracery, repetition, ornament, and variations that drifted between historical plausibility and controlled distortion. The Gothic language remained recognizable even when individual windows had never existed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FEbYqKlepOm6WVtAHy-wgA.png" /><figcaption><em>Gothic Diffusion, 2026. Generated rose-window matrix.</em></figcaption></figure><p>This is precisely where generative AI is extraordinary. It can synthesize the visible tendencies of a historical corpus without requiring those tendencies to be explicitly defined beforehand. I did not have to write a rule specifying how many lobes a rose window should contain, how tracery should branch, or how circles should intersect. The model acquired a probabilistic visual sense of those relationships from images.</p><p>Yet when I attempted to move from these images toward architectural geometry, the apparent intelligence began to dissolve. A generated rose window does not contain its geometric construction. There is no accessible compass sequence, no editable hierarchy of circles, no parametric relationship among radii, no explicit rotational symmetry, and no feature history through which the design might be systematically altered. Even when image-to-mesh systems reconstruct an approximate three-dimensional object, geometry is inferred from visual evidence rather than specified through architectural relations.</p><p>This gap becomes particularly clear when the same rose window is considered geometrically. Its construction can be described through operations: locating a center, establishing a radius, dividing a circle, rotating an element, intersecting arcs, repeating modules, generating secondary circles, and constraining relationships among them. The diffusion model captures appearance extraordinarily well, while the geometric description exposes construction. Neither representation is sufficient by itself, but architecture requires the second if generated forms are to remain editable, measurable, and relational after they appear.</p><p>This distinction suggests that the next stage of architectural AI cannot simply involve larger image models producing more spectacular renderings. A column must remain related to its grid, a window must retain its proportion, a wall must have thickness, a surface must remain connected to its boundary, and a design operation must be revisable without reconstructing the entire object from an image. The challenge is therefore to move from visual synthesis toward representations in which architectural relationships remain computationally accessible.</p><p>Gothic Diffusion became interesting to me precisely because of this limitation. It could learn how Gothic architecture <strong>looks</strong> before it could tell me how Gothic architecture <strong>works</strong>.</p><h3>After Warburg</h3><p>Warburg’s Atlas remains a relevant historical reference for how we use AI for design because there is nothing inherently problematic about thinking through images. His great insight was that images contain cultural memory and that their relationships can expose histories inaccessible to linear narrative. Generative AI extends this capacity at scales Warburg could never have imagined, allowing visual archives to be synthesized, reorganized, and destabilized with remarkable speed.</p><p>The danger begins only when architecture mistakes this visual intelligence for the entirety of computation.</p><p>The black panels of the <em>Mnemosyne Atlas</em> were instruments of inquiry, not substitutes for history. Perhaps the enormous screens appearing in contemporary AI laboratories should be understood similarly. They are windows into a new visual field, but they are not yet the architecture behind it.</p><p>The Vitruvian turn asks us to move through the image rather than abandon it. Architectural AI will eventually need to articulate the relations that images conceal: dimensions, constructions, symmetries, transformations, constraints, and sequences of operation. Generative AI has become remarkably good at imagining what architecture might look like; the more difficult task begins when we ask how that architecture comes into being.</p><p>The question is no longer whether a machine can picture architecture, but whether it can help us <strong>write it</strong>.</p><h3>Further Reading</h3><p>Aby Warburg, <em>Bilderatlas Mnemosyne</em></p><p>Vitruvius, <em>De architectura</em></p><p>Mario Carpo, <em>The Second Digital Turn</em></p><p>Robert Bork, <em>The Geometry of Creation</em></p><p>Robin Rombach et al., “High-Resolution Image Synthesis with Latent Diffusion Models”</p><p>Sabri Gokmen, <em>Gothic Diffusion</em>, 2026</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=55bfb4223992" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Digital Romanticism]]></title>
            <link>https://medium.com/@digitalromanticism/digital-romanticism-a2014ab62121?source=rss-24573cc2a3b2------2</link>
            <guid isPermaLink="false">https://medium.com/p/a2014ab62121</guid>
            <category><![CDATA[goethe]]></category>
            <category><![CDATA[romanticism]]></category>
            <category><![CDATA[morphology]]></category>
            <category><![CDATA[architecture]]></category>
            <category><![CDATA[digital-culture]]></category>
            <dc:creator><![CDATA[Digital Romanticism]]></dc:creator>
            <pubDate>Tue, 18 Aug 2026 15:32:43 GMT</pubDate>
            <atom:updated>2026-10-01T16:41:03.552Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Hv-5tWa1A8205t9CwuRoVA.png" /><figcaption>Romanticism at the edge of the digital sublime.</figcaption></figure><h3>A Technological Adventure of Reason</h3><p>Caspar David Friedrich’s <em>Wanderer above the Sea of Fog</em> presents a solitary figure at the edge of a landscape that resists complete comprehension. Mountains emerge from the mist only to disappear again, producing a world that seems simultaneously present and inaccessible. The wanderer surveys this terrain without mastering it, suspended between observation and speculation. The image accompanying this essay reimagines Friedrich’s scene for a different historical moment: the mountains have become meshes, the horizon dissolves into coordinates and signals, and the landscape itself appears computational. Yet the wanderer remains. For me, this is an appropriate point of departure for <strong>Digital Romanticism</strong>.</p><p>Digital Romanticism is not an attempt to revive Romanticism as a style or to dress contemporary technology in nineteenth-century imagery. It begins instead from the possibility that certain older questions about nature, form, transformation, imagination, and generation can acquire new relevance when examined through computation. More specifically, I am interested in what happens when historical ideas cease to be objects of interpretation alone and become operational propositions that can be diagrammed, reconstructed, varied, and tested.</p><h3>Computation without computers</h3><p>Computation is usually narrated through the history of machines, which makes it appear inseparable from the technological present. Architecture suggests a longer genealogy. Long before electronic computers, architects, geometers, and builders worked through procedures involving proportion, subdivision, repetition, rotation, projection, transformation, and rule. Gothic vaults developed through geometrical relationships among ribs, bays, surfaces, and supports; muqarnas systems produced extraordinary spatial differentiation through the organization and repetition of geometric units; Renaissance architectural thought established correspondences among bodies, measures, figures, and buildings.</p><p>It would be misleading to describe these practices as digital computation before the digital. What they reveal instead is that <strong>computational thinking can exist independently of computational machines</strong>. Geometry can operate procedurally: a line may be divided according to a relation, a module repeated under changing conditions, or a figure transformed through a sequence of prescribed operations. Computation, understood in this broader sense, concerns the explicit organization of relationships and the consequences that follow when operations are performed upon them.</p><p>This perspective changes how history and technology might be brought together. Rather than treating computation as a rupture separating contemporary architecture from its past, digital tools can become instruments for reopening historical systems of thought. Procedures embedded in drawings, buildings, geometric constructions, and theoretical texts can be reconstructed and subjected to forms of variation that would previously have been difficult or impossible to explore. History does not literally become software, but aspects of historical reasoning can become executable.</p><h3>Goethe and the rules of metamorphosis</h3><p>My interest in this possibility began with my doctoral research on Goethean morphology at Georgia Tech almost a decade ago. Goethe is often approached through his commitment to observation, intuition, and the irreducibility of living phenomena, qualities that have made his scientific writings particularly attractive to later phenomenological and humanistic interpretations. Yet <em>The Metamorphosis of Plants</em> revealed to me another dimension of his thinking that seemed considerably more systematic.</p><p>Goethe does not simply assemble and describe botanical specimens. He reads the plant as a sequence of transformations in which organs expand and contract, vegetative development passes into reproduction, and differentiated structures remain related to other states within the series. The identity of the plant is therefore not located in a single ideal image but in the relations that become visible across transformation. Form becomes intelligible through its capacity to change.</p><p>What interested me most was the presence of a small repertoire of principles guiding this investigation: polarity, expansion and contraction, seriality, correspondence, and intensification. These principles are not algorithms in the contemporary sense, but neither are they merely poetic metaphors. They organize Goethe’s observations, establish relationships among forms, and generate expectations that can be confronted with further phenomena. In this sense they behave remarkably like rules within an experimental mode of reasoning.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VNlgqDCfocM9BuqpVjBbbA.png" /><figcaption>The six cyclic stages of development in flowering plants showing the alternation of expansion and contraction.</figcaption></figure><p>My research began by asking what would happen if these relationships were translated into diagrams and eventually into computational operations. Expansion and contraction, for example, could become explicit transformations whose consequences could be repeatedly tested across a sequence. What had originally appeared as a philosophical or morphological proposition could then acquire a generative dimension. The framework that emerged from this work consequently treated forms as serial phenomena, identified polar relations, traced transformations and invariants, and used diagrams and algorithms to expose relationships that were otherwise difficult to see.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FLAtDjhm8LM9EItmns0v2w.png" /><figcaption>Computational reconstruction of a common buttercup (Ranunculus Acris) leaf through alternating stages of expansion and contraction, based on the metamorphic rules developed in <em>Metamorphic Leaves</em>.</figcaption></figure><p>Computation, from this perspective, does not replace observation with abstraction. It provides another medium through which the consequences of observation can be explored.</p><h3>The adventure of reason</h3><p>This problem reaches back to one of the central philosophical difficulties of the eighteenth century. Kant recognized that living organisms presented a peculiar challenge to mechanical explanation because their parts seemed to acquire meaning only through their relation to an organized whole. He consequently remained skeptical that human beings could ever possess a Newtonian science capable of fully explaining the production of even something as apparently simple as a blade of grass.</p><p>Yet Kant also entertained the possibility that apparently different organic forms might be related through common principles of transformation. He characterized this speculative undertaking as an <strong>“adventure of reason.”</strong> Goethe’s morphology can be understood as one attempt to enter precisely this territory. Rather than regarding formative nature as permanently inaccessible, he tried to approach it through disciplined comparison, observation, and the reconstruction of transformations across visible forms.</p><p>Here computation introduces an unexpected possibility. If observation allows relationships to be inferred, a computational model can allow those relationships to operate under altered conditions. A proposition about expansion and contraction can be extended beyond the original botanical sequence; an invariant can be tested against increasingly extreme variations; continuity can be pushed until a transformation produces rupture. Such experiments do not demonstrate that an algorithm has uncovered nature’s hidden machinery. Their value is epistemological rather than revelatory: they make the consequences of a hypothesis visible.</p><p>Technology, in this sense, need not be positioned against phenomenological or humanistic inquiry. It can become complementary to it. Computation expands the experimental space in which an idea can be followed, transformed, and challenged.</p><h3>The AI horizon</h3><p>Beginning Digital Romanticism at the present moment gives this older project an unexpected new horizon. Artificial intelligence is rapidly expanding our capacity to generate, transform, compare, classify, and navigate forms. Images that once required days of production can appear in seconds; immense visual and textual archives can be traversed computationally; patterns can be identified across collections that exceed the scale of individual scholarship; and increasingly sophisticated relationships can be established among language, images, data, and geometry.</p><p>Much of the discussion surrounding AI understandably concentrates on novelty and prediction: what will these technologies do next, which practices will they disrupt, and what new forms will they produce? Digital Romanticism is interested in a complementary question. <strong>What might these technologies allow us to revisit?</strong></p><p>The intellectual foundations of modernity are filled with partially resolved projects concerning morphology, organicism, archetypes, transformation, symmetry, typology, natural construction, and generative form. Some of these traditions were displaced by later scientific paradigms, while others survived in architecture, biology, mathematics, art, and computation under different names. The current technological transformation creates an unprecedented opportunity to return to such ideas with instruments their authors could never have possessed.</p><p>This does not mean proving that Goethe anticipated generative AI, that Gothic builders invented parametric design, or that historical philosophy can simply be translated into datasets. The more interesting task is to establish encounters between different modes of thought. A historical theory can be reconstructed as a simulation; a diagram can become the basis of a generator; an architectural corpus can be mapped as a field of related possibilities; and a philosophical proposition about transformation can be tested through computational variation. New technologies may therefore become instruments not only for producing futures, but for <strong>reopening unfinished histories</strong>.</p><h3>Another technological trajectory</h3><p>Technological history is often narrated as a progressive sequence in which each new instrument supersedes the intellectual world that preceded it. Digital Romanticism begins from a less linear proposition. Technology can move backward as well as forward, allowing abandoned concepts, speculative sciences, geometric practices, and alternative theories of form to re-enter contemporary thought under new conditions. The digital becomes interesting not simply because it promises something unprecedented, but because it can transform our access to what has already been thought.</p><p>The subjects of this series will therefore move freely across historical periods and disciplinary boundaries. Gothic geometry may sit beside generative algorithms; biological morphology beside architectural typology; Romantic science beside artificial intelligence. What connects these episodes is not chronology but an interest in how forms emerge, differentiate, mutate, and remain intelligible through change.</p><p>The contemporary wanderer in the image above consequently faces a peculiar landscape. Computation has made aspects of that territory increasingly visible while simultaneously expanding it beyond anything an individual observer could hope to survey. Artificial intelligence only intensifies this condition. The horizon is becoming more navigable and more uncertain at the same time.</p><p><strong>Digital Romanticism is an attempt to enter that landscape by bringing history, theory, and computation into the same field of inquiry.</strong> Its ambition is not to use technology to escape history, but to use it to return to historical ideas with different questions and different instruments. If Romantic science once sought an adventure of reason capable of following nature into processes of transformation, the digital present may offer an opportunity to resume that adventure along another trajectory.</p><h3>Further Reading</h3><p>Johann Wolfgang von Goethe, <em>The Metamorphosis of Plants</em></p><p>Immanuel Kant, <em>Critique of Judgment</em></p><p>Robert J. Richards, <em>The Romantic Conception of Life</em></p><p>Robert Bork, <em>The Geometry of Creation</em></p><p>George Stiny, <em>Shape: Talking about Seeing and Doing</em></p><p>Przemysław Prusinkiewicz and Aristid Lindenmayer, <em>The Algorithmic Beauty of Plants</em></p><p>Common buttercup leaf generated through alternating stages of expansion and contraction, based on the computational method developed in my article “Metamorphic Leaves,” <em>Leonardo</em> 53, no. 5 (2020).</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a2014ab62121" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>