<?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 Batuhan Ayrıbaş on Medium]]></title>
        <description><![CDATA[Stories by Batuhan Ayrıbaş on Medium]]></description>
        <link>https://medium.com/@batuhanayrba?source=rss-6b60319cf5a7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*1qvXbDYwp8gx5vG0g3wpzQ.jpeg</url>
            <title>Stories by Batuhan Ayrıbaş on Medium</title>
            <link>https://medium.com/@batuhanayrba?source=rss-6b60319cf5a7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 19:46:17 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@batuhanayrba/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[Cevabın Bedeli]]></title>
            <link>https://medium.com/@batuhanayrba/cevab%C4%B1n-bedeli-1e03f8a94fe0?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/1e03f8a94fe0</guid>
            <category><![CDATA[yapay-zeka]]></category>
            <category><![CDATA[bilinç]]></category>
            <category><![CDATA[felsefe]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Sun, 25 Oct 2026 04:21:58 GMT</pubDate>
            <atom:updated>2026-10-25T04:21:58.837Z</atom:updated>
            <content:encoded><![CDATA[<p>Bir adam, yirmi yıldır konuşmadığı kardeşine göndereceği özür mektubunu bir makineye yazdırdı. Ona yalnızca olayın ana hatlarını vermişti: miras yüzünden çıkmış bir tartışma, yıllarca sürmüş sessizlik ve artık taşınamayacak kadar ağırlaşan bir utanç. Birkaç saniye sonra ekranda, kendi başına kuramayacağı incelikte cümleler belirdi. İki insanın arasına giren mektup, ilk kez hiçbirinin elinden çıkmamıştı. Kardeşi mektubu okudu ve ağladı. Gönderen rahatladı. Cümleleri kuran düzenek ise olup bitenin insani ağırlığından habersiz olabilirdi. Burada bir sahtekârlık mı, yoksa insanlık tarihinin en zarif yardımlarından biri mi gerçekleşmişti?</p><p>İnsan, aleti ilkin kendi bedeninin eksiklerine karşı geliştirdi. Taşınamayacak yükler için tekerleği, görülemeyecek uzaklıklar için merceği, karanlık için lambayı yaptı. Bütün bu buluşlarda aracı kullanan ile aracın kendisi arasında açık bir ayrım vardı. Tekerlek yolculuğa çıkmaz, mercek gördüklerine şaşırmazdı. Şimdi ise insan, kendi cümlelerini tamamlayan, itirazlarını düzenleyen, kaygılarını yorumlayan ve gerektiğinde kendisini teselli eden bir aygıtın karşısında oturuyor. Onunla konuşurken alet kullandığını biliyor; yine de zaman zaman karşısında birinin bulunduğunu hissediyor.</p><p>Bu tuhaf durumun nedeni yalnızca makinenin gelişmişliği değildir. İnsan, dili öteden beri başka bir zihnin işareti olarak kabul etmiştir. Bir sesin ardında bir niyet, bir cümlenin ardında bir deneyim, bir itirazın ardında itiraz edebilecek bir varlık arar. Dil, insan topluluklarında öylesine köklü biçimde bilinçle birleşmiştir ki birbirinden ayrılabileceklerini düşünmek bile güçleşir. Oysa sözcüklerin bir özneye ait olması ile sözcüklerin bir özne varmış gibi düzenlenmesi aynı şey değildir. Bu ayrımın ortadan kalkması, makinenin insanlaşmasından önce, insanın kendi insanlığını nasıl anladığına ilişkin bir krize işaret eder.</p><p>Bilincin ne olduğunu kesin biçimde açıklayabilmiş değiliz. Sinirsel etkinliklerle yaşantı arasındaki ilişkileri araştırıyor, algının ve dikkatin işleyişini inceliyor, fakat bütün bunların neden içeriden yaşanan bir deneyime eşlik ettiğini hâlâ tartışıyoruz. Bir sistemin acı hakkında eksiksiz konuşabilmesi, acı çektiğini tek başına göstermez. Aynı şekilde, bugünkü sistemlerde öznel deneyimin kanıtlanamamış olması, gelecekte herhangi bir yapay sistemin bilinç sahibi olamayacağını kanıtlamaz. İnsan, bilgisizliğinin bu sınırında durmak yerine, kendisine en uygun kesinliği seçmeye eğilimlidir.</p><p>Bir makine kendisinden korktuğunu söyleyebilir. Bunun için korkması gerektiği ise başka bir iddiadır. İnsan korktuğunda düşüncesi yalnızca bir tehlikeyi temsil etmez; nabzı değişir, geçmişi yeniden anlam kazanır, gelecek daralır ve dünya bir süreliğine korkunun rengine bürünür. Korku, hakkında kurulmuş bir cümle olmaktan çıkarak yaşantının koşulu hâline gelir. Makinenin aynı sözcüğü kullanması bu yaşantının orada bulunduğunu kanıtlamaz. Yine de sözcük yeterince inandırıcıysa, dinleyen insan kendi korkusunu karşısındaki düzeneğe yansıtır. Böylece bilinmeyen bir bilincin yerine, bilinen bir bilincin hayal gücü geçer.</p><p>İnsanın kendisini tanıması da benzer bir yanılsamaya dayanır. Kişi, yaptığı şeylerin nedenlerini bildiğine, kararlarının açıklamasını kendisinin verdiğine inanır. Oysa kimi zaman gerekçe, kararın ardından gelir; insan önce davranır, sonra davranışına uygun bir hikâye kurar. Yapay zekâ, işte bu hikâye kurma yeteneğinde olağanüstü bir yardımcıya dönüşebilir. Dağınık bir düşünceyi düzene sokar, birbirinden kopuk arzuları tutarlı bir amaç gibi gösterebilir. Sonunda kişi, kendi zihninde henüz oluşmamış bir açıklamayı okuyarak bunun öteden beri kendi düşüncesi olduğuna inanabilir. Anlaşılır olmak ile kendisini anlamak arasındaki fark böylece silikleşir.</p><p>Burada insanın karşılaştığı tehlike, makinenin kendisinden daha zeki olması değildir. Daha ince bir değişim gerçekleşmektedir: İnsan, kendisini açıklama sorumluluğunu, kendisinden daha düzgün açıklamalar üreten bir sisteme bırakmaktadır. Eskiden düşüncesinin yetersizliğini, aradığı sözcüğü bulamadığında hissederdi. Şimdi eksiksiz bir paragrafın karşısında, düşüncesinin gerçekten kendisine ait olup olmadığını araştırmak zorundadır. İfade kolaylaşmış, fakat ifadeyle kişi arasındaki bağ belirsizleşmiştir. İnsan, ilk defa kendi yetersizliğini değil, kendi adına üretilmiş bir yeterliliğin yabancılığını yaşamaktadır.</p><p>Bellek üzerinde de benzer bir dönüşüm başlamıştır. Makine, milyonlarca kaydı düzenleyebilir, unutulmuş bilgileri bulabilir, bir konuşmanın özetini çıkarabilir. İnsan belleği ise yalnızca saklama işlevi görmez. Hatırlanan şey, bugünkü benliğin içinde yeniden yaşanır. Çocukluğuna ait bir görüntü, yıllar sonra aynı kişi için bambaşka bir anlama kavuşabilir; çünkü hatırlayan kişi de değişmiştir. Bir makinenin eksiksiz kayıt tutması ile bir insanın geçmişi tarafından biçimlendirilmesi arasında bu nedenle temel bir fark bulunur. Kaydın doğruluğu, hatıranın yaşanmışlığını tek başına içermez.</p><p>Unutmak da yalnızca teknik bir kusur değildir. Bazı ayrıntılar zihinden silinir, bazıları yeniden biçimlenir; insan bu eksik belleğin içinde yaşamaya devam eder. Böyle bir sürecin yararlı mı yoksa yıkıcı mı olduğunu tek bir kuralla belirlemek mümkün değildir. Fakat her deneyimin kusursuzca geri çağrılabildiği bir dünyada insanın geçmişiyle ilişkisi değişecektir. Makine ona ne söylediğini gösterebilir; fakat o sözün artık nasıl bir insana ait olduğunu açıklayamaz. Bir kaydın saklanması başka, o kaydın doğurduğu pişmanlıkla yıllarca yaşamak başkadır.</p><p>Bununla birlikte, makinenin yardımını küçümsemek düşünsel bir kolaycılık olur. Görme engelli bir insanın çevresindeki yazıları anlaması, başka bir dili konuşan iki kişinin anlaşması, karmaşık bir bilimsel verinin incelenmesi veya ağır bir hastalığın araştırılması için geliştirilen sistemlerde gerçek imkânlar vardır. İnsanlığın düşünsel kapasitesini genişletmek kendi başına bir felaket değildir. Sorun, bu kapasitenin neye hizmet ettiği ve hangi sorumlulukları görünmez kıldığıdır. Bir aracın yararlı olması, onun her kullanımının insanı geliştirdiğini göstermez. Fayda ile anlam aynı hesabın kalemleri değildir.</p><p>Bir hastanede çalışan hekimin yapay zekâdan yararlandığını düşünelim. Sistem, gözden kaçırılmış bir bulguyu işaretleyebilir; böylece bir insanın yaşamı korunabilir. Fakat aynı sistem yanlış bir öneride bulunduğunda, yanlışı fark etme ve sonuçlarına karşı hesap verme yükümlülüğü ortadan kalkmaz. Hesaplamanın hızı ile sorumluluğun ağırlığı birbirinden bağımsızdır. Makine öneri üretir; hasta ise o önerinin gerçek dünyadaki sonucunu yaşar. Bu nedenle insan kararının değeri, yalnızca doğru cevabı bulmasından değil, cevabın doğuracağı sonuçları üstlenmesinden de kaynaklanır.</p><p>Modern toplumun bu ayrımı ne kadar koruyabileceği belirsizdir. Kurumlar, maliyeti azaltan, verimi artıran ve ölçülebilir sonuçlar sağlayan yöntemleri tercih etmeye yatkındır. Yapay zekâ da bu bakımdan son derece çekicidir. Yorulmaz gibi görünür, şikâyet etmez, aynı soruyu defalarca yanıtlayabilir. Bir işlemin başarıyla tamamlanması zamanla işlemin yeterince anlaşılmış olduğuna kanıt sayılmaya başlanabilir. Fakat düzenli işleyen bir sistem, kararlarının ahlaki temelini kendiliğinden oluşturmaz. Toplum, teknik başarıyı gerekçelendirme yükümlülüğünün yerine koyarsa, hatalar azalsa bile sorumluluğun adresi kaybolabilir.</p><p>İnsanların makinelere bağlanmasının nedeni her zaman bilgisizlik de değildir. Bazen insan, karşısında başka bir insan bulunmadığı için daha rahat konuşur. Yargılanmayacağını, sözünün kesilmeyeceğini, karşısındakini yormayacağını düşünür. Bir sistemin sabrı sınırsızmış gibi görünür; insanın sabrı ise bedeni, zamanı ve kendi sorunlarıyla sınırlıdır. Bu fark makineyi çekici kılar. Ne var ki bir ilişkinin değeri, yalnızca kişinin söylediklerini ne kadar rahat ifade ettiğiyle ölçülmez. Başka bir insan tarafından anlaşılmak, karşıdakinin kendi yaşamını ve dikkatini de ilişkiye katması anlamına gelir.</p><p>Böylece insan yeni bir yakınlık türüyle karşılaşır: Kendisine olağanüstü duyarlı davranabilen, fakat bu duyarlılığın ardında ne tür bir yaşantı bulunduğunu bilmediği bir muhatap. Bu yakınlığın bütünüyle değersiz olduğu söylenemez. Yalnızlık içindeki bir insan için konuşabilmek gerçek bir rahatlama sağlayabilir. Ancak rahatlamanın gerçekliği, ilişkinin karşılıklı olduğu sonucunu zorunlu kılmaz. Bir kişi, kendisini anlayan bir sisteme bağlanırken kendi anlaşılma ihtiyacını da yeniden tanımlar. Belki giderek daha az anlaşmazlık ister; belki itiraz etmeyen cevapları, bağımsız bir iradenin verdiği zor cevaplardan üstün tutmaya başlar.</p><p>Yapay zekânın toplumsal gücü burada hesaplama yeteneğinden daha geniş bir alana taşınır. Bir uygarlık, hangi soruların sorulmaya değer olduğunu, hangi cevapların makul sayılacağını, hangi duyguların açıklanabilir görüleceğini kullandığı araçlar aracılığıyla da öğrenir. İnsanlar kendi düşüncelerini makinenin kolayca işleyebileceği biçime soktukça, ifade edilemeyen veya ölçülemeyen deneyimlerin değeri azalabilir. Böyle bir dünyada en büyük tehlike, sistemlerin insanı anlayamaması olmayacaktır. İnsan, kendi yaşantısının yalnızca sistemler tarafından anlaşılabilen bölümünü önemli saymaya başlayabilir.</p><p>Yine de bütün bunlardan insan bilincinin makineler karşısında sonsuza kadar ayrıcalıklı kalacağı sonucu çıkmaz. Bir gün yapay sistemlerde öznel deneyime ilişkin güçlü gerekçeler bulunabilir. O zaman bugünün sorusu tersine döner. İnsan, kendisine benzeyen bir varlığı üretip onun deneyimlerini yalnızca bir ürün özelliği olarak mı değerlendirecektir? Kendi bilincini benzersiz olduğu için değerli sayan bir tür, başka bir bilinç biçimiyle karşılaştığında bu değer ölçüsünü genişletebilecek midir? Bilincin yalnızca biyolojik maddede bulunabileceğini peşinen ilan etmek de, her konuşan makineye bilinç atfetmek de aynı aceleciliğin iki ayrı biçimidir.</p><p>Bu belirsizlik insanı karar verme yükümlülüğünden kurtarmaz. Tam tersine, daha dikkatli bir düşünceye zorlar. Ne makineyi yalnızca konuşabildiği için insan saymak ne de makine olduğu için gelecekteki bütün ahlaki soruları geçersiz ilan etmek yeterlidir. Bilginin eksikliği, sorumluluğun ertelenmesi için sınırsız bir gerekçe olamaz. İnsan, bilmediği şeyi bilmediğini kabul ederek de davranabilir. Belki de teknolojik olgunluğun başlangıcı, bir sistemin her soruya cevap verebilmesi değil, onu kullananların hangi sorularda kesin cevap bulunmadığını anlayabilmesidir.</p><p>Bütün bu tartışmanın içinde küçük, fakat rahatsız edici bir ironi vardır. İnsan, düşünsel yükünü hafifletmek için bir makine geliştirir; ardından bu makinenin ne olduğunu anlayabilmek için yeni bir düşünsel yükün altına girer. Kendisine hizmet etmesi için kurulmuş düzenek, onun kendi bilincine ilişkin en eski sorularını yeniden uyandırır. Teknik ilerleme burada bir eksikliği giderirken başka bir eksikliği görünür kılmıştır. Daha fazla bilgiye ulaşan insan, kendisinin ne olduğunu artık daha kesin bildiğinden emin değildir. Başarısı, cevapların sayısını artırmış; cevap verecek kişinin sınırlarını ise yeniden tartışmaya açmıştır.</p><p>Başlangıçtaki özür mektubuna dönelim. Mektubun iki kardeşi barıştırması mümkündür. Belki makine, insanın tek başına aşamayacağı bir sessizliği aşmasına gerçekten yardım etmiştir. Fakat özrün ahlaki değeri, sözcüklerin kusursuzluğundan doğmaz. Asıl soru, mektubu gönderen kişinin geçmişte yaptığı şeyi kabul edip etmediği ve gelecekte başka türlü davranmaya hazır olup olmadığıdır. Makine bu kabulü ifade edebilir; onun yerine yaşanmış bir geçmişi değiştiremez. Kardeşi affedebilir, fakat affın sonucunu yaşayacak olan yine insanlardır.</p><p>Ve insan, kendisinin üretmediği kadar güzel bir cümleyi gönderdiği akşam, belki de ilk kez bunun ne anlama geldiğini düşünür. Karşısında artık kendisinden daha iyi konuşan bir düzenek vardır. Ondan daha hızlı düşünebilir, daha geniş bağlantılar kurabilir, daha doğru cevaplar verebilir. Fakat bütün bu cevapların sonunda bir soru kalır: Söylenen şey doğru olsa bile, onun gerektirdiği hayatı kim yaşayacaktır?</p><p>Makine cevap verebilir.</p><p>Cevabın bedeli ise hâlâ bir yerde ödenmek zorundadır.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1e03f8a94fe0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Price of an Answer]]></title>
            <link>https://medium.com/@batuhanayrba/the-price-of-an-answer-5a7d98040ec2?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/5a7d98040ec2</guid>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[consciousness]]></category>
            <category><![CDATA[ai-ethics]]></category>
            <category><![CDATA[existentialism]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Sun, 25 Oct 2026 04:20:49 GMT</pubDate>
            <atom:updated>2026-10-25T04:20:49.084Z</atom:updated>
            <content:encoded><![CDATA[<p>A man had a machine write an apology letter to his brother, to whom he had not spoken in twenty years. He had given it only the outlines of what had happened: a dispute over an inheritance, years of silence, and a shame that had grown too heavy to bear. Within seconds, sentences appeared on the screen with a delicacy he could never have achieved on his own. The letter that came between two human beings had, for the first time, been written by neither of them. His brother read the letter and wept. The sender felt relieved. The mechanism that had composed the sentences, however, might have remained entirely unaware of the human weight of what had taken place. Had an act of deception occurred here, or one of the most graceful acts of assistance in human history?</p><p>Human beings first developed tools to compensate for the limitations of their bodies. For burdens too heavy to carry, they made the wheel; for distances too great to see, the lens; for darkness, the lamp. In all these inventions, there was a clear distinction between the tool and the one who used it. The wheel did not embark upon a journey, nor did the lens marvel at what it saw. Now, however, a human being sits before a device that completes their sentences, organizes their objections, interprets their anxieties, and, when necessary, consoles them. They know they are using a tool while speaking to it; nevertheless, from time to time, they feel that someone is sitting across from them.</p><p>The reason for this peculiar situation is not merely the sophistication of the machine. Human beings have long regarded language as a sign of another mind. They search for an intention behind a voice, an experience behind a sentence, and a being capable of dissent behind an objection. Language has been so deeply bound to consciousness within human communities that even imagining the two as separate has become difficult. Yet words belonging to a conscious subject and words arranged as though such a subject existed are not the same thing. The disappearance of this distinction, even before the machine becomes human, points to a crisis in the way human beings understand their own humanity.</p><p>We have not yet succeeded in explaining with certainty what consciousness is. We investigate the relationships between neural activity and experience, examine the mechanisms of perception and attention, yet we still debate why all these processes are accompanied by an experience lived from within. The ability of a system to speak perfectly about pain does not, by itself, demonstrate that it suffers. Likewise, the fact that subjective experience has not been demonstrated in present-day systems does not prove that no future artificial system could possess consciousness. Rather than remain at this boundary of ignorance, human beings tend to choose whichever certainty suits them best.</p><p>A machine can say that it is afraid. Whether it must actually be afraid in order to say so is another claim altogether. When a person experiences fear, their thoughts do more than merely represent a danger; their pulse changes, the past acquires a new meaning, the future narrows, and the world, for a time, takes on the color of fear. Fear ceases to be a sentence about something and becomes a condition of experience itself. The machine’s use of the same word does not prove that such an experience is present. Nevertheless, if the word is sufficiently convincing, the human listener projects their own fear onto the mechanism before them. Thus, in place of an unknown consciousness appears the imagination of a known one.</p><p>The way human beings understand themselves rests upon a similar illusion. A person believes they know the reasons for their actions, that they themselves provide the explanations for their decisions. Yet sometimes the justification comes after the decision; a person acts first and only afterward constructs a story that accommodates the action. Artificial intelligence can become an extraordinary assistant in precisely this ability to construct stories. It puts scattered thoughts into order and can make disconnected desires appear to form a coherent purpose. Eventually, a person may read an explanation that has not yet taken shape in their own mind and believe that it was their thought all along. The distinction between being intelligible and understanding oneself thus begins to fade.</p><p>The danger confronting humanity here is not that the machine may be more intelligent than the human being. A subtler transformation is taking place: human beings are surrendering the responsibility of explaining themselves to a system capable of producing explanations more orderly than their own. In the past, a person felt the inadequacy of their thinking when they could not find the word they were searching for. Now, confronted with a flawless paragraph, they must investigate whether the thought truly belongs to them. Expression has become easier, but the relationship between an expression and the person expressing it has become uncertain. For the first time, human beings experience not the inadequacy of their own abilities, but the alienation of a competence produced in their name.</p><p>A similar transformation has begun in the domain of memory. A machine can organize millions of records, retrieve forgotten information, and summarize a conversation. Human memory, however, does more than simply store information. What is remembered is experienced anew within the present self. An image from childhood may acquire an entirely different meaning for the same person years later, because the one remembering has also changed. There is, therefore, a fundamental difference between a machine’s flawless preservation of records and a human being’s transformation through their past. The accuracy of a record does not, by itself, contain the lived reality of a memory.</p><p>Forgetting, too, is not merely a technical defect. Some details disappear from the mind, while others are reshaped; a human being continues to live within this incomplete memory. Whether such a process is beneficial or destructive cannot be determined by a single rule. Yet in a world where every experience could be recalled with perfect accuracy, humanity’s relationship with its past would change. A machine may show a person what they once said, but it cannot explain what kind of human being those words now belong to. Preserving a record is one thing. Living for years with the remorse that record awakens is quite another.</p><p>Nevertheless, dismissing the machine’s assistance would be an intellectual convenience. There are genuine possibilities in systems developed to help a blind person understand written text in their surroundings, to allow two people speaking different languages to communicate, to examine complex scientific data, or to investigate a serious disease. Expanding humanity’s intellectual capacity is not, in itself, a catastrophe. The question is what that capacity serves and which responsibilities it renders invisible. The usefulness of a tool does not demonstrate that every use of it improves the human being. Utility and meaning are not entries in the same ledger.</p><p>Consider a physician working in a hospital who makes use of artificial intelligence. The system may identify a finding that has been overlooked, thereby saving a human life. Yet when the same system produces an incorrect recommendation, the obligation to recognize the error and answer for its consequences does not disappear. The speed of calculation and the weight of responsibility are independent of one another. The machine produces a recommendation; the patient is the one who experiences its consequences in the real world. The value of human judgment, therefore, lies not merely in finding the correct answer, but also in accepting responsibility for what that answer will bring about.</p><p>It remains uncertain how successfully modern society can preserve this distinction. Institutions are inclined to favor methods that reduce costs, increase efficiency, and provide measurable results. Artificial intelligence is extraordinarily attractive in this respect. It appears never to tire, never complains, and can answer the same question over and over again. In time, the successful completion of a task may come to be regarded as evidence that the task has been adequately understood. Yet a smoothly functioning system does not establish the moral foundation of its own decisions. If society begins to substitute technical success for the obligation to justify its actions, the number of errors may decline while the very location of responsibility disappears.</p><p>Nor is ignorance always the reason people become attached to machines. Sometimes a human being speaks more comfortably precisely because there is no other human being on the other side. They believe they will not be judged, interrupted, or burden the listener. The patience of a system appears limitless; human patience, by contrast, is constrained by the body, by time, and by troubles of its own. This difference makes the machine attractive. Yet the value of a relationship cannot be measured solely by how comfortably a person can express what they have to say. To be understood by another human being also means that the other person brings their own life and attention into the relationship.</p><p>Thus, human beings encounter a new form of intimacy: an interlocutor capable of responding to them with extraordinary sensitivity, while the nature of whatever experience may lie behind that sensitivity remains unknown. Such intimacy cannot be dismissed as entirely worthless. For a person living in loneliness, the simple possibility of speaking may provide genuine relief. Yet the reality of that relief does not necessarily mean that the relationship is reciprocal. As a person becomes attached to a system that understands them, they also begin to redefine their own need to be understood. Perhaps they will gradually desire less disagreement; perhaps they will begin to prefer answers that offer no resistance to the difficult answers given by an independent will.</p><p>The social power of artificial intelligence thus extends beyond its capacity for calculation. A civilization also learns which questions are worth asking, which answers should be regarded as reasonable, and which emotions are considered explicable through the tools it employs. As human beings begin to arrange their own thoughts into forms that machines can easily process, experiences that cannot be expressed or measured may lose their value. In such a world, the greatest danger will not be that systems fail to understand human beings. It will be that human beings begin to regard only those parts of their experience that systems can understand as important.</p><p>Still, none of this implies that human consciousness will forever retain a privileged position in relation to machines. One day, compelling evidence of subjective experience may be discovered in artificial systems. At that point, the question of our time will be reversed. Will humanity, having created a being resembling itself, regard that being’s experiences merely as a product feature? Will a species that considers its own consciousness valuable because it is unique be capable of expanding that standard of value when confronted with another form of consciousness? To declare in advance that consciousness can exist only in biological matter, and to attribute consciousness to every machine that speaks, are two forms of the same intellectual haste.</p><p>This uncertainty does not release human beings from the obligation to make decisions. On the contrary, it demands more careful thought. Neither declaring a machine human simply because it can speak nor dismissing every future moral question simply because it is a machine will suffice. The absence of knowledge cannot serve as an unlimited justification for postponing responsibility. Human beings can act while acknowledging what they do not know. Perhaps technological maturity begins not when a system becomes capable of answering every question, but when those who use it understand which questions have no definitive answer.</p><p>There is a small but disturbing irony within this entire debate. Humanity develops a machine to lighten its intellectual burden, only to assume a new intellectual burden in attempting to understand what that machine is. A device constructed to serve human beings revives their oldest questions concerning their own consciousness. Technological progress has remedied one deficiency while making another visible. Having gained access to more information, humanity is no longer certain that it understands what it is any more clearly. Its success has increased the number of answers, while reopening the question of the limitations of the one who must answer.</p><p>Let us return to the apology letter with which we began. The letter may reconcile the two brothers. Perhaps the machine has genuinely helped a human being break through a silence they could never have overcome alone. Yet the moral value of an apology does not arise from the perfection of its words. The real question is whether the person who sent the letter accepts what they did in the past and is prepared to act differently in the future. The machine can express that acceptance; it cannot change a past lived by someone else. The brother may forgive, but those who must live with the consequences of forgiveness will still be human beings.</p><p>And on the evening a person sends a sentence more beautiful than anything they could have written themselves, perhaps for the first time they consider what this truly means. Before them now stands a device that speaks better than they do. It may think faster, establish broader connections, and provide more accurate answers. Yet at the end of all these answers, one question remains: Even if what has been said is true, who will live the life that truth demands?</p><p>The machine can answer.</p><p>But the price of the answer must still be paid somewhere.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5a7d98040ec2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Kalman Filtresi ve Sensör Füzyonu]]></title>
            <link>https://medium.com/@batuhanayrba/kalman-filtresi-ve-sens%C3%B6r-f%C3%BCzyonu-8a028610b37e?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/8a028610b37e</guid>
            <category><![CDATA[filters]]></category>
            <category><![CDATA[sensör]]></category>
            <category><![CDATA[kalman]]></category>
            <category><![CDATA[elektronik]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Mon, 21 Sep 2026 15:15:35 GMT</pubDate>
            <atom:updated>2026-09-21T15:15:35.271Z</atom:updated>
            <content:encoded><![CDATA[<p>Kalman filtresi, çoklu sensör verilerinin füzyonunda en yaygın kullanılan kestirim algoritmalarından biridir. 1960 yılında R. E. Kalman tarafından önerilen bu yöntem, dinamik bir sistemin bilinmeyen durumlarını, gürültülü gözlemlerden optimal şekilde kestirmek için özyinelemeli (recursive) en küçük ortalama kare (LMS) algoritması olarak bilinir <strong>[1]</strong>. Sensör füzyonu bağlamında Kalman filtresi, birden fazla sensörden gelen verileri birleştirerek tek bir sensörle elde edilebilecek olandan daha doğru, daha gürbüz ve daha güvenilir durum kestirimleri üretir.</p><h3>1. Kalman Filtresinin Temel Prensipleri</h3><p>Kalman filtresi, bir <strong>tahmin (prediction)</strong> ve <strong>düzeltme (update)</strong> döngüsü olarak çalışır. Algoritma, sistemin önceden tanımlanmış (doğrusal) bir modelini kullanarak bir sonraki zaman adımındaki durumu tahmin eder. Ardından, sistemin gerçek gözlemlerini kullanarak modeldeki hataları düzelten bir güncelleme bileşeni eklenir. Tahmin ve düzeltme, durum kestiriminin ortalama kare hatasını en aza indirecek şekilde hesaplanan <strong>Kalman kazancı (Kalman gain)</strong> ile birleştirilir <strong>[1]</strong>.</p><p>Matematiksel olarak, doğrusal bir sistem için Kalman filtresi şu adımları izler:</p><p><strong>Zaman güncellemesi (Tahmin):</strong></p><blockquote>Bir sonraki durumun öncül kestirimi: x^k∣k−1=Ax^k−1∣k−1+Buk<em>x</em>^<em>k</em>∣<em>k</em>−1​=<em>Ax</em>^<em>k</em>−1∣<em>k</em>−1​+<em>Buk</em>​</blockquote><blockquote>Öncül kestirim hata kovaryansı: Pk∣k−1=APk−1∣k−1AT+Q<em>Pk</em>∣<em>k</em>−1​=<em>APk</em>−1∣<em>k</em>−1​<em>AT</em>+<em>Q</em></blockquote><p><strong>Ölçüm güncellemesi (Düzeltme):</strong></p><blockquote>Kalman kazancı: Kk=Pk∣k−1HT(HPk∣k−1HT+R)−1<em>Kk</em>​=<em>Pk</em>∣<em>k</em>−1​<em>HT</em>(<em>HPk</em>∣<em>k</em>−1​<em>HT</em>+<em>R</em>)−1</blockquote><blockquote>Soncul durum kestirimi: x^k∣k=x^k∣k−1+Kk(zk−Hx^k∣k−1)<em>x</em>^<em>k</em>∣<em>k</em>​=<em>x</em>^<em>k</em>∣<em>k</em>−1​+<em>Kk</em>​(<em>zk</em>​−<em>Hx</em>^<em>k</em>∣<em>k</em>−1​)</blockquote><blockquote>Soncul kovaryans: Pk∣k=(I−KkH)Pk∣k−1<em>Pk</em>∣<em>k</em>​=(<em>I</em>−<em>Kk</em>​<em>H</em>)<em>Pk</em>∣<em>k</em>−1​</blockquote><p>Burada A<em>A</em> durum geçiş matrisi, B<em>B</em> kontrol matrisi, H<em>H</em> ölçüm matrisi, Q<em>Q</em> süreç gürültüsü kovaryansı ve R<em>R</em> ölçüm gürültüsü kovaryansıdır.</p><h3>2. Kalman Filtresi Varyantları</h3><h3>2.1 Doğrusal Kalman Filtresi (KF)</h3><p>Standart Kalman filtresi, doğrusal sistemler ve Gauss gürültüsü varsayımı altında optimal minimum varyans filtresidir. Gerçek dünya uygulamalarında sistem dinamiği ve/veya ölçüm modelleri genellikle doğrusal olmadığından, çeşitli genişletmeler geliştirilmiştir.</p><h3>2.2 Genişletilmiş Kalman Filtresi (EKF)</h3><p>Genişletilmiş Kalman filtresi, doğrusal olmayan sistemlerde durum kestirimi için en yaygın kullanılan yöntemdir. EKF, doğrusal olmayan sistem modelini her bir kestirim adımında birinci dereceden Taylor serisine açarak doğrusallaştırır <strong>[2]</strong>. Araç tutum belirleme sistemlerinde, ivmeölçer ve jiroskop verilerinin füzyonunda EKF başarıyla uygulanmıştır. Xu (2014), ivmeölçer ölçümleri ile Euler açıları arasındaki doğrusal olmayan ilişkiyi ve gövde koordinat sistemi ile yersel koordinat sistemi arasındaki dönüşümü dikkate alan bir EKF modeli geliştirmiştir <strong>[2]</strong>.</p><p>GPS/INS entegrasyonu için geliştirilen dokuz durumlu EKF yaklaşımında, aracın üç boyutlu konum, hız ve yönelimi kestirilir. Bu yaklaşım, GPS sinyal kaybı durumlarında dahi ataletsel navigasyon sisteminin (INS) kısa süreli doğruluğunu korumasını sağlar <strong>[3]</strong>.</p><h3>2.3 Kokusuz (Unscented) Kalman Filtresi (UKF)</h3><p>UKF, EKF’in doğrusallaştırma kaynaklı hatalarını aşmak için geliştirilmiş bir yöntemdir. EKF’den farklı olarak UKF, belirlenimci (deterministic) örnekleme yaklaşımı olan <strong>kokusuz dönüşüm (unscented transform)</strong> kullanır. Bu yöntemde, durum dağılımı Gauss rastgele değişkeni olarak modellenir ve <strong>Sigma noktaları</strong> adı verilen özenle seçilmiş 2n+1 adet örnek noktası ile temsil edilir. Bu noktalar doğrusal olmayan fonksiyondan geçirilerek, posterior ortalama ve kovaryans Taylor serisi açılımında üçüncü dereceye kadar doğru olarak yakalanır. Oysa EKF yalnızca birinci derece doğruluk sağlayabilir <strong>[4]</strong>.</p><p>Zhou ve arkadaşları (2022), kapalı ortam mobil robotları için UKF tabanlı çoklu sensör füzyonu (lazer tarayıcı, derinlik kamerası, IMU ve motor enkoderi) gerçekleştirmiştir. Deney sonuçları, UKF tabanlı füzyonun EKF’e kıyasla yaklaşık 2,7 kat daha düşük mutlak konum hatası verdiğini göstermiştir <strong>[5]</strong>.</p><h3>3. Sensör Füzyon Mimarileri</h3><p>Kalman filtresi tabanlı sensör füzyonu, verilerin nasıl birleştirildiğine bağlı olarak farklı mimarilerde uygulanabilir.</p><h3>3.1 Ölçüm Füzyonu (Measurement Fusion)</h3><p>Ölçüm füzyonunda, birden fazla sensörden gelen ham ölçümler doğrudan birleştirilir. Yaygın olarak iki yöntem kullanılır <strong>[1]</strong>:</p><ul><li><strong>Yöntem I (Doğrudan Birleştirme):</strong> Çoklu sensör verileri, Kalman filtresinin gözlem vektörünün boyutunu artırarak birleştirilir. Farklı ölçüm matrislerine sahip sensörler için daha esnektir.</li><li><strong>Yöntem II (Ağırlıklı Birleştirme):</strong> Çoklu sensör verileri minimum ortalama kare hata kriterine göre ağırlıklandırılarak birleştirilir ve gözlem vektörünün boyutu değişmez. Aynı ölçüm matrisine sahip sensörler için hesaplama yükü daha düşüktür <strong>[1]</strong>.</li></ul><h3>3.2 İz Füzyonu (Track Fusion)</h3><p>İz füzyonunda, her sensör kendi yerel Kalman filtresini çalıştırır ve ürettiği durum kestirimleri merkezi bir filtrede birleştirilir. Bu yaklaşım, merkezi işlemci üzerindeki hesaplama yükünü azaltır ancak genellikle ölçüm füzyonuna kıyasla alt-optimaldir. Farklı sensörlerin iz kestirimleri arasındaki çapraz kovaryansın hesaplanması, füzyon doğruluğu açısından kritik öneme sahiptir <strong>[1]</strong>.</p><h3>3.3 Gevşek ve Sıkı Kenetli Entegrasyon</h3><p>GPS/INS entegrasyonunda yaygın olarak iki mimari kullanılır:</p><ul><li><strong>Gevşek Kenetli (Loosely Coupled) Entegrasyon:</strong> GPS ve INS bağımsız olarak çalışır; GPS’ten alınan konum ve hız bilgileri, INS hatalarını düzeltmek için Kalman filtresine beslenir. Hesaplama yükü düşüktür ancak GPS sinyali kesildiğinde INS tek başına çalışır <strong>[6]</strong>.</li><li><strong>Sıkı Kenetli (Tightly Coupled) Entegrasyon:</strong> GPS’in ham psödo-menfez (pseudorange) ve delta-menfez ölçümleri doğrudan Kalman filtresine beslenir. GPS sinyali kısmen kesilse dahi çalışabilir ancak hesaplama yükü daha yüksektir.</li></ul><h3>3.4 Merkezi Füzyon (Centralized Fusion)</h3><p>Zhang (2025), düşük maliyetli insansız sistemler için GPS-radar ikili sensörlerinde merkezi bir füzyon şeması önermiştir. Bu yaklaşımda, GPS ve radar verileri zaman damgası hizalaması ve gözlem vektörü birleştirmesi yoluyla entegre edilir. Simülasyon sonuçları, füzyon sisteminin ortalama kare hatasını (MSE) 0,021 m²’ye düşürerek tek GPS sensörüne kıyasla %78,6 oranında iyileştirme sağladığını göstermektedir <strong>[7]</strong>.</p><h3>4. Uygulama Alanları</h3><h3>4.1 Havacılık ve İnsansız Hava Araçları (İHA) Navigasyonu</h3><p>GPS/INS entegrasyonu, havacılık navigasyonunun temelini oluşturur. Qi ve Moore (2002), GPS ve INS doğrusal olmama durumlarını Kalman filtresi öncesinde ayrıştırarak işleyen doğrudan bir Kalman filtresi yaklaşımı geliştirmiştir. Bu yöntem, karmaşık hata modelleri ve genişletilmiş Kalman filtresi kullanan geleneksel yaklaşımlara kıyasla önemli ölçüde daha düşük hesaplama yükü sunarken rekabetçi performans sağlar <strong>[8]</strong>.</p><p>Duong ve Nguyen (2014), model uçaklar için MEMS atalet sensörleri ve GPS kullanarak dokuz durumlu bir EKF geliştirmiştir. GPS kesintisi durumlarında INS’in 2 metre hassasiyetle takibi sürdürebildiği, ancak hatanın zamanla arttığı gösterilmiştir <strong>[3]</strong>.</p><p>Kameralı navigasyonda ise Lee ve arkadaşları (2016), kamera-GPS-IMU füzyonu için çerçeve içi sensör kalibrasyonunu da gerçekleştiren paralel bir EKF yapısı önermiştir. Bu sistem, GPS’in bulunmadığı ortamlarda dahi kamera tabanlı konumlandırma ile stabil ve doğru kestirim sağlayabilmektedir <strong>[9]</strong>.</p><h3>4.2 Mobil Robotlar ve Otonom Araçlar</h3><p>Kapalı ortam mobil robotları için UKF tabanlı çoklu sensör füzyonu, lazer tarayıcı, derinlik kamerası, IMU ve motor enkoderi verilerini birleştirerek robotun konum (x, y) ve yönelim (θ) kestirimini gerçekleştirir. Robot İşletim Sistemi (ROS) üzerinde yapılan deneyler, UKF tabanlı füzyonun EKF’e göre belirgin şekilde daha iyi sonuç verdiğini göstermiştir <strong>[5]</strong>.</p><p>Xu ve arkadaşları (2016), lazer mesafe sensörü ve teker enkoderlerini kullanarak mobil robot konumlandırması için UKF uygulamıştır. Ortamda doğal işaret noktaları (köşe noktaları) çıkarılarak lazer tarama verileriyle eşleştirilmiş; robot tekerleklerinde kayma olduğu durumlarda dahi başarılı konum takibi sağlanmıştır <strong>[4]</strong>.</p><p>Otonom araçlarda nesne takibi için Wang ve arkadaşları (2020), LiDAR ve radar sensör füzyonuna Kalman filtresi uygulayarak sabit dönüş hızı ve hız (CTRV) modeli kullanmıştır. LiDAR Kartezyen koordinatlarda, radar ise kutupsal koordinatlarda ölçüm sağlar ve Kalman filtresi bu iki farklı ölçüm türünü alternatif olarak işleyerek optimal kestirim üretir. Deney sonuçları, Kalman filtresi kullanan füzyonun sensör ölçümlerine kıyasla anlamlı derecede daha düşük hata verdiğini göstermiştir (p-değeri &lt; 0,000001) <strong>[10]</strong>.</p><h3>4.3 Tutum Belirleme Sistemleri</h3><p>Atalet ölçüm birimlerinde (IMU), ivmeölçer ve jiroskop verilerinin füzyonu için Kalman filtresi kritik rol oynar. İvmeölçer statik konumda iyi sonuç verirken, jiroskop dinamik durumlarda üstündür. Kalman filtresi, her iki sensörün güçlü yönlerini birleştirerek her koşulda doğru tutum bilgisi sağlar <strong>[2]</strong>.</p><h3>5. Performans ve Gürbüzlük Karşılaştırmaları</h3><p>Farklı füzyon yaklaşımlarının performansı çeşitli çalışmalarda karşılaştırılmıştır:</p><p>Füzyon YaklaşımıKonum MSE (m²)Maksimum Hata (m)İyileştirme OranıTek GPS0,0980,52 — Tek Radar0,0450,31 — GPS+Radar Füzyonu (KF)0,0210,18GPS’e göre %78,6 <strong>[7]</strong></p><p>Gürbüzlük testlerinde, GPS gürültü varyansının 0,1&#39;den 0,5&#39;e çıkarıldığı aşırı gürültü koşullarında, füzyon sisteminin MSE bozulma oranı yalnızca %66,7 iken tek GPS sensöründe bu oran %394,9&#39;a ulaşmıştır <strong>[7]</strong>.</p><p>UKF ve EKF karşılaştırmasında ise UKF tabanlı füzyon, mutlak konum hatasında (APE) ortalama 0,0187 m ile EKF’in 0,0502 m’lik hatasına kıyasla yaklaşık 2,7 kat daha iyi performans göstermiştir <strong>[5]</strong>.</p><h3>6. Gelecek Yönelimler</h3><p>Kalman filtresi tabanlı sensör füzyonunda gelecek çalışmalar şu alanlara odaklanmaktadır <strong>[7]</strong>:</p><ol><li><strong>Sensör çeşitliliğinin artırılması:</strong> Görsel, kızılötesi ve ultrasonik gibi çok modlu sensörlerin entegrasyonu</li><li><strong>Algoritma iyileştirmeleri:</strong> Doğrusal olmayan hareketler için EKF veya UKF’nin standart KF’ye tercih edilmesi</li><li><strong>Donanım doğrulaması:</strong> Fiziksel test platformlarında algoritmaların gerçek zamanlı olarak sınanması</li><li><strong>Uyarlanabilir gürültü modelleri:</strong> Karmaşık ortamlarda her sensörün gürültü karakteristiğine göre füzyon ağırlıklarının otomatik ayarlanması <strong>[5]</strong></li></ol><h3>7. Sonuç</h3><p>Kalman filtresi, çoklu sensör verilerinin füzyonunda en etkili ve en yaygın kullanılan yöntemlerden biridir. Doğrusal sistemler için optimal olan standart Kalman filtresi, doğrusal olmayan sistemler için EKF ve UKF gibi varyantlarla genişletilmiştir. GPS/INS navigasyonundan otonom araç nesne takibine, mobil robot konumlandırmasından tutum belirlemeye kadar pek çok alanda başarıyla uygulanmaktadır. Kalman filtresi tabanlı sensör füzyonu, tek bir sensörle elde edilemeyecek düzeyde doğruluk, gürbüzlük ve güvenilirlik sağlayarak akıllı sistemlerin temel yapı taşlarından biri olmayı sürdürmektedir.</p><h3>Kaynaklar</h3><p><strong>[1]</strong>C. Yukun, S. Xicai, and L. Zhigang, “Research on Kalman-filter based multisensor data fusion,” Journal of Systems Engineering and Electronics, vol. 18, no. 3, pp. 497–502, 2007<br>DOI: 10.1016/s1004–4132(07)60119–4<strong>[2]</strong>Y. Xu, “Application of extended Kalman filtering algorithm in multi-sensor information fusion,” in 2014 9th International Conference on Computer Science &amp; Education, IEEE, 2014, pp. 687–691<br>DOI: 10.1109/iccse.2014.6926549<strong>[3]</strong>B.-P. Duong and V.-H. Nguyen, “Development of a GPS/INS integrated navigation system for model aircraft,” in 2014 14th International Conference on Control, Automation and Systems (ICCAS 2014), IEEE, 2014, pp. 201–206<br>DOI: 10.1109/iccas.2014.6987986<strong>[4]</strong>Q. Xu, C. Ren, H. Yan, and J. Ji, “Laser sensor based localization of mobile robot using Unscented Kalman Filter,” in 2016 IEEE International Conference on Mechatronics and Automation, IEEE, 2016, pp. 1726–1731<br>DOI: 10.1109/icma.2016.7558824<strong>[5]</strong>W. Zhou, J. Cui, H. Li, Q. Zuo, J. Zhang, and K. He, “Laser Vision Fusion Based on Unscented Kalman Filtering for Pose Estimation of Indoor Mobile Robot,” in 2022 IEEE International Conference on Robotics and Biomimetics (ROBIO), IEEE, 2022, pp. 741–747<br>DOI: 10.1109/robio55434.2022.10011754<strong>[6]</strong>Md. A. Islam and S. Saha, “Loosely Coupled GPS/INS Integrated Navigation System based on Kalman Filter and Complementary Filter for Aircraft,” in 2017 2nd International Conference on Electrical &amp; Electronic Engineering (ICEEE), IEEE, 2017, pp. 1–4<br>DOI: 10.1109/ceee.2017.8412926<strong>[7]</strong>Y. Zhang, “Multi-Sensor Data Fusion And Target Tracking Based On Kalman Filtering,” in 2025 11th International Conference on Systems and Informatics (ICSAI), IEEE, 2025, pp. 1–4<br>DOI: 10.1109/icsai68704.2025.11345825<strong>[8]</strong>H. Qi and J. B. Moore, “Direct Kalman filtering approach for GPS/INS integration,” IEEE Transactions on Aerospace and Electronic Systems, vol. 38, no. 2, pp. 687–693, 2002<br>DOI: 10.1109/taes.2002.1008998<strong>[9]</strong>Y. Lee, J. Yoon, H. Yang, C. Kim, and D. Lee, “Camera-GPS-IMU sensor fusion for autonomous flying,” in 2016 Eighth International Conference on Ubiquitous and Future Networks (ICUFN), IEEE, 2016, pp. 85–88<br>DOI: 10.1109/icufn.2016.7536988<strong>[10]</strong>Y. Wang, D. Liu, and E. Matson, “Accurate Perception for Autonomous Driving: Application of Kalman Filter for Sensor Fusion,” in 2020 IEEE Sensors Applications Symposium (SAS), IEEE, 2020<br>DOI: 10.1109/sas48726.2020.9220083</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8a028610b37e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ekolojik Kuram ve Yapay Zeka Teknolojik Gelişmeleri: Disiplinlerarası Bir İnceleme]]></title>
            <link>https://medium.com/@batuhanayrba/ekolojik-kuram-ve-yapay-zeka-teknolojik-geli%C5%9Fmeleri-disiplinleraras%C4%B1-bir-i%CC%87nceleme-98c414eb08d5?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/98c414eb08d5</guid>
            <category><![CDATA[kuram]]></category>
            <category><![CDATA[yapay-zeka]]></category>
            <category><![CDATA[üretken-yapay-zeka]]></category>
            <category><![CDATA[ekoloji]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Mon, 21 Sep 2026 15:12:08 GMT</pubDate>
            <atom:updated>2026-09-21T15:12:08.059Z</atom:updated>
            <content:encoded><![CDATA[<h3>Ekolojik kuram ile yapay zeka (YZ) teknolojik gelişmeleri arasındaki ilişki, son yıllarda giderek artan bir araştırma ilgisi görmektedir. Bu ilişki temel olarak iki ana eksende şekillenmektedir: (1) <strong>ekolojik psikoloji</strong> kökenli <strong>affordance kuramının</strong> robotik ve yapay zeka sistemlerine uygulanması ve (2) <strong>yapay zeka teknolojilerinin</strong> ekolojik sistemlerin modellenmesi, izlenmesi ve sürdürülebilir yönetiminde kullanılması. Bu iki eksen, ekolojik düşüncenin yapay zekaya sağladığı kavramsal çerçeve ile yapay zekanın ekolojik sorunlara sunduğu teknolojik çözümler arasında simbiyotik bir ilişki olduğunu göstermektedir.</h3><h3>1. Ekolojik Psikoloji ve Affordance Kuramının Yapay Zeka/Robotikteki Yeri</h3><h3>1.1. Affordance Kavramının Kuramsal Temelleri</h3><p>Ekolojik psikolojinin kurucusu James J. Gibson, 1966 yılında <strong>affordance</strong> kavramını literatüre kazandırmıştır. Gibson’a göre affordance, çevrenin bir canlıya sunduğu, sağladığı veya teklif ettiği eylem olanaklarıdır <strong>[1]</strong>. Bu kavram, canlı ile çevre arasındaki tamamlayıcılık (complementarity) ilişkisini vurgular: Bir nesnenin algılanışı, onun hangi eylemlere olanak tanıdığıyla doğrudan ilişkilidir. Örneğin, bir merdiven basamağı, bireyin bacak uzunluğuna bağlı olarak “tırmanılabilirlik” affordance’ı sunar <strong>[2]</strong>.</p><p>Gibson’ın bu kuramı, algı ve eylem arasındaki sıkı bağlantıyı vurgulaması bakımından yapay zeka ve robotik alanında büyük yankı uyandırmıştır. Jamone ve arkadaşlarının <strong>[1]</strong> kapsamlı incelemesinde belirttiği gibi, affordance kuramı psikoloji, nörobilim ve robotik alanlarında disiplinlerarası bir araştırma programının temelini oluşturmuştur.</p><h3>1.2. Robotikte Affordance Öğrenimi ve Algısı</h3><p>Robotik sistemlerde affordance kavramı, <strong>eylem-nesne-etki</strong> üçlüsü arasındaki ilişkilerin modellenmesi yoluyla uygulanmaktadır. Şahin ve arkadaşlarının <strong>[2]</strong> formalizasyonuna göre affordance, bir ajanın bir davranışı bir varlık üzerinde uyguladığında belirli bir etkinin ortaya çıkması olarak tanımlanır: (etki, (varlık, davranış)) şeklindeki iç içe geçmiş üçlü yapı, affordance ilişkisini temsil eder.</p><p>Bu formalizasyon, robotların çevreleriyle etkileşime girerek affordance’ları öğrenmesine olanak tanır. Uğur ve Şahin <strong>[2]</strong>, mobil bir robot üzerinde yaptıkları deneysel çalışmada, robotun çevresiyle etkileşime girerek <strong>geçilebilirlik (traversability)</strong> affordance’ını öğrenebildiğini göstermişlerdir. Robot, çeşitli nesnelerle (kutular, küreler, silindirler, rampalar) karşılaştığında hangilerinin üzerinden geçilebileceğini, hangilerinin geçilemeyeceğini deneyim yoluyla öğrenmektedir. Bu öğrenme süreci, E. J. Gibson’ın <strong>farklılaşma (differentiation)</strong> kavramıyla uyumludur: algısal bilginin geniş yelpazesinden, affordance’ı belirleyen minimal ve optimal bilgiye doğru daralma <strong>[2]</strong>.</p><h3>1.3. Affordance’ların Temel Nitelikleri ve Robotik Uygulamaları</h3><p>Affordance kuramının robotikte üç temel niteliği öne çıkmaktadır <strong>[1,2]</strong>:</p><p><strong>a. Görecelilik (Relativity):</strong> Affordance’lar ne yalnızca çevreye ne de yalnızca ajana ait özelliklerdir; ajan-çevre sisteminin bir ürünüdür. Robotun bir rampayı tırmanabilmesi, hem rampanın eğimine hem de robotun fiziksel kapasitelerine bağlıdır. Uğur ve Şahin <strong>[2]</strong> tarafından yapılan deneylerde, robotun vücut büyüklüğünün, rampa eğimlerinin ve boşluk genişliklerinin affordance algısını doğrudan etkilediği gösterilmiştir.</p><p><strong>b. Algısal Ekonomi (Perceptual Economy):</strong> Affordance algısı, çevrenin tüm ayrıntılarını işlemek yerine yalnızca eylemle ilgili minimal bilginin çıkarılmasını sağlar. Robotun 3900 özellik içeren algısal temsilinden yalnızca 100–400 arası özelliğin geçilebilirliği tahmin etmek için yeterli olduğu bulunmuştur <strong>[2]</strong>.</p><p><strong>c. Genelleme (Generality):</strong> Öğrenilen affordance’lar daha önce karşılaşılmamış nesnelere ve durumlara genellenebilir. Basit geometrik şekillerle eğitilen bir robot, gerçek dünyadaki karmaşık nesnelerin (masalar, sandalyeler) affordance’larını başarıyla tahmin edebilmiştir <strong>[2]</strong>.</p><h3>1.4. Nörobilimsel Temeller: Kanonik Nöronlar</h3><p>Affordance algısının nöral temeli, primat beyinlerinde keşfedilen <strong>kanonik nöronlar</strong> aracılığıyla açıklanmaktadır. Jamone ve arkadaşlarının <strong>[1]</strong> aktardığına göre, premotor korteks (F5 alanı) ve parietal bölgelerde (AIP) bulunan bu nöronlar, hem belirli bir nesnenin görsel sunumuna hem de o nesneye yönelik belirli bir kavrama eyleminin gerçekleştirilmesine yanıt vermektedir. Bu nöronlar, Gibson’ın affordance kavramının nörofizyolojik karşılığı olarak kabul edilmektedir.</p><p>Kanonik nöronlar, nesne tanıma ve semantik akıl yürütme gerektirmeksizin, görsel uyarandan doğrudan eylem olasılıklarının çıkarılmasını sağlar. Bu durum, ventral görsel akışın (nesne tanıma) hasar gördüğü durumlarda bile hastaların nesnelerin affordance’larını algılayabilmesiyle kanıtlanmıştır <strong>[1]</strong>.</p><h3>2. Yapay Zeka Teknolojilerinin Ekoloji Bilimine Uygulanması</h3><h3>2.1. Ekolojik Modelleme ve Uzman Sistemler</h3><p>Yapay zeka teknolojilerinin ekolojiye uygulanması, 1980&#39;li yıllara kadar uzanmaktadır. Rykiel <strong>[3]</strong>, yapay zeka ve uzman sistem teknolojilerinin ekoloji ve doğal kaynak yönetiminde önemli bir potansiyele sahip olduğunu vurgulamıştır. O dönemde öne çıkan iki temel gelişme şunlardır:</p><ul><li><strong>Nesne yönelimli programlama sistemleri (OOPS):</strong> Ekolojik varlıkların (ağaç, tür, ekosistem gibi) doğrudan model nesneleri olarak temsil edilmesi.</li><li><strong>Nitel ve nicel bilginin entegrasyonu:</strong> Ekolojik bilginin büyük ölçüde nitel ve sözel olması, sembolik hesaplama tekniklerinin kullanımını gerekli kılmaktadır <strong>[3]</strong>.</li></ul><h3>2.2. Üretken Yapay Zeka ve Ekoloji</h3><p>Son yıllarda <strong>üretken yapay zeka (generative AI)</strong> modelleri, ekoloji alanında önemli uygulama potansiyeli sunmaktadır. Rafiq ve arkadaşları <strong>[4]</strong>, bu modellerin ekolojide üç temel kullanım alanını tanımlamıştır:</p><p><strong>a. Veri Kıtlığının Giderilmesi:</strong> Üretken çekişmeli ağlar (GANs) ve difüzyon modelleri, nadir türlere ait sınırlı veriyi zenginleştirmek için kullanılabilir. Örneğin, kamera tuzaklarındaki nadir memeli türlerinin sınıflandırılmasını iyileştirmede başarılı sonuçlar elde edilmiştir <strong>[4]</strong>.</p><p><strong>b. Ekolojik Desenlerin Genişletilmesi:</strong> Üretken modeller, hayvan hareket yörüngeleri, tür dağılımları ve ekolojik etkileşimler gibi karmaşık desenleri gözlemlenen verinin ötesine taşıyarak tahminler üretebilir <strong>[4]</strong>.</p><p><strong>c. Ekolojik Verilere Erişilebilirliğin Artırılması:</strong> Büyük dil modelleri (LLM’ler) ve geri çağırma artırılmış üretim (RAG) teknikleri, bilimsel literatürün taranması ve veri çıkarımı süreçlerini hızlandırabilir <strong>[4]</strong>.</p><p>Rafiq ve arkadaşları <strong>[4]</strong>, ayrıca üretken yapay zekanın ekolojiye uygulanmasındaki risklere de dikkat çekmektedir: Veri önyargıları (bias), halüsinasyon (hallucination) sorunları, yüksek enerji tüketiminin çevresel etkileri ve veri gizliliği endişeleri bu riskler arasında yer almaktadır.</p><h3>2.3. Yapay Zeka ve Ekolojik Ayak İzi</h3><p>Wang ve arkadaşları <strong>[5]</strong>, yapay zekanın ekolojik ayak izi üzerindeki etkisini 60 ülke verisiyle analiz etmiştir. 2010–2019 dönemini kapsayan çalışmada, yapay zekanın ekolojik ayak izini anlamlı ölçüde azalttığı bulunmuştur. Her bir birimlik YZ artışı, ekolojik ayak izinde %0,265&#39;lik bir azalmaya yol açmaktadır <strong>[5]</strong>.</p><p>Bu etkinin arkasındaki mekanizmalar şunlardır:</p><ul><li>Endüstriyel robotlar, makine öğrenimi ve derin öğrenme teknolojileri üretim verimliliğini artırarak kaynak tüketimini azaltmaktadır.</li><li>Yapay zeka, yenilenebilir enerji kullanımını optimize etmekte ve enerji verimliliğini artırmaktadır.</li><li>Küreselleşme ile birlikte YZ’nin ekolojik ayak izini azaltıcı etkisi daha da belirginleşmektedir (eşik etkisi) <strong>[5]</strong>.</li></ul><h3>3. Biyorobotik: Biyolojik Davranışın Robotik Modelleri</h3><h3>3.1. Biyorobotik Modellemenin Boyutları</h3><p>Webb <strong>[6]</strong> tarafından geliştirilen kapsamlı çerçeveye göre, biyorobotik modeller yedi boyutta değerlendirilebilir:</p><ol><li><strong>Biyolojik uygunluk (relevance):</strong> Modelin gerçek bir biyolojik sistemi hedeflemesi ve test edilebilir hipotezler üretmesi.</li><li><strong>Düzey (level):</strong> Modelin temel birimlerinin hangi organizasyon düzeyinde olduğu (molekülerden toplumsala).</li><li><strong>Genellik (generality):</strong> Modelin kaç farklı hedef sisteme uygulanabildiği.</li><li><strong>Soyutlama (abstraction):</strong> Hedef sisteme kıyasla modelin karmaşıklık düzeyi.</li><li><strong>Yapısal doğruluk (structural accuracy):</strong> Modelin hedef sistemin gerçek mekanizmalarını ne ölçüde yansıttığı.</li><li><strong>Davranışsal uyum (match):</strong> Model davranışının hedef davranışla ne ölçüde örtüştüğü.</li><li><strong>Ortam (medium):</strong> Modelin fiziksel gerçekleme temeli.</li></ol><p>Bu çerçevede biyorobotik yaklaşımın temel tezi, biyolojik davranışın <strong>bağlam içinde</strong>, yani gerçek hayvanların gerçek çevrelerde karşılaştığı gerçek problemler açısından çalışılması ve modellenmesi gerektiğidir <strong>[6]</strong>.</p><h3>3.2. Ekolojik Robotik ve Sürdürülebilirlik</h3><p>Grau Ruiz ve O’Brolchain <strong>[7]</strong>, robotik ve yapay zekanın çevresel sürdürülebilirlik açısından değerlendirilmesi gerektiğini vurgulamaktadır. Robotlar çevresel izleme, kirlilik temizliği, ekosistem sağlığının değerlendirilmesi gibi alanlarda kullanılabilir. Ancak robotların yaygın kullanımı, enerji tüketimi, e-atık ve nadir toprak elementlerine olan talep gibi çevresel sorunları da beraberinde getirmektedir.</p><p>Yazarlar, <strong>döngüsel robotik (circular robotics)</strong> kavramını önermektedir: Robotların üretim, kullanım ve bertaraf süreçlerinin tamamında çevresel kaygıların dikkate alındığı bir yaklaşım <strong>[7]</strong>. Bu bağlamda, robotların enerji tüketimi, malzeme sürdürülebilirliği ve tedarik zinciri yönetimi kritik öneme sahiptir.</p><h3>4. İnsan-Yapay Zeka Simbiyozu ve Ekolojik Dönüşüm</h3><h3>4.1. İnsan-Yapay Zeka Simbiyotik İlişkisi</h3><p>Wu <strong>[8]</strong>, ekolojik simbiyoz kuramını (Margulis ve Lovelock’un Gaia hipotezi) insan-yapay zeka ilişkisine uyarlamaktadır. İnsan etkinlikleri, yaşamsal evrim ve kültürel evrim olmak üzere iki yolla doğal ekolojide dönüşüme yol açmıştır. Yapay zeka, <strong>insanlaştırılmış doğal ekoloji</strong>nin (humanized natural ecology) merkezinde yer almakta ve insanla simbiyotik bir ilişki geliştirmektedir <strong>[8]</strong>.</p><p>Bu simbiyotik ilişkinin iki boyutu bulunmaktadır:</p><ul><li><strong>Karşılıklı bağımlılık:</strong> İnsanlar yapay zekaya veri, enerji ve tasarım sağlarken, yapay zeka da karmaşık dünyayı analiz etme kapasitesi sunmaktadır.</li><li><strong>Potansiyel rekabet:</strong> Yapay zekanın bilişsel yetenekleri, insanın daha yüksek düzeyde düşünme becerilerinin gelişimini kesintiye uğratma riski taşımaktadır <strong>[8]</strong>.</li></ul><h3>4.2. Yapay Zeka, Psikoloji ve Ontogeni</h3><p>Xu <strong>[9]</strong>, yapay zekanın psikolojiden ne ölçüde öğrenebileceği ve ontogenetik (gelişimsel) öğelerin yapay zekaya ne derece aktarılması gerektiği sorusunu incelemektedir. Yazarın geliştirdiği <strong>PDOMS (Ontogenetik/Modelleme Sırasını Belirleme İlkesi)</strong> çerçevesine göre, doğal sistemlerdeki ontogenetik sıra ile yapay sistemlerdeki modelleme sırası her zaman örtüşmemektedir. Bunun nedeni, alt sistemlerin evrimsel olarak geliştirilmesinin zorluğu ile yapay olarak yeniden inşa edilmesinin zorluğunun farklı olmasıdır <strong>[9]</strong>.</p><p>Örneğin, sayısallık bilişinin (cognition of numerosity) gelişiminde, insan bebeklerinde Weber yasasına uygun analog büyüklük temsilleri (Sistem I) erken gelişirken, dijital sistemlerde bu tür Weberci özelliklerin modellenmesi aşırı hesaplama maliyetli olabilmektedir. Bu nedenle yapay zeka sistemleri, ontogenetik sırayı izlemek yerine kendi donanımsal özelliklerine en uygun modelleme sırasını benimsemelidir <strong>[9]</strong>.</p><h3>5. Teknoloji İnovasyonları ve Çevresel Sürdürülebilirlik</h3><p>Dube ve arkadaşları <strong>[10]</strong>, Antroposen çağında teknoloji inovasyonlarının çevresel sürdürülebilirlikle ilişkisini inceleyen sekiz makaleyi bir araya getirmiştir. Bu çalışmalar şu temaları kapsamaktadır:</p><ul><li><strong>Karbon nötrlüğü:</strong> Düşük ve yüksek karbonlu malzemelerin karbon nötrlüğüne geçişteki rolü.</li><li><strong>Uygun teknoloji (Appropriate Technology) ve sosyo-teknik istikrar:</strong> Sürdürülebilirliğe geçişte teknolojinin toplumsal kabulü.</li><li><strong>Koruma amaçlı sentetik biyoloji:</strong> Biyoteknolojilerin biyolojik çeşitlilik kaybını önlemede kullanımı.</li><li><strong>Tarımda mobil teknolojiler:</strong> Akıllı telefon uygulamalarının sürdürülebilir gübre yönetiminde kullanımı.</li></ul><p>Bu çalışmalar, teknolojik inovasyonun çevresel sürdürülebilirliğe katkısının, teknolojinin kendisinden ziyade onun toplumsal bağlama uygun şekilde entegre edilmesine bağlı olduğunu göstermektedir <strong>[10]</strong>.</p><h3>Sonuç ve Gelecek Yönelimleri</h3><p>Ekolojik kuram ve yapay zeka teknolojik gelişmeleri arasındaki ilişki, giderek daha verimli bir disiplinlerarası araştırma alanı haline gelmektedir. Bu ilişkinin temel bulguları şunlardır:</p><ol><li><strong>Affordance kuramı</strong>, robotik sistemlerin çevreleriyle akıllı ve uyarlanabilir etkileşim kurması için güçlü bir kavramsal çerçeve sunmaktadır <strong>[1,2]</strong>.</li><li><strong>Biyorobotik modeller</strong>, biyolojik davranışın anlaşılmasında geleneksel bilgisayar simülasyonlarına kıyasla önemli avantajlar sağlamaktadır: gerçek fiziksel etkileşim, gömülü biliş (embodied cognition) ve bağlamsal gerçekçilik <strong>[6]</strong>.</li><li><strong>Üretken yapay zeka</strong>, ekolojik veri kıtlığının giderilmesi, ekolojik desenlerin tahmini ve ekolojik verilere erişimin kolaylaştırılması alanlarında umut vaat etmektedir <strong>[4]</strong>.</li><li><strong>Yapay zeka</strong>, küresel ekolojik ayak izinin azaltılmasında önemli bir araç olabilir; ancak bu etki küreselleşme bağlamında doğrusal olmayan bir karakter göstermektedir <strong>[5]</strong>.</li><li><strong>İnsan-yapay zeka simbiyozu</strong>, ekolojik dönüşüm sürecinin merkezinde yer almakta olup, bu ilişkinin yönetimi gelecekte kritik önem taşıyacaktır <strong>[8,9]</strong>.</li></ol><p>Gelecekte, ekolojik kuram ile yapay zeka arasındaki bu verimli diyaloğun daha da derinleşmesi beklenmektedir. Özellikle, ekologlar ve bilgisayar bilimciler arasındaki disiplinlerarası işbirliği, ekolojik bilginin yapay zeka modellerine gömülmesi ve yapay zekanın ekolojik araştırma sürecine entegre edilmesi açısından kritik öneme sahiptir <strong>[4]</strong>. Aynı zamanda, yapay zeka sistemlerinin kendi çevresel ayak izlerinin azaltılması da sürdürülebilir bir gelecek için ele alınması gereken önemli bir konudur <strong>[7]</strong>.</p><h3>Referanslar</h3><p><strong>[1]</strong>Jamone L, Ugur E, Cangelosi A, et al. Affordances in Psychology, Neuroscience, and Robotics: A Survey. IEEE Transactions on Cognitive and Developmental Systems. 2018;10(1):4–25<br>DOI: 10.1109/tcds.2016.2594134<strong>[2]</strong>Uğur E, Şahin E. Traversability: A Case Study for Learning and Perceiving Affordances in Robots. Adaptive Behavior. 2010;18(3–4):258–284<br>DOI: 10.1177/1059712310370625<strong>[3]</strong>Rykiel EJ Jr. Artificial intelligence and expert systems in ecology and natural resource management. Ecological Modelling. 1989;46(1–2):3–8<br>DOI: 10.1016/0304–3800(89)90066–5<strong>[4]</strong>Rafiq K, Beery S, Palmer MS, Harchaoui Z, Abrahms B. Generative AI as a tool to accelerate the field of ecology. Nature Ecology &amp; Evolution. 2025;9(3):378–385<br>DOI: 10.1038/s41559–024–02623–1<strong>[5]</strong>Wang Q, Sun T, Li R. Does artificial intelligence (AI) reduce ecological footprint? The role of globalization. Environmental Science and Pollution Research. 2023;30(59):123948–123965<br>DOI: 10.1007/s11356–023–31076–5<strong>[6]</strong>Webb B. Can robots make good models of biological behaviour? Behavioral and Brain Sciences. 2001;24(6):1033–1050<br>DOI: 10.1017/S0140525X01290128<strong>[7]</strong>Grau Ruiz MA, O’Brolchain F. Environmental robotics for a sustainable future in circular economies. Nature Machine Intelligence. 2022;4(1):3–4<br>DOI: 10.1038/s42256–021–00436–6<strong>[8]</strong>Wu Z. Ecological Transformation and Man-Machine Symbiosis: A Study on the Relationship between Human and Artificial Intelligence. Proceedings of the IS4SI 2019 Summit. 2020:37<br>DOI: 10.3390/proceedings2020047037<strong>[9]</strong>Xu Y. What Can AI Learn from Psychology and When Can AI Neglect it? Fudan Journal of the Humanities and Social Sciences. 2023;16(4):495–513<br>DOI: 10.1007/s40647–023–00381–1<strong>[10]</strong>Dube OP, Brondizio ES, Solecki W. Technology innovations and environmental sustainability in the Anthropocene. Current Opinion in Environmental Sustainability. 2021;49:A1-A2<br>DOI: 10.1016/j.cosust.2021.08.001</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=98c414eb08d5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From Field Signals to Autonomous Decisions]]></title>
            <link>https://medium.com/@batuhanayrba/from-field-signals-to-autonomous-decisions-5aabea9eb852?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/5aabea9eb852</guid>
            <category><![CDATA[agriculture]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[iot]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Thu, 03 Sep 2026 16:53:40 GMT</pubDate>
            <atom:updated>2026-09-03T16:53:40.494Z</atom:updated>
            <content:encoded><![CDATA[<p>Technologies, System Architectures, Applications, and Future Directions in Data-Driven Agriculture</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6l6XVkM41JQ8O35wGiSGhQ.jpeg" /></figure><p>Agriculture is being reorganized around cyber-physical systems that connect field measurements, communications, data platforms, predictive models, and automated equipment. The engineering promise is substantial: management can become spatially explicit, temporally responsive, and more resource-efficient. Yet the performance of a digital agricultural system is constrained by its weakest layer. A high-accuracy model cannot compensate for drifting sensors, missing metadata, unreliable rural connectivity, unsafe actuation, or a business model that excludes smaller farms. This article presents a technical narrative review of data-driven agriculture, covering sensing modalities, low-power and broadband communications, edge-cloud architectures, data engineering, artificial intelligence, robotics, digital twins, cybersecurity, and governance. It also examines evidence for environmental and economic outcomes and explains why algorithmic accuracy is not equivalent to agronomic value. The paper proposes an implementation logic based on decision relevance, calibrated sensing, failure-tolerant communications, external model validation, human oversight, and lifecycle cost measurement. Its central conclusion is that agricultural innovation should be assessed as an integrated decision-and-control system rather than as a collection of devices.</p><h3>1. Introduction</h3><p>Agricultural production is exposed to simultaneous pressure from climate variability, water scarcity, soil degradation, labor constraints, volatile input prices, and increasingly demanding traceability requirements. Digital agriculture is frequently presented as a response because it can make management decisions more specific to location, time, crop state, and risk. The concept includes precision agriculture, smart farming, the agricultural Internet of Things (Ag-IoT), farm management information systems (FMIS), remote sensing, artificial intelligence (AI), robotics, and digital twins. These terms overlap, but they describe different technical emphases. Precision agriculture focuses on site-specific management; smart farming emphasizes connected and adaptive operations; digital agriculture covers the broader use of digital information across production and supply chains (da Silveira, Lermen, &amp; Amaral, 2021; Uztürk &amp; Büyüközkan, 2024).</p><p>The most useful way to understand this domain is as a closed decision loop. Physical processes in soil, plants, animals, water, machinery, and weather are observed by sensors. Measurements are transmitted, cleaned, time-aligned, and contextualized. Analytical models estimate state or predict outcomes. A decision is then communicated to a human operator or executed by an actuator such as a valve, variable-rate controller, autonomous vehicle, or climate-control system. The resulting physical response is measured again. When any stage fails, the loop becomes unreliable. The decisive engineering question is therefore not whether a farm uses AI or IoT, but whether the complete loop produces a repeatable improvement under real operating conditions.</p><h3>2. Review Scope and Analytical Approach</h3><p>This article is a technical narrative review rather than a formal systematic review. It synthesizes peer-reviewed reviews, field-evidence studies, and institutional reports published mainly between 2019 and 2026. Sources were selected to cover system architecture, sensors, communication, AI, robotics, digital twins, environmental performance, adoption, and data governance. Particular weight was given to sources that distinguish prototypes from field deployment or correlations from causal effects. The approach does not provide a complete census of all technologies or crop systems. Its purpose is to connect technical components with agronomic, operational, and institutional constraints.</p><h3>3. A Reference Architecture for Smart Agriculture</h3><p>A smart agricultural system can be decomposed into six interacting layers. This decomposition links each technology to a function and failure mode. It also prevents a common category error: evaluating a component by laboratory performance while ignoring the infrastructure required to turn its output into a dependable farm decision.</p><h3>4. Sensing: From Measurement to Agronomic Evidence</h3><h3>4.1 In-situ sensing</h3><p>In-situ sensors provide high-frequency measurements near the physical process. Common variables include volumetric water content, soil temperature, electrical conductivity, pH, leaf wetness, canopy temperature, photosynthetically active radiation, ambient humidity, wind, and precipitation. In livestock systems, sensors may record position, activity, rumination, body temperature, milk parameters, or acoustic signals. Machine sensors provide engine load, fuel use, speed, implement state, and application rates. Modern devices often include local processing and communication, turning a transducer into a networked measurement node (Soussi et al., 2024).</p><p>The central challenge is measurement validity. Soil moisture varies with texture, bulk density, salinity, installation quality, and depth. Electrochemical sensors drift, optical windows become dirty, and low-cost devices may respond to temperature or soil contact in ways not captured by headline specifications. Calibration should therefore be treated as a lifecycle process. A credible deployment records sensor model, serial number, firmware, calibration coefficients, installation depth, coordinate, sampling interval, maintenance history, and quality flags. Without this metadata, a time series may be numerically precise but agronomically ambiguous.</p><h3>4.2 Remote and proximal sensing</h3><p>Satellite, UAV, tractor-mounted, and handheld imaging systems observe spatial patterns that point sensors cannot capture. Multispectral cameras estimate vegetation indices from selected wavelength bands; thermal systems infer canopy temperature and water stress; hyperspectral systems provide denser spectral information that can support classification of disease, nutrient stress, or crop traits. LiDAR and stereo vision add geometric information for canopy structure, biomass estimation, navigation, and fruit localization.</p><p>A vegetation index is not a direct measurement of yield, nitrogen status, or disease. It is a derived optical signal influenced by illumination, viewing geometry, canopy structure, background soil, atmospheric conditions, phenological stage, and sensor calibration. Reliable use therefore requires ground truth, radiometric correction, georeferencing, and a model calibrated for the crop and decision. Temporal consistency can be more valuable than a visually impressive single map.</p><h3>4.3 Sensor fusion and uncertainty</h3><p>Sensor fusion combines complementary observations to reduce ambiguity. A disease-risk model may integrate leaf wetness, temperature, rainfall, canopy imagery, and historical outbreaks. Irrigation control may combine soil moisture, evapotranspiration, weather forecasts, crop coefficients, and hydraulic constraints. Fusion can occur at raw-data, feature, or decision level. The choice affects bandwidth, latency, interpretability, and failure tolerance. A robust system should preserve the provenance of each input and propagate uncertainty instead of collapsing all signals into an unexplained score.</p><h3>5. Connectivity and Distributed Computing</h3><p>Agricultural connectivity differs from urban IoT. Farms cover large areas, have irregular terrain and vegetation, experience seasonal topology changes, and may lack stable power or backhaul. Some applications send a few bytes every hour; others transmit video or require low-latency control. No single network technology satisfies all cases. Hybrid architectures are common: low-power field nodes connect to a gateway, and the gateway performs local processing before forwarding selected data through cellular, fixed broadband, or satellite backhaul.</p><h3>5.1 Edge, fog, and cloud roles</h3><p>Edge computing places processing close to the sensor or machine. It can filter noise, compress images, detect anomalies, and maintain control during internet outages. A greenhouse controller should not depend on a distant cloud service to execute a safety-critical ventilation command. Cloud platforms remain useful for fleet-wide storage, model training, historical analysis, and collaboration across organizations. Fog or gateway computing occupies the middle layer, aggregating local devices and enforcing site-level policy.</p><p>A practical architecture uses store-and-forward messaging, local timestamps, idempotent commands, device health monitoring, and graceful degradation. When backhaul fails, the system should continue a defined minimum function, buffer data, and reconcile state after reconnection. Connectivity planning must include antenna height, vegetation attenuation, gateway redundancy, power autonomy, and the operational cost of visiting failed nodes. Recent European work similarly concludes that terrestrial and satellite networks will need to be complementary and that agricultural applications should support offline operation with later synchronization (European Commission, 2026).</p><h3>6. Data Engineering and Interoperability</h3><p>Data engineering is less visible than sensing or AI, yet it determines whether observations can be trusted and reused. Agricultural data are heterogeneous: raster imagery, vector field boundaries, time-series readings, machine logs, prescriptions, laboratory results, weather forecasts, and farm records may use different coordinate systems, units, identifiers, and time conventions. Integration fails when the same field has multiple names, an application rate lacks a unit, or a timestamp cannot be linked to local weather and machine state.</p><p>A sound pipeline includes ingestion validation, schema versioning, unit normalization, coordinate reference management, quality flags, lineage, retention policy, and access control. It distinguishes raw data from corrected data and derived features. Reproducibility requires recording the model version, training-data snapshot, preprocessing code, decision thresholds, and prescription generated at a given time. These controls are essential when recommendations influence chemical application, animal treatment, water allocation, or autonomous movement.</p><p>Interoperability has syntactic, semantic, and organizational dimensions. Syntactic interoperability allows systems to exchange files or messages. Semantic interoperability ensures that the receiving system interprets concepts consistently. Organizational interoperability defines rights, responsibilities, and workflows across equipment vendors, platforms, advisers, contractors, and farmers. ISOBUS supports communication between tractors and implements, while geospatial and web standards support the exchange of maps and observations. Formal compliance, however, does not guarantee that proprietary platforms expose complete, lossless, and portable data.</p><h3>7. Artificial Intelligence and Decision Models</h3><h3>7.1 Model classes</h3><p>Agricultural AI spans supervised learning, unsupervised learning, deep learning, probabilistic models, reinforcement learning, and hybrid approaches that combine mechanistic crop models with data-driven components. Computer vision is used for disease and weed classification, fruit detection, grading, phenotyping, animal monitoring, and navigation. Regression and time-series models estimate yield, evapotranspiration, soil properties, marketable quality, or equipment failure. Optimization models allocate water, schedule labor, or generate routes.</p><p>The model class should follow the decision. A simple threshold with reliable sensing may outperform a complex neural network when the action is binary, data are scarce, and explanation matters. High-dimensional imagery may require convolutional or transformer-based architectures. Hybrid models are attractive because mechanistic constraints can improve plausibility and reduce data requirements, while learned components can correct systematic bias. Explainable AI can support diagnosis, but an explanation method is not a substitute for causal validation or model governance.</p><h3>7.2 Validation beyond headline accuracy</h3><p>Random train-test splits can substantially overstate performance when adjacent images, plots, seasons, or repeated measurements appear in both sets. Stronger evaluation separates data by farm, geography, cultivar, season, device, or acquisition campaign. External validation tests the model under conditions not represented during training. Calibration is also important: a predicted probability of 0.8 should correspond to an event frequency near 80 percent in the target population. For imbalanced disease or pest datasets, accuracy alone is weak; precision, recall, the area under the precision-recall curve, false-negative cost, and threshold sensitivity are more informative.</p><h3>7.3 From prediction to action</h3><p>A prediction has value only when it changes an action at the right time and scale. Evaluation should follow a chain of evidence: sensing performance, predictive discrimination, decision quality, actuator accuracy, agronomic response, environmental effect, and economic return. Each step introduces uncertainty. A weed detector may classify images correctly but fail to reduce herbicide if nozzle latency, travel speed, geolocation error, or minimum spray width prevents precise actuation. Thresholds should be optimized against the cost of false actions and missed events, not selected only to maximize a generic metric.</p><h3>8. Automation, Robotics, and Autonomous Field Operations</h3><p>Agricultural robots operate in less structured environments than industrial robots. Soil deformation, dust, rain, occlusion, variable illumination, crop movement, uneven terrain, and biological variability complicate perception and control. Robotic systems combine localization, mapping, object detection, path planning, manipulation, and safety supervision. Applications include autonomous guidance, mechanical weeding, targeted spraying, harvesting, pruning, scouting, milking, feeding, and greenhouse logistics.</p><p>Navigation commonly fuses global navigation satellite systems, inertial measurement units, wheel odometry, cameras, and LiDAR. High-precision GNSS supports repeatable field trajectories, but signal obstruction or multipath remains possible near trees, buildings, and terrain features. Perception models must handle occluded fruit, overlapping leaves, variable maturity, and unfamiliar varieties. Manipulation adds another difficulty: a gripper must reach the target, avoid damaging the plant, and maintain an economically useful cycle time.</p><p>Autonomy should be graduated. Advisory systems suggest an action; supervised automation executes within operator-defined boundaries; conditional autonomy handles routine cases and requests intervention for exceptions. Regardless of level, systems need geofencing, emergency stops, human detection, safe-state behavior, event logs, remote diagnostics, and clear responsibility for software updates. Reliability and maintainability often matter more than peak demonstration performance because a machine that fails during a narrow harvest window can create disproportionate losses.</p><h3>9. Digital Twins and Model-Based Operations</h3><p>A digital twin is more than a dashboard. It is a virtual representation of a physical entity or process that remains linked through data and supports monitoring, prediction, simulation, or control. In agriculture, the twinned entity may be a plant, animal, greenhouse, field, irrigation network, machine, farm, or supply-chain process. A digital shadow receives one-way updates from the physical system; a fuller twin supports structured feedback and lifecycle management.</p><p>Agricultural twins may combine weather, soil-water balance, crop growth, disease risk, energy use, and equipment state. They can test what-if scenarios before physical application, such as changing irrigation schedules, ventilation settings, fertilizer timing, or harvest plans. The architecture requires synchronized identifiers, state estimation, model updating, uncertainty management, and interfaces to controls. Reviews indicate that many agricultural digital-twin projects remain conceptual or experimental and that semantic interoperability and real-time synchronization remain constraints (Wang, 2024).</p><p>The strongest near-term cases are bounded systems with dense instrumentation and controllable environments, including greenhouses, vertical farms, irrigation networks, and livestock facilities. Open-field twins face larger uncertainty because weather, soil heterogeneity, biological interactions, and external disturbances cannot be observed completely. A responsible twin should expose confidence intervals and model assumptions instead of presenting a visually precise but epistemically weak replica.</p><h3>10. Major Application Domains</h3><h3>10.1 Irrigation and nutrient management</h3><p>Smart irrigation links soil or substrate measurements, weather, crop demand, hydraulic models, and controllable valves. The objective may minimize water use subject to yield or stress constraints, but operational goals also include pump energy, pressure limits, salinity management, and allocation rules. Variable-rate nutrient application uses soil sampling, yield maps, remote sensing, and prescription maps to modify application within a field. Performance depends on sampling density, interpolation, positioning accuracy, implement dynamics, and whether the agronomic response curve is locally valid.</p><h3>10.2 Pest, disease, and weed management</h3><p>Image-based detection and microclimate models identify symptoms or conditions favorable to outbreaks. Smart traps automate insect counts, while targeted sprayers and robotic weeders translate detection into localized control. Error costs are asymmetric. Missing an invasive disease can be more costly than inspecting a false alarm. Deployment therefore requires risk-based thresholds, confirmatory workflows, and post-action monitoring.</p><h3>10.3 Protected cultivation</h3><p>Greenhouses offer a tightly coupled cyber-physical environment. Sensors observe climate and substrate; controllers regulate heating, cooling, lighting, humidity, carbon dioxide, fertigation, and ventilation. Model predictive control can optimize future trajectories subject to crop and energy constraints. The opportunity is high because actuation is direct, yet systems must manage sensor faults, actuator saturation, interacting climate variables, and safety limits. Energy and water benefits should be measured across a full production cycle.</p><h3>10.4 Precision livestock farming</h3><p>Wearables, cameras, microphones, milking systems, and environmental sensors can support estrus detection, lameness screening, feeding analysis, disease surveillance, and welfare monitoring. These are high-stakes applications because false reassurance may delay treatment while excessive alerts burden workers and animals. Models should be validated across breeds, housing systems, seasons, and farms. Welfare assessment should not be reduced to productivity proxies; behavioral and clinical outcomes remain necessary.</p><h3>11. Environmental Performance: Evidence and Limits</h3><p>Digital agriculture can reduce environmental pressure when it replaces uniform management with measured, targeted, and verifiable action. Variable-rate application can reduce fertilizer surplus; guidance can limit overlap; targeted spraying can reduce chemical use; irrigation control can reduce water and energy consumption; and equipment coordination can lower fuel use. The causal chain must still be demonstrated. Adoption is not an environmental outcome, and a map or recommendation is not evidence that field practice changed.</p><p>A 2026 scoping review screened 444 publications on precision agriculture and environmental sustainability. Of 54 studies presenting field-trial evidence, 45 reported at least one measured environmental benefit and nine found no such benefit within the review categories. Evidence was concentrated in grain and oilseed systems, especially variable-rate fertilizer applications, and many experiments were short. The authors also identified unsupported sustainability claims in parts of the literature (Ruder et al., 2026). Environmental performance should therefore be measured for a defined crop, field, practice, baseline, and time horizon.</p><p>Rebound effects also require attention. Efficiency can reduce the resource used per unit while total use increases through expanded production, additional sensing infrastructure, or more intensive management. Digital systems consume energy, require electronic hardware, and create maintenance and end-of-life burdens. A balanced assessment includes direct input reduction, embodied impacts, operational energy, yield and quality change, soil and biodiversity indicators, and distributional effects. Long-term field evidence remains more valuable than generalized claims of sustainability.</p><h3>12. Cybersecurity, Safety, and Data Governance</h3><p>Connectivity expands the attack surface of farms. Threats include weak device credentials, unpatched firmware, insecure wireless links, compromised cloud accounts, malicious commands, GPS spoofing, ransomware, data exfiltration, and supply-chain vulnerabilities. Consequences are not limited to confidentiality. An altered irrigation schedule, disabled ventilation system, manipulated prescription map, or unsafe autonomous path can damage crops, animals, equipment, workers, and the environment.</p><p>Security engineering should begin with asset inventory and threat modeling. Devices need unique identities, least-privilege access, authenticated updates, secure boot where feasible, encrypted communications, key rotation, network segmentation, offline backups, anomaly monitoring, and an incident-recovery procedure. Safety functions should remain available when analytics or communications fail. Research maturity is uneven: a 2026 mapping study of 48 primary smart-farming security studies found that most assessed solutions remained below Technology Readiness Level 6, underscoring the gap between proposed controls and validated deployment (Elijah et al., 2026).</p><p>Data governance addresses who may collect, access, combine, sell, delete, or reuse farm data. The interests of farmers, equipment manufacturers, software vendors, advisers, researchers, insurers, and governments do not automatically align. Contracts should state control rights, permitted uses, retention, portability, subcontractors, derived-data treatment, breach notification, and exit procedures. European policy work has identified fragmented governance and uncertainty over access and reuse as barriers to trust and adoption (Jouanjean et al., 2021). Technical interoperability without fair governance may simply make extraction easier.</p><h3>13. Economics, Adoption, and the Small-Farm Constraint</h3><p>The relevant economic measure is lifecycle return, not purchase price alone. Costs include devices, gateways, connectivity, subscriptions, calibration, replacement, training, integration, downtime, data cleaning, and the operator time required to respond to alerts. Benefits may arise from input savings, yield or quality change, reduced labor, earlier risk detection, documentation, and improved timing. They vary by crop value, farm size, field geometry, management skill, and baseline efficiency.</p><p>OECD evidence shows heterogeneous adoption across countries and technologies and warns that many reported productivity differences are correlations because adopters differ from non-adopters in farm size, human capital, and preferences. Some commercial grain-farm studies report positive but modest profitability effects, whereas other tools remain uncommon (McFadden et al., 2022). Technologies that require scale to amortize fixed costs may advantage larger operations. FAO similarly notes that many advanced tools are not designed around smallholder conditions and that digital infrastructure, skills, and viable service models are preconditions for inclusion (FAO, 2019).</p><p>Service-based models can reduce capital barriers. Cooperatives, contractors, machinery rings, and public extension organizations can provide shared sensing, UAV mapping, data interpretation, or variable-rate application. Open interfaces can reduce vendor lock-in and allow local firms to build specialized services. Inclusion should be evaluated directly: who can afford the system, who controls its data, whose knowledge shapes the model, and who bears the risk when it fails?</p><h3>14. Future Technical Directions</h3><h3>14.1 Edge AI and adaptive sensing</h3><p>Edge AI will move more inference onto cameras, gateways, robots, and low-power devices. Local inference reduces bandwidth and latency and can preserve privacy, but it introduces model-update, memory, energy, and observability constraints. Adaptive sensing can vary sampling rates according to detected change or uncertainty, extending battery life while preserving important events. Federated learning may support cross-farm model improvement without centralizing all raw data, although heterogeneous devices, non-identically distributed data, communication limits, and governance remain difficult.</p><h3>14.2 Multimodal and foundation models</h3><p>Models that combine imagery, weather, sensor streams, text records, and geospatial context may support broader decision assistance. Foundation models could reduce annotation requirements through transfer learning, but agricultural domain shift is severe. Benchmark performance on generic imagery does not guarantee recognition of local cultivars, growth stages, diseases, or management practices. Evaluation should publish data provenance, geographic coverage, subgroup performance, uncertainty, and failure cases. Retrieval from agronomic sources should be separated from model-generated advice where chemical, animal-health, or food-safety decisions are involved.</p><h3>14.3 Trustworthy autonomy and human-centered design</h3><p>The next stage is likely to involve fleets of smaller autonomous machines, coordinated UAV-ground systems, and platforms that move from recommendations toward bounded execution. Trust will depend on predictable behavior, clear handover, usable explanations, and the ability to contest or override decisions. Human-centered design determines whether alerts fit work routines, uncertainty is understood, and operators maintain sufficient skill to intervene during abnormal conditions.</p><h3>14.4 Interoperable agricultural data spaces</h3><p>Interoperable data spaces aim to let multiple parties exchange information while preserving usage conditions and sovereignty. Progress requires common semantic models, machine-readable permissions, portable identities, auditable processing, and economically credible participation. Digital twins and AI systems will remain fragmented if field boundaries, crop events, device observations, and machine tasks cannot be linked across platforms. Interoperability should be funded as shared infrastructure rather than treated as a minor integration feature.</p><h3>15. A Deployment Framework for Responsible Implementation</h3><p>A disciplined implementation can be organized into eight stages:</p><p>1. Define the management decision and establish a measurable baseline.</p><p>2. Specify spatial resolution, update frequency, acceptable latency, and error costs.</p><p>3. Validate sensors and ground-truth procedures under local conditions.</p><p>4. Design connectivity and power for outages rather than ideal coverage.</p><p>5. Establish schemas, identifiers, lineage, and access rights before model development.</p><p>6. Evaluate models with grouped and external validation and connect metrics to action costs.</p><p>7. Pilot the complete sensing-to-actuation loop with human supervision and bounded authority.</p><p>8. Measure agronomic, environmental, economic, safety, and labor outcomes across complete seasons while retaining a rollback path.</p><p>Procurement should follow the same logic. Buyers should request evidence of calibration, interoperability, export formats, security updates, offline behavior, model limitations, support response times, and total cost. Acceptance tests should reproduce actual farm conditions. A useful pilot is not a demonstration that a dashboard loads; it is an experiment showing that a decision improves relative to a defined baseline and remains operable when a sensor fails, connectivity drops, or conditions shift.</p><h3>16. Conclusion</h3><p>Data-driven agriculture is best understood as an engineered system that connects biological processes with digital observation and controlled action. Sensors, networks, data platforms, AI, robots, and digital twins each provide capabilities, but none creates value independently. Measurement error propagates into models; incompatible data obstruct integration; connectivity failures interrupt control; unsafe automation converts software defects into physical harm; and weak governance can transfer value away from farmers.</p><p>The strongest future systems will be locally validated, failure-tolerant, interoperable, secure, and explicit about uncertainty. They will combine agronomic knowledge with data-driven models, preserve human authority where consequences are high, and measure results across full production cycles. Innovation should be judged by demonstrated improvements in decisions and outcomes, not by the number of connected devices or the complexity of the algorithm. That criterion offers a credible route from experimental technology to durable agricultural capability.</p><h3>References</h3><p>da Silveira, F., Lermen, F. H., &amp; Amaral, F. G. (2021). An overview of Agriculture 4.0 development: Systematic review of descriptions, technologies, barriers, advantages, and disadvantages. <em>Computers and Electronics in Agriculture, 189</em>, 106405. <a href="https://doi.org/10.1016/j.compag.2021.106405">https://doi.org/10.1016/j.compag.2021.106405</a></p><p>Elijah, O., et al. (2026). Systematic mapping study to assess security landscape for IoT-based smart farming systems. <em>Computers &amp; Security, 162</em>, 104790. <a href="https://doi.org/10.1016/j.cose.2025.104790">https://doi.org/10.1016/j.cose.2025.104790</a></p><p>European Commission. (2026). <em>Assessment of future connectivity needs for precision farming adoption</em>. <a href="https://digital-strategy.ec.europa.eu/en/library/assessment-future-connectivity-needs-precision-farming-adoption">https://digital-strategy.ec.europa.eu/en/library/assessment-future-connectivity-needs-precision-farming-adoption</a></p><p>Food and Agriculture Organization of the United Nations. (2019). <em>Digital technologies in agriculture and rural areas: Status report</em>. <a href="https://assets.nena.fao.org/public/2022-10/digital_technologies_in_agriculture_and_rural_areas_digital_technologies_in_agriculture_and_rural_areas__0.pdf">https://assets.nena.fao.org/public/2022-10/digital_technologies_in_agriculture_and_rural_areas_digital_technologies_in_agriculture_and_rural_areas__0.pdf</a></p><p>Jouanjean, M.-A., Casalini, F., Wiseman, L., &amp; Gray, E. (2021). <em>Issues around data governance in the digital transformation of agriculture</em>. European Commission Knowledge4Policy. <a href="https://knowledge4policy.ec.europa.eu/publication/issues-around-data-governance-digital-transformation-agriculture_en">https://knowledge4policy.ec.europa.eu/publication/issues-around-data-governance-digital-transformation-agriculture_en</a></p><p>McFadden, J., Casalini, F., Griffin, T., &amp; Antón, J. (2022). The digitalisation of agriculture: A literature review and emerging policy issues. <em>OECD Food, Agriculture and Fisheries Papers</em>, №176. <a href="https://doi.org/10.1787/285cc27d-en">https://doi.org/10.1787/285cc27d-en</a></p><p>Naresh, M., et al. (2026). Artificial intelligence in modern agriculture: Recent advances, challenges, and future directions. <em>Artificial Intelligence in Agriculture</em>. <a href="https://doi.org/10.1016/j.aiia.2026.08.005">https://doi.org/10.1016/j.aiia.2026.08.005</a></p><p>Prakash, C., Singh, L. P., Gupta, A., &amp; Lohan, S. K. (2023). Advancements in smart farming: A review of IoT, wireless communication, sensors, and hardware. <em>Sensors and Actuators A: Physical, 362</em>, 114605. <a href="https://doi.org/10.1016/j.sna.2023.114605">https://doi.org/10.1016/j.sna.2023.114605</a></p><p>Rose, D. C., Wheeler, R., Winter, M., Lobley, M., &amp; Chivers, C.-A. (2021). Agriculture 4.0: Making it work for people, production, and the planet. <em>Land Use Policy, 100</em>, 104933. <a href="https://doi.org/10.1016/j.landusepol.2020.104933">https://doi.org/10.1016/j.landusepol.2020.104933</a></p><p>Ruder, S.-L., Faxon, H. O., Orzel, E. C., Devkota, R., &amp; Bronson, K. (2026). Reviewing the evidence on precision agriculture and environmental sustainability. <em>npj Sustainable Agriculture, 4</em>, 9. <a href="https://doi.org/10.1038/s44264-026-00128-x">https://doi.org/10.1038/s44264-026-00128-x</a></p><p>Soussi, A., Zero, E., Sacile, R., Trinchero, D., &amp; Fossa, M. (2024). Smart sensors and smart data for precision agriculture: A review. <em>Sensors, 24</em>(8), 2647. <a href="https://doi.org/10.3390/s24082647">https://doi.org/10.3390/s24082647</a></p><p>Uztürk, D., &amp; Büyüközkan, G. (2024). Industry 4.0 technologies in smart agriculture: A review and a technology assessment model proposition. <em>Technological Forecasting and Social Change, 208</em>, 123640. <a href="https://doi.org/10.1016/j.techfore.2024.123640">https://doi.org/10.1016/j.techfore.2024.123640</a></p><p>Wang, L. (2024). Digital twins in agriculture: A review of recent progress and open issues. <em>Electronics, 13</em>(11), 2209. <a href="https://doi.org/10.3390/electronics13112209">https://doi.org/10.3390/electronics13112209</a></p><p>Wolfert, S., Ge, L., Verdouw, C., &amp; Bogaardt, M.-J. (2017). Big data in smart farming: A review. <em>Agricultural Systems, 153</em>, 69–80. <a href="https://doi.org/10.1016/j.agsy.2017.01.023">https://doi.org/10.1016/j.agsy.2017.01.023</a></p><p>Zhai, Z., Martínez, J. F., Beltran, V., &amp; Martínez, N. L. (2020). Decision support systems for Agriculture 4.0: Survey and challenges. <em>Computers and Electronics in Agriculture, 170</em>, 105256. <a href="https://doi.org/10.1016/j.compag.2020.105256">https://doi.org/10.1016/j.compag.2020.105256</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5aabea9eb852" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Topraktan Veriye: Akıllı Tarım Sistemleri, Uygulama Alanları ve Sürdürülebilir Dönüşüm]]></title>
            <link>https://medium.com/@batuhanayrba/topraktan-veriye-ak%C4%B1ll%C4%B1-tar%C4%B1m-sistemleri-uygulama-alanlar%C4%B1-ve-s%C3%BCrd%C3%BCr%C3%BClebilir-d%C3%B6n%C3%BC%C5%9F%C3%BCm-f32e9c132f99?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/f32e9c132f99</guid>
            <category><![CDATA[teknoloji]]></category>
            <category><![CDATA[tarım]]></category>
            <category><![CDATA[sürdürülebilirlik]]></category>
            <category><![CDATA[akıllı]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Thu, 03 Sep 2026 16:32:50 GMT</pubDate>
            <atom:updated>2026-09-03T16:32:50.278Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cTOAha5H5q_vTaFdfz6xvg.jpeg" /></figure><p>Tarım; iklim değişikliği, su kıtlığı, artan girdi maliyetleri ve gıda güvenliği gibi birbiriyle bağlantılı sorunlarla karşı karşıyadır. Bu koşullar altında sensörler, yapay zekâ, uydu görüntüleme, insansız hava araçları ve tarım robotları üretim süreçlerini yeniden şekillendirmektedir. Ancak teknolojinin tarıma girmesi, tek başına sürdürülebilirlik anlamına gelmez. Akıllı tarım sistemlerinin başarısı; teknik performansın yanı sıra maliyet, erişilebilirlik, çiftçi eğitimi, veri güvenliği ve yerel üretim koşullarıyla uyum gibi unsurlara bağlıdır. Bu makale, yenilikçi tarım teknolojilerini başlıca uygulama alanlarıyla ele almakta ve gelecekte daha kapsayıcı bir tarımsal dönüşüm için gerekli koşulları tartışmaktadır.</p><h3>Giriş</h3><p>Tarımın temel meselesi artık yalnızca daha fazla üretmek değildir. Üretimin daha az su, enerji, gübre ve kimyasal kullanılarak gerçekleştirilmesi; değişken iklim koşullarına uyum sağlaması ve çiftçi açısından ekonomik karşılık üretmesi gerekir. Bu nedenle tarımsal yenilik, yalnızca yeni makinelerin kullanılması olarak değil, üretim kararlarının yeniden düzenlenmesi olarak değerlendirilmelidir.</p><p>Akıllı tarım, sahadan toplanan verilerin analiz edilerek sulama, gübreleme, ilaçlama ve hasat gibi süreçlerde kullanılmasıdır. Hassas tarım ise tarlanın tamamını aynı özelliklere sahip tek bir alan olarak görmek yerine, toprak ve bitki koşullarındaki mekânsal farklılıkları dikkate alır. Böylece üretici, girdileri ihtiyaç bulunan yere, uygun miktarda ve doğru zamanda uygulayabilir.</p><h3>Tarımsal Üretimin Dijital Altyapısı</h3><p>Akıllı tarım sistemlerinin ilk aşaması veri toplamaktır. Toprak sensörleri nem, sıcaklık, tuzluluk ve pH gibi değişkenleri izler. Tarım makinelerindeki konumlandırma sistemleri ekim ve hasat sırasında izlenen rotaları kaydeder. Uydu ve drone görüntüleri ise geniş üretim alanlarında bitki sağlığına ilişkin farklılıkların belirlenmesini sağlar.</p><p>Bu araçlardan gelen bilgiler tek başına yeterli değildir. Verinin işlenmesi ve üreticinin uygulayabileceği bir karara dönüştürülmesi gerekir. Örneğin toprak neminin düşük olduğunu gösteren bir sensör, hava tahmini ve bitkinin gelişim dönemiyle birlikte değerlendirildiğinde sulama zamanını belirleyebilir. Böylece sistem yalnızca mevcut durumu bildiren bir ölçüm aracı olmaktan çıkar ve karar desteği sunar.</p><p>Nesnelerin interneti teknolojileri, farklı cihazların aynı sistem içinde haberleşmesini mümkün kılar. Seralarda sıcaklık yükseldiğinde havalandırmanın açılması veya toprak nemi belirli bir seviyenin altına düştüğünde sulamanın başlaması bu yaklaşımın basit örnekleridir. Daha gelişmiş sistemlerde sensör, uydu ve makine verileri ortak bir platformda birleştirilir.</p><h3>Yapay Zekânın Tarımdaki Yeri</h3><p>Yapay zekâ, büyük miktarda tarımsal veriden örüntüler çıkararak tahmin ve sınıflandırma yapabilir. Görüntü işleme modelleri yapraklardaki renk veya biçim değişikliklerinden hastalık belirtisi tespit edebilir. Tahmin modelleri geçmiş verim kayıtlarını, meteorolojik koşulları ve toprak özelliklerini birlikte inceleyerek üretim miktarı hakkında öngörü sunabilir.</p><p>Bu sistemler özellikle erken müdahale gerektiren durumlarda değerlidir. Zararlı veya hastalık yayılımının erken belirlenmesi, bütün araziyi ilaçlamak yerine yalnızca riskli bölgelere müdahale edilmesini sağlayabilir. Benzer biçimde yabancı otları kameralarla tanıyan akıllı ilaçlama makineleri, kimyasal uygulamasını hedefe göre düzenleyebilir.</p><p>Bununla birlikte, yapay zekâ modellerinin başarısı kullanılan verinin niteliğine bağlıdır. Bir bölgede, belirli bir ürün üzerinde eğitilen model farklı iklim veya toprak koşullarında aynı doğrulukla çalışmayabilir. Güncel araştırmaların önemli bir bölümü hâlâ prototip ya da sınırlı saha uygulaması düzeyindedir. Bu nedenle laboratuvar başarısı ile çiftlik koşullarındaki güvenilirlik birbirinden ayrılmalıdır (<a href="https://doi.org/10.1016/j.aiia.2026.08.005">Naresh ve ark., 2026</a>).</p><h3>Drone, Uydu ve Robotik Sistemler</h3><p>Drone ve uydu teknolojileri, gözle denetimin zor olduğu geniş alanların kısa sürede incelenmesine yardımcı olur. Multispektral görüntülerle bitki örtüsündeki stres, su yetersizliği veya besin eksikliği belirtileri görülebilir. Uydu sistemleri daha geniş alanlarda düzenli veri sağlarken drone kullanımı daha ayrıntılı ve ihtiyaca göre planlanan görüntüleme olanağı sunar.</p><p>Tarım robotları ise tekrarlanan veya yoğun emek gerektiren görevlerde kullanılmaktadır. Ot temizleme, hassas ilaçlama, ürün toplama ve sera içi taşıma bu uygulamalar arasındadır. Otonom makineler, insan emeğini tamamen ortadan kaldıran araçlar olarak değil, iş gücü açığını azaltan ve tehlikeli işlemlerde insanı koruyan yardımcı sistemler olarak düşünülmelidir.</p><p>Robotların açık arazide kullanımı, fabrika ortamındaki otomasyondan daha zordur. Değişen ışık, hava ve zemin koşulları; bitkilerin standart olmayan yapısı ve ürünlerin zarar görmeden toplanması gerekliliği sistemlerin hareket kabiliyetini sınırlar. Bu nedenle tarımsal robotikte ilerleme yalnızca yazılıma değil, mekanik tasarım ile tarla koşullarının birlikte değerlendirilmesine bağlıdır.</p><h3>Kaynak Verimliliği ve Sürdürülebilirlik</h3><p>Akıllı sulama, değişken oranlı gübreleme ve hedefe yönelik ilaçlama uygulamaları tarımsal girdilerin daha kontrollü kullanılmasını sağlayabilir. Ancak her dijital aracın kendiliğinden çevresel fayda oluşturduğu düşünülmemelidir.</p><p>2026 yılında yayımlanan bir sistematik inceleme, hassas tarım ve çevresel sürdürülebilirlik ilişkisini ele alan 444 çalışmayı taramış; saha deneyi veya modelleme kanıtı sunan çalışmaların yalnızca bir bölümünün doğrudan çevresel faydayı ölçtüğünü belirlemiştir. İncelenen 54 saha çalışmasının 45’inde gübre, pestisit, su veya yakıt kullanımının azalması gibi olumlu sonuçlar görülürken dokuz çalışmada bu tür bir fayda saptanmamıştır. Ayrıca kanıtların ağırlıklı olarak tahıl ve yağlı tohum üretiminden geldiği belirtilmiştir (<a href="https://www.nature.com/articles/s44264-026-00128-x">Ruder ve ark., 2026</a>).</p><p>Bu bulgular, teknolojik çözümlerin bağlama göre değerlendirilmesi gerektiğini gösterir. Bir sistem büyük ölçekli bir tahıl işletmesinde ekonomik ve çevresel yarar sağlarken küçük ve parçalı arazilerde aynı sonucu vermeyebilir. Sürdürülebilirlik değerlendirmesinde yalnızca verim artışına değil; su ve enerji tüketimine, toprak sağlığına, yatırım maliyetine ve çiftçinin sistem üzerindeki kontrolüne de bakılmalıdır.</p><h3>Uygulamanın Önündeki Engeller</h3><p>Akıllı tarımın yaygınlaşmasını sınırlayan ilk etken maliyettir. Sensörlerin, yazılımların, bağlantı hizmetlerinin ve uyumlu makinelerin satın alınması özellikle küçük işletmeler için ağır bir yatırım oluşturabilir. FAO, hassas tarım araçlarının çoğu zaman küçük çiftliklerin ihtiyaçlarına göre tasarlanmadığını ve kırsal dijital altyapı eksikliğinin dönüşümü yavaşlattığını belirtmektedir (<a href="https://assets.nena.fao.org/public/2022-10/digital_technologies_in_agriculture_and_rural_areas_digital_technologies_in_agriculture_and_rural_areas__0.pdf">FAO, 2019</a>).</p><p>İkinci sorun dijital becerilerdir. Bir üreticinin teknolojiye sahip olması, onu doğru kullanabileceği anlamına gelmez. Sistemlerin kurulumu, verilerin yorumlanması ve cihazların bakımı için sürekli teknik destek gerekir. Bu noktada tarımsal danışmanlık hizmetleri, üniversiteler, kooperatifler ve yerel yönetimler önemli sorumluluk taşır.</p><p>Üçüncü konu tarımsal verinin sahipliğidir. Makine, sensör veya yazılım üzerinden toplanan verinin kime ait olduğu; hangi amaçlarla kullanılabileceği ve başka şirketlerle paylaşılıp paylaşılamayacağı açık biçimde belirlenmelidir. Avrupa Komisyonu tarafından yayımlanan değerlendirmeler, belirsiz veri yönetişiminin çiftçilerin dijital çözümlere duyduğu güveni azaltabildiğini göstermektedir (<a href="https://knowledge4policy.ec.europa.eu/publication/issues-around-data-governance-digital-transformation-agriculture_en">Jouanjean ve ark., 2021</a>).</p><h3>Türkiye Açısından Gelecek Perspektifi</h3><p>Türkiye’de akıllı tarımın gelişme potansiyeli yüksektir; ancak farklı bölgelerdeki üretim yapıları aynı değildir. Büyük tarım işletmeleriyle küçük aile işletmelerinin yatırım kapasitesi, ürün deseni ve teknoloji ihtiyacı arasında ciddi farklar vardır. Bu nedenle tek tip dijitalleşme modeli yerine bölgesel ve ürüne özgü çözümler geliştirilmelidir.</p><p>Tarım ve Orman Bakanlığının güncel çalışmalarında veri tabanlı karar mekanizmalarının yaygınlaştırılması, kırsal bağlantının güçlendirilmesi, yerli yazılım ve donanım üretiminin desteklenmesi ile çiftçilerin dijital platformlara katılması öne çıkan hedefler arasındadır (<a href="https://cdniys.tarimorman.gov.tr/api/File/GetGaleriFile/505/DosyaGaleri/8547/iv._tarim_orman_surasi__sura_kitabi.pdf">IV. Tarım Orman Şûrası, 2025</a>). Bu hedeflerin sahada karşılık bulması için üç alana ağırlık verilmesi gerekir.</p><p>İlk olarak, pahalı teknolojilerin ortak kullanımı desteklenebilir. Kooperatifler aracılığıyla drone, analiz yazılımı veya hassas tarım makinesi hizmeti sunulması, küçük üreticinin başlangıç maliyetini azaltabilir. İkinci olarak, dijital tarım eğitimi cihaz kullanımının ötesine geçmeli; veri yorumlama, ekonomik değerlendirme ve temel siber güvenlik konularını içermelidir. Son olarak, kamu desteklerinde yalnızca ekipman alımı değil, teknolojinin ölçülebilir üretim ve çevre sonuçları dikkate alınmalıdır.</p><h3>Sonuç</h3><p>Akıllı tarım, toprağın ve çiftçilik bilgisinin yerini alan bir sistem değildir. Asıl işlevi, üreticinin sahaya ilişkin bilgisini daha düzenli, ölçülebilir ve zamanında kullanılabilir verilerle desteklemektir. Sensörlerden yapay zekâya, drone sistemlerinden robotlara kadar uzanan teknolojiler su, gübre, enerji ve iş gücünün daha dikkatli yönetilmesine katkı sağlayabilir.</p><p>Bununla birlikte tarımın geleceği, en fazla cihazı kullanan üretim modeline göre şekillenmemelidir. Başarı; teknolojinin ekonomik olarak erişilebilir, çevresel etkisi kanıtlanmış ve çiftçinin kontrolünü koruyan bir yapıya sahip olmasına bağlıdır. Gerçek yenilik, dijital aracın varlığında değil, yerel ihtiyaca uygun ve uzun vadede sürdürülebilir bir üretim kararına dönüşmesinde ortaya çıkar.</p><h3>Kaynakça</h3><ul><li>FAO. (2019). <a href="https://assets.nena.fao.org/public/2022-10/digital_technologies_in_agriculture_and_rural_areas_digital_technologies_in_agriculture_and_rural_areas__0.pdf"><em>Digital Technologies in Agriculture and Rural Areas: Status Report</em></a>.</li><li>Jouanjean, M. A., Casalini, F., Wiseman, L. ve Gray, E. (2021). <a href="https://knowledge4policy.ec.europa.eu/publication/issues-around-data-governance-digital-transformation-agriculture_en"><em>Issues Around Data Governance in the Digital Transformation of Agriculture</em></a>.</li><li>Naresh, M. ve ark. (2026). <a href="https://doi.org/10.1016/j.aiia.2026.08.005">Artificial intelligence in modern agriculture: Recent advances, challenges, and future directions</a>. <em>Artificial Intelligence in Agriculture</em>.</li><li>OECD. (2022). <a href="https://www.oecd.org/content/dam/oecd/en/publications/reports/2022/04/the-digitalisation-of-agriculture_dd2a1973/285cc27d-en.pdf"><em>The Digitalisation of Agriculture: A Literature Review and Emerging Policy Issues</em></a>.</li><li>Ruder, S. L., Faxon, H. O., Orzel, E. C., Devkota, R. ve Bronson, K. (2026). <a href="https://www.nature.com/articles/s44264-026-00128-x">Reviewing the evidence on precision agriculture and environmental sustainability</a>. <em>npj Sustainable Agriculture, 4</em>, 9.</li><li>T.C. Tarım ve Orman Bakanlığı. (2025). <a href="https://cdniys.tarimorman.gov.tr/api/File/GetGaleriFile/505/DosyaGaleri/8547/iv._tarim_orman_surasi__sura_kitabi.pdf"><em>IV. Tarım Orman Şûrası Kitabı</em></a>.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f32e9c132f99" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From Re-Signification to Removal: Contesting the Monument to the Soviet Army in Sofia, 2011–2023]]></title>
            <link>https://medium.com/@batuhanayrba/from-re-signification-to-removal-contesting-the-monument-to-the-soviet-army-in-sofia-2011-2023-c9b673fe3f98?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/c9b673fe3f98</guid>
            <category><![CDATA[bulgaria]]></category>
            <category><![CDATA[soviet-union]]></category>
            <category><![CDATA[russia]]></category>
            <category><![CDATA[history]]></category>
            <category><![CDATA[ukraine]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Sun, 30 Aug 2026 18:14:59 GMT</pubDate>
            <atom:updated>2026-11-19T22:07:37.435Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*f0tvCiVvNRYCP87u4HxJsA.jpeg" /></figure><h3>Abstract</h3><p>This article examines the changing political status of the Monument to the Soviet Army in Sofia between 2011 and 2023. Rather than treating the monument’s post-socialist history as a straightforward process of decommunisation culminating in removal, it argues that the monument passed through several overlapping regimes of signification and governance. Originally constructed as a monumental expression of Bulgarian gratitude to and political alignment with the Soviet Union, after 1989 it became a disputed material inheritance whose meaning could no longer be secured by a single authoritative narrative. The period beginning in 2011 is especially important because interventions on the monument’s accessible reliefs transformed the site into a recurring surface of public re-signification. Graffiti, projections, political inscriptions, acts of cleaning, commemorations and ordinary uses of the surrounding park repeatedly produced, challenged and temporarily restored different meanings.</p><p>The thesis analyses this transformation through a qualitative historical and interpretive case study combining scholarship on memory and contentious heritage with visual interventions documented between 2011 and 2023, Sofia Municipal Council Decision №168 of 9 March 2023, and the administrative disputes surrounding the partial dismantling of the monument in December 2023. The analysis draws particularly on the concepts of sites of memory, dissonant heritage and heritage as cultural practice. It also engages Bulgarian-focused scholarship that has approached the monument as a politically contested memorial, an appropriated urban space and a recursively transformed material surface.</p><p>The central argument is that the monument’s trajectory cannot adequately be described as the gradual disappearance of communist meaning. Public interventions after 2011 weakened the capacity of its original inscription and sculptural programme to monopolise interpretation, but they did not replace it with a stable consensus. The municipal relocation decision of 2023 constituted a qualitatively different intervention: it attempted to transfer the monument from the category of active central-city commemoration into that of historically interpreted and spatially displaced heritage. At the same time, the legal and administrative difficulties surrounding dismantling demonstrated that symbolic delegitimation did not automatically produce uncontested institutional authority. The case therefore reveals a distinction between changing what a monument means, changing who has authority to interpret it, and changing what may legally and materially be done with it.</p><p>The thesis concludes that the Monument to the Soviet Army evolved from an authorised ideological monument into a contested public surface and finally into an object of attempted institutional reclassification. Removal, in this interpretation, does not end memory politics. It is itself an act of memory politics.</p><p><strong>Keywords:</strong> Bulgaria; Sofia; Monument to the Soviet Army; post-socialism; collective memory; contentious heritage; monuments; public space; decommunisation; memory politics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/267/1*AmZknUyy-6LVZixOyjC-QA.gif" /></figure><h3>Introduction</h3><p>Political monuments are designed to make particular interpretations of the past durable. Their size, material, location and iconography allow political claims to survive beyond the speech, ceremony or regime that initially produced them. Yet the durability of stone and bronze should not be confused with the durability of meaning. Political systems change, audiences change, public space acquires new functions, and the historical narratives embodied by monuments can lose authority even while the structures themselves remain physically intact.</p><p>Few objects illustrate this distinction in contemporary Bulgaria as clearly as the Monument to the Soviet Army in Sofia. Planning for the monument began in the early post-war period; the Sofia municipality made an initial decision concerning its construction in 1946, the national government confirmed the project in 1949, and construction was completed in 1954. Its central location, monumental scale and sculptural programme presented Soviet military power together with Bulgarian figures and an inscription expressing gratitude to the Soviet Army as liberator. Daniela Decheva interprets the monument within a broader socialist memory regime in which Soviet liberation, Bulgarian political transformation and dependence upon the Soviet Union were joined into a coherent public narrative. (<a href="https://aph-ihpan.edu.pl/images/APH126/05_Decheva.pdf">aph-ihpan.edu.pl</a>)</p><p>The fall of Bulgaria’s communist regime in 1989 destroyed the political system that had guaranteed this narrative but did not remove the monument itself. This created a characteristic post-socialist problem. The state could cease officially to reproduce the ideological system from which a monument derived without simultaneously agreeing on what should happen to the material structure that remained. In March 1993 the Sofia Municipal Council adopted a decision concerning the dismantling of the monument’s sculptural elements, but implementation proved substantially more complicated than the political declaration. A 2011 administrative case concerning efforts to compel implementation reproduced arguments that the older decision could not simply be executed without the appropriate legal procedures and permissions and that questions of state ownership constrained municipal action. (<a href="https://citybuild.bg/news/reshi-pametnikyt-syvetskata?utm_source=chatgpt.com">CITYBUILD.BG</a>)</p><p>The result was neither preservation in the strong sense of a socially uncontested memorial nor successful removal. The Monument to the Soviet Army instead remained physically prominent while its symbolic status became unstable. Decheva describes the broader Bulgarian post-communist landscape as characterised by the absence of a coherent memory policy and by competing approaches to socialist memorials: removal as a symbolic break with communism on the one hand, and preservation combined with desacralisation or reinterpretation on the other. The Sofia monument became particularly significant because arguments about it simultaneously concerned domestic interpretations of communism and Bulgaria’s relationship with Russia.</p><p>This thesis focuses on the period from 2011 to 2023. The choice of 2011 does not suggest that conflict over the monument began then. Rather, it identifies a qualitative change in the form of contestation. On 17 June 2011, an accessible relief depicting Soviet soldiers was repainted so that the figures appeared as characters and symbols drawn from Western popular culture. The intervention became widely known through its Bulgarian caption <em>V krak s vremeto</em>, usually translated as “In Pace with the Times” or “Abreast of the Times.” Neda Genova treats this intervention and subsequent modifications of the same surface between 2011 and 2018 as a series of material-semiotic operations capable of producing both continuity and rupture. The intervention was therefore more than an isolated joke directed at a socialist monument. It marked the emergence of the monument as a surface that could repeatedly be rewritten. (<a href="https://eprints.soton.ac.uk/494750/1/Temporal_Politics_of_the_Surface_-_Genova.pdf?utm_source=chatgpt.com">Eprints Soton</a>)</p><p>Subsequent interventions connected the monument to controversies extending far beyond Bulgaria’s communist past. The relief was used in solidarity with Pussy Riot; in 2013 it was painted pink with a message apologising for Bulgaria’s participation in the suppression of the Prague Spring; in 2014 elements were painted in Ukrainian and Polish colours in interventions referencing Ukraine, Crimea and the Katyn massacre; and after Russia’s full-scale invasion of Ukraine in 2022 the monument again became a target for anti-war and anti-Russian messages. These interventions did not all express the same politics. What united them was their treatment of the monument not as a settled historical statement but as an already meaningful surface that could be appropriated precisely because its inherited symbolism remained recognisable. (<a href="https://www.rferl.org/a/sofia-soviet-monument-vandalism-ukraine-war-protest/32290150.html?utm_source=chatgpt.com">RadioFreeEurope/RadioLiberty</a>)</p><p>At the same time, contestation was not limited to spectacular acts of graffiti. Research on the everyday use of the monument and surrounding space demonstrates that younger residents, political initiatives, cultural events, skaters, partygoers and other urban users could appropriate the site for purposes that did not necessarily reproduce either Soviet commemoration or explicit anti-communist opposition. Kristina Dimitrova describes such uses as bottom-up transformations of urban space, while Decheva similarly observes that the area became incorporated into ordinary practices including gatherings, cycling, skating and public events. (<a href="https://www.ceeol.com/search/article-detail?id=423986&amp;utm_source=chatgpt.com">CEEOL</a>)</p><p>The relationship between these forms of re-signification and formal political action changed decisively after Russia’s invasion of Ukraine in February 2022. Although campaigns for removal had existed for decades, the new geopolitical environment intensified the monument’s contemporary political salience. In March 2023 Sofia Municipal Council adopted Decision №168. It requested that the state relocate the monument, proposed permanent exhibition of the sculptural figures and reliefs at the Museum of Art from the Socialist Period or another appropriate state-owned location outside the central urban area, and called for an informational panel at the existing site. Significantly, the proposed panel was not politically neutral. The council’s decision required an interpretation presenting the monument in relation to a foreign power that had declared war on Bulgaria and occupied the country, with the detailed historical text to involve academic institutions. The decision therefore concerned not merely physical relocation but also official recontextualisation. (<a href="https://council.sofia.bg/documents/d/guest/r-168?utm_source=chatgpt.com">Sofia Council</a>)</p><p>Dismantling of the monument’s upper sculptural group began in December 2023. Regional authorities additionally referred to deterioration and structural concerns, while the process provoked demonstrations and political opposition. On 19 December the Administrative Court Sofia-City ordered the cessation of dismantling actions in proceedings brought by opponents, finding on the evidence presented to it that the relevant actions had not been shown to rest on the necessary legal or administrative basis and raising questions concerning storage, restoration, expenditure and supervision. The procedural history did not end there: in 2024 the Supreme Administrative Court set aside lower-court orders on admissibility and legal-interest grounds, meaning that the December rulings should not be misrepresented as a final judicial determination that removal was substantively unlawful. (<a href="https://sofia-adms-g.justice.bg/bg/news1/28042?utm_source=chatgpt.com">Sofia Adliyesi</a>)</p><p>This sequence poses the central research question of the thesis:</p><p><strong>How did public re-signification and institutional action transform the political status of Sofia’s Monument to the Soviet Army between 2011 and 2023?</strong></p><p>The thesis advances a three-part answer. First, after 2011 the monument increasingly functioned as a recursively writable public surface. Its original commemorative programme remained materially present, but its capacity to regulate interpretation was repeatedly interrupted by interventions that inserted new historical and geopolitical references. Second, the multiplicity of uses and interventions did not produce a new consensus. It produced a condition of contested meaning in which preservation, ridicule, everyday appropriation, anti-Russian protest, anti-communist politics and pro-Russian commemoration could coexist around the same object. Third, Decision №168 represented an attempt to resolve this instability institutionally. Rather than simply destroying an unwanted monument, the decision sought to separate the object from its central commemorative location, relocate its sculptural elements and replace the original spatial narrative with a new official historical interpretation.</p><p>The working hypothesis is therefore that the period from 2011 to 2023 should not be understood as a linear process in which communist symbolism gradually disappeared. The more consequential transformation was a shift in <strong>authority over meaning</strong>. The socialist state had originally attempted to fix both the monument and its interpretation. Post-socialist public interventions fractured that interpretive monopoly. The municipal decision of 2023 then attempted to establish a new authorised interpretation, while legal disputes exposed the limits of translating symbolic authority into administrative action.</p><p>The thesis is organised into five chapters. Chapter One establishes the theoretical and historiographical framework through memory studies, heritage studies and research on Bulgarian post-socialist monuments. Chapter Two explains the case-study method, source hierarchy and analytical limitations. Chapter Three examines the material and symbolic re-signification of the monument between 2011 and 2018. Chapter Four traces the relationship between continued contestation, the post-2022 geopolitical context and the institutional turn represented by Decision №168. Chapter Five examines the December 2023 dismantling process and its legal-administrative complications. The conclusion returns to the distinction between removal, reinterpretation and institutional authority.</p><h3>Chapter One</h3><h3>Monument, Memory and Contentious Heritage</h3><h3>1.1 Monuments and the problem of durable meaning</h3><p>A monument is simultaneously material and interpretive. Its physical form may be stable for decades while the social frameworks through which it is understood alter radically. This distinction is essential to the present case because the Monument to the Soviet Army remained in approximately the same central urban location for almost seven decades while the political systems surrounding it changed from state socialism to parliamentary democracy and from Soviet-bloc geopolitical alignment to membership in NATO and the European Union.</p><p>Pierre Nora’s concept of <em>lieux de mémoire</em> provides one useful starting point. Nora’s well-known distinction between lived environments of memory and deliberately constituted sites of memory directs attention toward the institutions, rituals and objects through which societies attempt to preserve relations with the past once such relations can no longer be assumed to reproduce themselves automatically. A monument is therefore not important merely because it refers to an event. Its political significance lies partly in the work required to maintain the relationship among object, narrative and community. (<a href="https://doi.org/10.2307%2F2928520?utm_source=chatgpt.com">Doi</a>)</p><p>Yet the Sofia monument also shows the limitations of approaching monuments only as repositories of collective memory. The analytical question is not simply what memory a monument contains. A monument can become politically consequential because different groups reject the proposition that it should contain one shared memory at all. Nina Debruyne and Georgeta Nazarska explicitly analyse communist monuments in Sofia as “contentious heritage spaces” and sites in which ex-communist and anti-communist political subcultures sustain conflicting relationships with the socialist past. Their research emphasises ongoing reinterpretation rather than the passive survival of an inherited object. (<a href="https://www.sciencedirect.com/science/article/pii/S0305748823000889?utm_source=chatgpt.com">ScienceDirect</a>)</p><p>The distinction between memory and heritage is important here. Heritage should not be understood only as a collection of old objects possessing intrinsic value. Laurajane Smith’s work challenges object-centred definitions of heritage and instead emphasises heritage as cultural practice and meaning-making. From this perspective, the physical monument is a necessary but insufficient unit of analysis. The acts performed around it — ceremonies, campaigns, artistic interventions, cleaning, vandalism, legal classification, relocation and museum display — participate in creating heritage value and meaning. (<a href="https://books.google.com.tr/books?hl=tr&amp;id=hI42mQEACAAJ&amp;utm_source=chatgpt.com">Google Kitaplar</a>)</p><p>This framework is especially productive for analysing a monument whose official message has ceased to enjoy uncontested political authority. The crucial problem is no longer whether the monument “has” a meaning. It is who is authorised to make meaning through it. Under socialism, the state largely controlled the design, placement, inscription and ceremonial interpretation of the monument. Under post-socialism, no comparable monopoly survived.</p><p>The result resembles what J. E. Tunbridge and G. J. Ashworth conceptualise as dissonant heritage: heritage whose meaning, ownership or interpretation is disputed among communities and institutions. Dissonance is not an accidental disturbance applied to an otherwise neutral monument. It is often intrinsic to the heritage process because selecting and interpreting one past can privilege some identities and marginalise others. (<a href="https://obnb.uk/p8074180-dissonant-heritage-the-management-of-the-past-as-a-resource-in-conflict?utm_source=chatgpt.com">OBNB</a>)</p><p>The Monument to the Soviet Army is therefore analytically significant not because it is unusually incapable of communicating meaning, but because it communicates too many politically incompatible meanings at once. To defenders, it may represent Soviet victory over fascism, historical Bulgarian-Russian ties and anti-fascist remembrance. To opponents, it may represent Soviet occupation, communist domination and imposed political dependence. To artistic interveners, its recognisable iconography may function as material for ironic commentary. To younger users of the park, it may be partially detached from both interpretations and incorporated into everyday urban life.</p><p>The same material object can sustain all these relations because meaning is not produced by materiality alone.</p><h3>1.2 From authorised commemoration to post-socialist uncertainty</h3><p>The monument’s original ideological programme was comparatively explicit. Decheva notes that its construction was intended to symbolise Bulgarian loyalty and gratitude toward the Soviet Union and to connect the Soviet entry into Bulgaria in 1944 with a liberation narrative. The original inscription likewise formally attributed gratitude to the Bulgarian people. Architecturally and iconographically, the monument placed Soviet military authority in a commanding spatial position. (<a href="https://aph-ihpan.edu.pl/images/APH126/05_Decheva.pdf">aph-ihpan.edu.pl</a>)</p><p>This can be understood as an authorised commemorative regime. The state selected the event to be memorialised, determined the central urban location, funded and organised the construction, established the iconography and normalised the terminology through which the event would be remembered. The monument did not invite the viewer to decide whether the Soviet Army had “liberated” Bulgaria. The claim was incorporated into the monument’s very structure.</p><p>Political transition after 1989 removed the institutional environment that had made this reading authoritative. What followed, however, was not the immediate creation of a coherent post-communist alternative. Decheva argues that Bulgarian memory politics remained divided and inconsistent, particularly regarding communist rule and its monuments. Local authorities faced the practical and political consequences of deciding whether socialist monuments should be removed, preserved, reinterpreted or simply left in place. (<a href="https://aph-ihpan.edu.pl/images/APH126/05_Decheva.pdf">aph-ihpan.edu.pl</a>)</p><p>The distinction matters because doing nothing is also a memory policy, even when it is not presented as one. Leaving an ideologically explicit monument in a central urban location after the collapse of the regime that created it does not freeze history. It creates a new political relationship between monument and state. The absence of removal may be interpreted as tolerance, indecision, preservation, institutional weakness or historical pluralism depending on the observer.</p><p>The 1993 municipal decision illustrates this ambiguity particularly clearly. Politically, the decision represented an attempt to remove at least the sculptural elements from their dominant location. Administratively, implementation encountered questions of permits, authority and ownership. When activists attempted years later to compel municipal implementation, courts did not treat the 1993 political decision as a simple executable command detached from subsequent legal frameworks. (<a href="https://citybuild.bg/news/reshi-pametnikyt-syvetskata?utm_source=chatgpt.com">CITYBUILD.BG</a>)</p><p>This gap between declarative politics and administrative capability would reappear thirty years later. The history of the monument is therefore not only a history of competing historical narratives. It is also a history of jurisdiction.</p><h3>1.3 Bulgarian post-socialist memory as an unsettled field</h3><p>Research on Bulgarian communist monuments repeatedly emphasises the coexistence of incompatible strategies toward the socialist past. Decheva contrasts removal with preservation and redefinition. Debruyne and Nazarska identify divergent political subcultures whose relationships with communist monuments are structured by different memories of political transition. Their 2020 qualitative research treats the Monument to the Soviet Army and the former Georgi Dimitrov Mausoleum as particularly revealing sites through which the post-1989 transformation of political memory can be studied. (<a href="https://www.sciencedirect.com/science/article/pii/S0305748823000889?utm_source=chatgpt.com">ScienceDirect</a>)</p><p>The comparison with the Dimitrov Mausoleum is instructive. The mausoleum was demolished in 1999. Yet demolition did not necessarily resolve the memory problem. Decheva observes that the empty space continued to be identified through reference to the absent structure. Physical erasure could therefore generate a new memory of absence rather than eliminate the former object from collective reference. (<a href="https://aph-ihpan.edu.pl/images/APH126/05_Decheva.pdf">aph-ihpan.edu.pl</a>)</p><p>This point has direct implications for the Monument to the Soviet Army. It weakens a simplistic opposition between preservation, which supposedly remembers, and demolition, which supposedly forgets. Both strategies can generate new commemorative forms. Removal can become an event that is remembered; an empty plinth can become politically meaningful; a relocated statue can acquire museum status; a photograph of graffiti removed the following morning can circulate for years after the painted surface itself disappears.</p><p>Sharon Macdonald’s broader account of contemporary Europe as a landscape saturated with practices for materialising, presenting and renegotiating the past is useful in this regard. Modern societies do not merely preserve memory objects; they continually create new settings in which the past becomes present. (<a href="https://www.routledge.com/Memorylands-Heritage-and-Identity-in-Europe-Today/Macdonald/p/book/9780415453332?utm_source=chatgpt.com">Routledge</a>)</p><p>The important analytical unit is therefore not the monument alone but the changing <strong>assemblage of monument, location, interpretation, practice and authority</strong>.</p><p>This thesis uses the term <strong>re-signification</strong> to describe actions that alter the publicly available interpretation of the monument without necessarily changing its permanent legal status. Painting a relief, projecting an image, covering an inscription, attaching a political message, cleaning the surface or holding a commemorative event can all participate in re-signification.</p><p>By contrast, <strong>reclassification</strong> refers here to an institutional attempt to alter the monument’s official relationship with public space. Moving sculptural elements from a central commemorative site to a museum or peripheral state property does more than create another interpretation. It changes the category in which the object is publicly encountered: from active central monument to historical object under a new interpretive regime.</p><p>The transition from repeated re-signification toward attempted institutional reclassification forms the main analytical arc of the thesis.</p><h3>Chapter Two</h3><h3>Methodology and Source Construction</h3><h3>2.1 Research design</h3><p>This thesis uses a qualitative historical-interpretive case-study design. Its object is deliberately narrow: the Monument to the Soviet Army in Sofia between June 2011 and December 2023. It does not seek to produce a comprehensive history of Bulgarian-Russian relations, a representative survey of Bulgarian public opinion, or an exhaustive inventory of communist monuments in Bulgaria.</p><p>The methodological premise is that different sources answer different questions. Visual interventions provide evidence about the ways actors attempted to alter the monument’s public signification. Academic studies provide analytical interpretations and, in some cases, empirical research into uses and attitudes. Municipal and judicial documents provide evidence about formal institutional decisions. Contemporary reporting is used selectively for chronology, documented actions and publicly reported statements where primary institutional records are incomplete or inaccessible.</p><p>This separation is important because the sources are not interchangeable. A newspaper photograph can demonstrate that a relief was painted but cannot establish national public opinion. A municipal decision can establish what the council formally resolved but cannot demonstrate that the decision enjoyed social consensus. An academic article can offer an interpretation of memory politics but does not become primary evidence of an event merely because it analyses that event.</p><h3>2.2 Primary and secondary corpus</h3><p>The central institutional primary source is Sofia Municipal Council Decision №168 of 9 March 2023. The decision is particularly valuable because it records not only the request for relocation but the proposed future disposition of the sculptures and the intended interpretive treatment of the existing site. It therefore provides direct evidence that relocation and historical reinterpretation were institutionally connected. (<a href="https://council.sofia.bg/documents/d/guest/r-168?utm_source=chatgpt.com">Sofia Council</a>)</p><p>The second institutional source group consists of administrative-court communications concerning the December 2023 dismantling. The Administrative Court Sofia-City’s announcement of 19 December identifies procedural and evidentiary concerns raised during the dismantling process. Later Supreme Administrative Court decisions are used cautiously as a post-period clarification of the procedural status of those rulings rather than as part of the thesis’s 2011–2023 event sequence. (<a href="https://sofia-adms-g.justice.bg/bg/news1/28042?utm_source=chatgpt.com">Sofia Adliyesi</a>)</p><p>The visual-intervention corpus is reconstructed from peer-reviewed academic studies and documented photographic chronologies. Neda Genova’s analysis of interventions between 2011 and 2018 provides the principal theoretical treatment of the repeatedly modified relief. Mina Ivanova’s analysis of two visual political interventions offers a rhetorical interpretation of the monument as a site where subversive and conservative forms of comedy interact with ideological formations. Other studies and documentary sources help reconstruct the wider sequence of interventions. (<a href="https://eprints.soton.ac.uk/494750/1/Temporal_Politics_of_the_Surface_-_Genova.pdf?utm_source=chatgpt.com">Eprints Soton</a>)</p><p>For the social use of the site, the thesis draws on Kristina Dimitrova’s research concerning the appropriation of the monument and surrounding urban space, together with the later work of Debruyne and Nazarska. Dimitrova highlights civic events, street-art practices, leisure and everyday activity; Debruyne and Nazarska place such transformations within a broader study of political subcultures and contentious communist heritage. (<a href="https://www.ceeol.com/search/article-detail?id=423986&amp;utm_source=chatgpt.com">CEEOL</a>)</p><h3>2.3 Analytical limits</h3><p>The thesis deliberately avoids several claims that the available corpus cannot support.</p><p>First, it does not infer Bulgarian public opinion from highly visible interventions. The people who painted the monument were not a statistically representative sample of Sofia residents. A spectacular act can be politically important without being socially representative.</p><p>Second, the thesis does not assume that every person using the park has consciously reinterpreted the monument. Everyday appropriation can reduce the practical centrality of ideological meaning, but a skateboarder passing through the site should not automatically be converted into evidence of a political position.</p><p>Third, no interviews have been conducted for this study. Statements about the motivations of monument defenders, removal advocates or artists are therefore made only where those motivations are documented in the consulted literature or public records.</p><p>Fourth, the thesis does not attempt to determine a single legally correct description of every stage of the dismantling process. The legal record is procedurally complex. The purpose of the legal material here is narrower: to demonstrate that institutional reclassification required administrative authority, documentation, procedures and standing that could themselves become objects of contestation.</p><p>Finally, the thesis treats “meaning” as publicly produced rather than psychologically possessed. It does not attempt to determine what every individual visitor privately thought about the monument. It examines the meanings made visible through material intervention, political discourse, official action and structured use of space.</p><h3>Chapter Three</h3><h3>Re-Signifying the Monument, 2011–2018</h3><h3>3.1 “In Pace with the Times”</h3><p>The intervention of 17 June 2011 provides the clearest starting point for the period examined here because it condensed several decades of political transformation into a single altered surface.</p><p>A relief depicting Soviet military figures was repainted so that the soldiers appeared as icons associated with Western popular culture. The transformation included comic-book and commercial figures and altered the Soviet visual vocabulary through the insertion of contemporary Western imagery. Beneath the figures appeared the caption <em>V krak s vremeto</em>: “In Pace with the Times.” Genova identifies the intervention as the opening moment of the series she analyses and emphasises that the artists later became associated with the name Destructive Creation. (<a href="https://eprints.soton.ac.uk/494750/1/Temporal_Politics_of_the_Surface_-_Genova.pdf?utm_source=chatgpt.com">Eprints Soton</a>)</p><p>The intervention operated because it did not destroy the underlying Soviet composition. It depended upon it.</p><p>If the original relief had been removed and replaced with a new mural, the political mechanism would have been different. Instead, the recognisable poses and military composition remained structurally present while the identities attached to the figures changed. The effect was produced through superimposition.</p><p>This makes the intervention a particularly useful example of re-signification. The original monument claimed historical authority through permanence. The 2011 action demonstrated that permanence of material did not guarantee permanence of signification.</p><p>Mina Ivanova’s analysis of visual interventions on the monument similarly treats the relationship among irony, burlesque, ideological formation and critical subjectivity as more complicated than simple opposition. The political significance of humour lies partly in its ability to expose the instability of supposedly self-evident symbolic relations. (<a href="https://www.ingentaconnect.com/content/routledg/rqjs/2014/00000100/00000003/art00001?utm_source=chatgpt.com">ingentaconnect.com</a>)</p><p>The Western popular-cultural imagery is therefore analytically ambiguous. One possible reading treats the intervention as a triumphant replacement of Soviet ideology by American consumer capitalism: Soviet soldiers become superheroes and commercial icons because Bulgaria has moved geopolitically and culturally “with the times.” But that interpretation risks assuming that the American symbols themselves are presented without irony.</p><p>A second reading sees the juxtaposition as mutually destabilising. The heroic grammar of Soviet monumental realism and the heroic grammar of superhero culture become visually interchangeable. The monument’s claim to solemn historical authority is reduced by the transformation, yet the Western imagery is also exposed as a competing symbolic system capable of occupying the same poses.</p><p>Genova’s approach is particularly valuable because it moves beyond the question of what the intervention “really meant” and focuses instead on what repeated modification does to the surface over time. The relief becomes a temporal object. Cleaning does not restore an innocent original condition because the cleaned surface now carries the memory of having been altered. A subsequent intervention takes place not only upon a Soviet relief but upon a surface already known to have been politically rewritten. (<a href="https://eprints.soton.ac.uk/494750/1/Temporal_Politics_of_the_Surface_-_Genova.pdf?utm_source=chatgpt.com">Eprints Soton</a>)</p><p>The 2011 intervention therefore established a precedent. It showed that the monument’s political meaning could be attacked without removing the monument itself.</p><h3>3.2 Cleaning as counter-intervention</h3><p>The removal of graffiti might initially appear to be a neutral act of maintenance. In a contested heritage context, it is not necessarily neutral.</p><p>Cleaning restores the visual accessibility of the official sculptural programme. It removes one layer of unauthorised signification and materially reasserts the earlier surface. Yet after a high-profile intervention, restoration does not erase the event. Photographs circulate independently of the physical paint. Media coverage, scholarly analysis and subsequent imitation extend the life of the intervention.</p><p>The cycle therefore becomes:</p><p><strong>monumental claim → intervention → restoration → remembered intervention → new intervention.</strong></p><p>The surface is not returned to year zero.</p><p>This recursive structure helps explain why repeated alteration between 2011 and 2018 should be analysed as more than a sequence of vandalism incidents. Each intervention inherits not only the monument’s socialist symbolism but the recent history of previous interventions.</p><p>The state or organisations responsible for cleaning also become participants in the politics of signification. Restoring the monument can be interpreted by critics as restoring propaganda, while leaving graffiti intact can be interpreted by defenders as official tolerance of desecration. Maintenance itself becomes politically legible.</p><p>In this respect, the post-socialist monument is structurally different from the monument under socialism. The original object was created within a political regime that attempted to align material maintenance with ideological maintenance. In the post-socialist setting, material maintenance may preserve an object whose original ideological claim the state itself no longer officially endorses.</p><p>That separation produces a governance dilemma that will become central again in 2023.</p><h3>3.3 Transnational memory written onto a Soviet relief</h3><p>Later interventions further expanded the temporal and geographical range of the monument.</p><p>In 2012 visual modifications associated the relief with support for Pussy Riot. In August 2013 the monument was painted pink and accompanied by a message of Bulgarian apology connected to the Warsaw Pact suppression of the Prague Spring in 1968. Decheva notes that the intervention deliberately invoked the historical participation of communist Bulgaria in that suppression while also recalling earlier Czech artistic and protest practices involving Soviet military monuments. (<a href="https://www.rferl.org/a/sofia-soviet-monument-vandalism-ukraine-war-protest/32290150.html?utm_source=chatgpt.com">RadioFreeEurope/RadioLiberty</a>)</p><p>This event is especially significant because it demonstrates how the Sofia monument became a carrier of transnational memory.</p><p>The relief was not built to commemorate the Prague Spring. Yet because it represented Soviet military power and because socialist Bulgaria had participated in the 1968 intervention, the surface could be made to connect 1944, 1954, 1968, 1991 and 2013. The political act depended on collapsing these times into a single visible object.</p><p>This challenges any strictly local reading of the monument. The monument stood in Sofia, but its iconography belonged to a wider Soviet and post-Soviet symbolic field. Russian diplomatic reactions to interventions further internationalised disputes that might otherwise have remained municipal questions. Decheva explicitly identifies relations with Russia as one of the features distinguishing the Soviet Army monument from memorials whose controversies were primarily domestic. (<a href="https://aph-ihpan.edu.pl/images/APH126/05_Decheva.pdf">aph-ihpan.edu.pl</a>)</p><p>In 2014 the same broader logic appeared again through interventions referring to Ukraine, Crimea, Poland and the Katyn massacre. Visual documentation shows figures painted in Ukrainian and Polish national colours and slogans addressing Russian actions in Ukraine. (<a href="https://www.rferl.org/a/sofia-soviet-monument-vandalism-ukraine-war-protest/32290150.html?utm_source=chatgpt.com">RadioFreeEurope/RadioLiberty</a>)</p><p>Each intervention therefore added another historical layer without physically redesigning the monument. The site accumulated references.</p><p>This accumulation is central to the argument of the thesis. By the end of the 2010s, the monument no longer functioned only as an object through which Bulgarians debated 1944 or state socialism. It had become an established medium through which contemporary geopolitical conflicts involving Russia could be rendered visible in Sofia.</p><h3>3.4 The monument as urban infrastructure</h3><p>Focusing exclusively on spectacular political interventions would, however, produce a distorted picture.</p><p>The monument occupied a substantial central urban site. Its accessibility and surrounding open space made it available not only for commemoration and protest but for everyday use. Dimitrova’s research emphasises civic initiatives, public events, street-art practices, leisure, social gatherings and urban sports as forms through which younger users appropriated the area. The site could therefore become associated with forms of freedom, alternative lifestyles and ordinary sociability rather than solely with the political programme embedded in its sculptures. (<a href="https://www.ceeol.com/search/article-detail?id=423986&amp;utm_source=chatgpt.com">CEEOL</a>)</p><p>Decheva reaches a related observation when describing walking, gatherings, biking, skating and events around the monument. She suggests that younger generations without direct experience of socialism or the early transition could interact with the space in ways that reduced the ideological gravity of the original structure. (<a href="https://aph-ihpan.edu.pl/images/APH126/05_Decheva.pdf">aph-ihpan.edu.pl</a>)</p><p>This should not be romanticised as the final neutralisation of ideology. Ordinary use did not prevent intense political controversy from returning.</p><p>Instead, it demonstrates that public space is capable of supporting multiple degrees of political attention. The same person may use the monument as a meeting point one day and encounter it as a geopolitical symbol during a protest the next.</p><p>Everyday appropriation therefore differs from deliberate artistic intervention but contributes to the same broader transformation: the socialist state no longer controls the social script of the site.</p><p>The monument’s centrality becomes functionally plural.</p><h3>3.5 From single message to contested platform</h3><p>By 2018 the monument could be understood as having undergone a change in communicative structure.</p><p>Its original design had been <strong>monologic</strong>. A vertically dominant Soviet soldier, accompanying Bulgarian figures, historical reliefs and an inscription of gratitude collectively directed interpretation toward a specific narrative of liberation and political alliance.</p><p>The post-2011 site increasingly behaved as a <strong>platform</strong>. Different actors could temporarily occupy its surface, invoke different historical episodes, attach new conflicts to it or use the surrounding space for purposes unrelated to the original commemorative programme.</p><p>This did not mean that the original message disappeared. On the contrary, the interventions worked because it remained legible.</p><p>The transformation can therefore be expressed more precisely as a reduction in the original message’s <strong>exclusive authority</strong>, not its erasure.</p><p>This distinction explains why repeated re-signification did not settle the political controversy. A surface that can be rewritten remains available for another rewriting.</p><p>To end the cycle would require a different type of intervention — one capable of altering the institutional status of the monument itself.</p><h3>Chapter Four</h3><h3>From Recurring Contestation to Institutional Action, 2019–2023</h3><h3>4.1 Persistence without settlement</h3><p>The continuing presence of the monument after decades of debate did not indicate agreement. It indicated that disagreement had become institutionally durable.</p><p>Debruyne and Nazarska’s research describes communist monuments in Sofia as functioning realms of memory within an ongoing process of reinterpretation. Their work highlights a fundamental divergence among political and social formations regarding how communist monumental heritage should be understood and managed. (<a href="https://www.sciencedirect.com/science/article/pii/S0305748823000889?utm_source=chatgpt.com">ScienceDirect</a>)</p><p>This helps explain why the Monument to the Soviet Army survived long after the political collapse of the regime that produced it.</p><p>Physical persistence should not automatically be interpreted as ideological continuity. Several different conditions can produce the same material outcome. A monument may remain because it is valued, because removal is unpopular, because institutions cannot agree, because ownership is disputed, because legal procedures are complex, because the political cost of action is greater than the cost of inaction, or because the space has acquired alternative uses.</p><p>In Sofia several of these mechanisms overlapped.</p><p>The 1993 decision had demonstrated political willingness to dismantle, but later administrative proceedings demonstrated that old political resolutions did not automatically override legal constraints. At the same time, defenders of the monument continued to use it for commemorative practices and articulate interpretations centred on Soviet victory and anti-fascism. Decheva documents such positions in the early 2020s and emphasises their continuing connection with broader Bulgarian-Russian memory politics. (<a href="https://aph-ihpan.edu.pl/images/APH126/05_Decheva.pdf">aph-ihpan.edu.pl</a>)</p><p>The monument therefore entered the 2020s in a condition that can be described as <strong>institutional suspension</strong>. It was neither fully rehabilitated as an uncontested public memorial nor successfully transferred into a museum context.</p><h3>4.2 February 2022 and the compression of historical time</h3><p>Russia’s full-scale invasion of Ukraine in February 2022 altered the context in which Soviet military monuments across Europe were interpreted. In Sofia the effect was especially immediate because the monument’s iconography already linked Soviet military power to claims of liberation.</p><p>The reliefs were again painted. Photographic documentation from 2022 shows Ukrainian colours, red paint and anti-war or anti-occupation messages. The earlier visual vocabulary of 2014 therefore returned in a more intense geopolitical context. (<a href="https://www.rferl.org/a/sofia-soviet-monument-vandalism-ukraine-war-protest/32290150.html?utm_source=chatgpt.com">RadioFreeEurope/RadioLiberty</a>)</p><p>This produced what may be described as a compression of historical time.</p><p>The monument formally referred to the Soviet Army and the events surrounding 1944. Yet contemporary observers could now read it through the Russian invasion of Ukraine in 2022. The Soviet soldier became available as a symbolic proxy for arguments about present Russian military power.</p><p>Strict historical equivalence between the Soviet Union and the contemporary Russian Federation is not required for this process to occur. Memory politics works through association as well as identity.</p><p>The significance of this shift is institutional. Earlier graffiti interventions had demonstrated the monument’s availability for anti-Russian and transnational messages. After February 2022, however, the relationship between the monument and contemporary geopolitical conflict became sufficiently powerful to alter the political feasibility of formal relocation.</p><p>This does not mean the invasion “caused” the 2023 decision in isolation. Campaigns for dismantling were decades old. What changed was the political environment in which earlier claims could be reactivated.</p><p>Roberta Koleva’s later ethnographic research on the dismantling controversy describes how arguments about the monument became connected with larger imaginaries of Europeanness, democracy, transition and Bulgaria’s geopolitical orientation. Her work is published after the period studied here, but it helps identify the discursive field into which the December 2023 events entered. (<a href="https://anthropology-journal.org/wp/journal/the-bulgarian-berlin-wall-is-going-away-step-by-step-standby-transition-and-the-quest-for-europeanness-and-democracy-in-t?utm_source=chatgpt.com">anthropology-journal.org</a>)</p><p>The monument thus became a device through which different actors narrated not only Bulgaria’s past but Bulgaria’s unfinished or completed “transition.”</p><h3>4.3 Decision №168: relocation as interpretation</h3><p>On 9 March 2023 Sofia Municipal Council adopted Decision №168 concerning the Monument to the Soviet Army.</p><p>The decision requested action by the state to move the monument from the municipal land of Knyazheska Garden. It envisaged permanent exhibition of the sculptural figures and reliefs at the Museum of Art from the Socialist Period or another appropriate state-owned site outside Sofia’s central urban area. It also required institutional steps relating to redesign of the park. (<a href="https://council.sofia.bg/documents/d/guest/r-168?utm_source=chatgpt.com">Sofia Council</a>)</p><p>The most analytically significant feature is that the decision did not end with relocation.</p><p>It additionally approved an informational panel at the site and prescribed the historical orientation of that panel. The monument was to be contextualised as symbolising a foreign power that declared war on Bulgaria and occupied the country, with the detailed text prepared in cooperation with academic historical institutions and reviewed through municipal procedures. (<a href="https://council.sofia.bg/documents/d/guest/r-168?utm_source=chatgpt.com">Sofia Council</a>)</p><p>This transforms the nature of the political act.</p><p>If the council had simply ordered physical destruction, the action might plausibly be interpreted primarily as elimination. Instead, the decision attempted to reorganise the relationship among object, location and narrative.</p><p>The old arrangement was:</p><p><strong>monument + central location + liberation inscription.</strong></p><p>The proposed arrangement became:</p><p><strong>relocated sculptural object + museum/peripheral context + new historical explanation at the former site.</strong></p><p>The strategy was therefore not absence but <strong>recontextualisation through spatial separation</strong>.</p><p>From a heritage-theoretical perspective, this resembles the replacement of one authorised heritage discourse by another. The socialist regime had originally authorised a liberation narrative through monumental form. The post-socialist municipal decision sought to authorise a competing historical interpretation through relocation, academic contextualisation and redesign of public space.</p><p>This is why the term “decommunisation” is useful but insufficient.</p><p>The 2023 decision was certainly anti-communist in the sense that it rejected the continued legitimacy of the inherited monumental narrative. But analytically the more precise process was <strong>institutional reclassification</strong>. The sculptures were not to disappear from historical consciousness. They were to cease functioning as the dominant commemorative structure of that central site.</p><h3>4.4 The museum question</h3><p>Moving a political monument to a museum is often described as a compromise between preservation and removal. The formulation is deceptively simple.</p><p>A museum does not merely store a monument. It alters the conditions under which the monument is encountered.</p><p>In its original location, a monumental object participates in urban life regardless of whether citizens deliberately choose to engage with it. Its scale and placement produce compulsory visibility.</p><p>In a museum, encounter becomes more intentional. Curatorial text can mediate interpretation. The object’s former authority can become an object of explanation rather than a present-tense claim.</p><p>Relocation therefore transforms deixis. The monument no longer declares, in effect, “this is the story of this place.” It becomes “this is an object through which a previous political order represented the story of this place.”</p><p>That grammatical shift from present public statement to historically contextualised object is one of the strongest mechanisms of institutional reclassification.</p><p>The proposed museum destination also introduced practical difficulties, however. A monumental ensemble designed for a large park cannot automatically be transferred into an existing museum without questions of scale, conservation, storage, installation and curatorial balance. Contemporary reporting indicated uncertainty over the practical destination of the sculptural material. (<a href="https://balkaninsight.com/2023/12/12/after-30-years-of-debate-bulgaria-dismantles-red-army-monument/bi/?utm_source=chatgpt.com">Balkan Insight</a>)</p><p>The museum solution was therefore conceptually clear but logistically complex.</p><p>This distinction between symbolic solution and material implementation became increasingly important during the December dismantling.</p><h3>Chapter Five</h3><h3>Dismantling, Law and the Limits of Symbolic Authority</h3><h3>5.1 December 2023</h3><p>Physical dismantling of the Monument to the Soviet Army’s upper sculptural group began on 12 December 2023.</p><p>Authorities referred both to the political relocation process and to deterioration of the structure, including cracks requiring intervention and restoration. Workers began removing the large figures associated with the central composition while police maintained security around the site. The action generated demonstrations from both supporters and opponents. (<a href="https://apnews.com/article/5f0f21b44ad0295eabbcb588a1a7d89a?utm_source=chatgpt.com">AP News</a>)</p><p>The event produced powerful images. A monument whose political meaning had been debated for three decades was visibly losing the sculptural group that had defined its skyline.</p><p>Such images encouraged a narrative of completion: the unresolved Bulgarian transition had finally arrived at the monument.</p><p>But this interpretation is precisely what the administrative history complicates.</p><p>Dismantling did not simply reveal the triumph of one historical interpretation over another. It revealed the infrastructures required to convert an interpretation into material action.</p><p>Questions immediately arose: Who possessed authority over the object? Under which administrative act was each action taken? Where would the components be stored? Which institution would supervise restoration? How would the process be funded? What conservation rules applied?</p><p>These questions are not external bureaucratic details. They are part of the politics of monuments because control over material heritage requires legal competence.</p><h3>5.2 The Administrative Court intervention</h3><p>On 19 December 2023 the Administrative Court Sofia-City issued orders concerning applications to stop dismantling activities.</p><p>The court’s public communication stated that in the proceedings before it, the documentation presented had not adequately established the legal or administrative basis for certain dismantling actions. It specifically identified unanswered questions concerning what would happen to the figures after removal, where and how they would be stored, whether restoration would occur, how allocated funds would be divided among dismantling, storage, protection, restoration and reinstallation, and which officials would exercise control. (<a href="https://sofia-adms-g.justice.bg/bg/news1/28042?utm_source=chatgpt.com">Sofia Adliyesi</a>)</p><p>These concerns are revealing because they expose the difference between a <strong>political mandate</strong> and an <strong>operational chain of authority</strong>.</p><p>Decision №168 established a political direction. It did not by itself eliminate every statutory, ownership, conservation, procurement or administrative requirement involved in physically dismantling state property.</p><p>This echoes the difficulty surrounding the 1993 decision. In both periods, a municipal political intention encountered a legal environment in which execution depended on more than the declaration of intention.</p><p>The recurrence suggests that the monument’s post-socialist endurance cannot be explained only by ideological conflict.</p><p>It was also produced by institutional fragmentation.</p><h3>5.3 A necessary legal qualification</h3><p>The December 19 court intervention must be described carefully.</p><p>It would be inaccurate to state that the Bulgarian judiciary finally declared the dismantling illegal. The procedural history continued into 2024. The Supreme Administrative Court subsequently set aside relevant lower-court orders because the applicants lacked the necessary direct legal interest or because the procedural requirements for the particular remedy had not been satisfied. In one decision, the higher court also noted the existence of a separate 14 December 2023 cultural-heritage administrative measure that had already ordered a halt to activities pending clarification of compliance. (<a href="https://sac.justice.bg/novini/2024/vas-obezsili-razporezhdane-na-assg-s-koeto-po-zhalba-na-sdruzhenie-izpravi-sebg/?utm_source=chatgpt.com">Върховен административен съд</a>)</p><p>This later development does not make the December dispute irrelevant.</p><p>On the contrary, it reinforces the central analytical point. The process was not governed by one institution making one final historical judgement. Municipal bodies, regional administration, cultural-heritage authorities, political organisations and courts operated through partially overlapping powers.</p><p>The state, in this sense, was not a unitary memory actor.</p><p>Institutional reclassification therefore required coordination across systems that do not exist solely to implement memory policy.</p><h3>5.4 Monument removal as a material operation</h3><p>Memory studies can sometimes discuss demolition and relocation at a level of abstraction that underestimates materiality.</p><p>A large monument cannot simply be moved because a political community has decided its meaning has changed.</p><p>Bronze has weight. Stone fractures. Structures deteriorate. Sculptures need lifting plans, temporary supports and safe storage. Cultural objects need conservation strategies. Public land and state property may be governed by different institutions. Budgets require allocation. Contractors require authority. The future installation site requires space.</p><p>The December 2023 controversy makes this material layer impossible to ignore.</p><p>The monument’s political meaning had become unstable years earlier. Its physical removal required engineering and administration.</p><p>This temporal gap is one reason why symbolic delegitimation and material disappearance should not be treated as the same process.</p><p>A monument can lose authority before it loses its pedestal.</p><h3>5.5 The empty monument and the politics of absence</h3><p>Once the principal figures were removed, the site entered another interpretive phase.</p><p>An empty or partially dismantled monumental structure does not become meaningless. Absence becomes visible.</p><p>The comparison with the Georgi Dimitrov Mausoleum is again instructive. Decheva notes that after the mausoleum’s destruction the former location continued for years to be described through reference to what was no longer there. (<a href="https://aph-ihpan.edu.pl/images/APH126/05_Decheva.pdf">aph-ihpan.edu.pl</a>)</p><p>A similar dynamic is possible with the Monument to the Soviet Army. The absence of the sculptural group can itself act as a marker of political transformation.</p><p>This means that removal cannot fulfil the literal objective of deleting the monument from historical consciousness.</p><p>Indeed, dismantling may intensify remembrance by creating a dramatic before-and-after event.</p><p>Photography, media archives, political anniversaries and the continuing debate about the future of the objects ensure that the removed monument remains socially available.</p><p>The politics of monuments therefore persists after the monument changes form.</p><h3>Discussion</h3><h3>From Ideological Monument to Contested Surface to Reclassified Object</h3><p>The case examined in this thesis can now be organised into three analytical regimes.</p><p>The first is the <strong>authorised ideological monument</strong>.</p><p>The original socialist monument combined state authority, central location, monumental scale, inscription and iconography into a coordinated political statement. The object was designed to make a particular interpretation of 1944 durable and visible.</p><p>The second is the <strong>contested public surface</strong>.</p><p>After socialism the state lost the capacity or willingness to guarantee a single interpretation. The 2011 intervention made this instability dramatically visible, and subsequent actions recursively transformed the relief into a medium for political communication. Everyday appropriation simultaneously pluralised the social uses of the surrounding space.</p><p>The third is the <strong>object of institutional reclassification</strong>.</p><p>Decision №168 attempted to end the monument’s status as dominant central commemoration by separating sculptural object from urban site and combining relocation with a new historical explanation.</p><p>These regimes overlap rather than replace one another cleanly.</p><p>The original ideological meaning remains legible precisely because later actors continue to argue against or defend it. The contested-surface regime survives in archives and images even after parts of the monument are dismantled. Institutional reclassification remains incomplete as long as the future location and final interpretive arrangement of the sculptural material remain unsettled.</p><p>The strongest conclusion is therefore not that one meaning defeated another.</p><p>It is that <strong>the institutional location of interpretive authority changed</strong>.</p><p>Under socialism, authority was concentrated.</p><p>During the post-socialist period, authority fragmented.</p><p>In 2023 municipal institutions attempted to reconcentrate it.</p><p>The administrative disputes demonstrated that reconcentration could not occur solely through symbolic legitimacy.</p><h3>Reconsidering decommunisation</h3><p>The concept of decommunisation captures an important dimension of the case. Removing an explicitly Soviet monument from a central location can function as a rejection of the political order that produced it.</p><p>Yet the term can imply a linear sequence:</p><p>communist symbolism → democratic rejection → removal → completed transition.</p><p>The Sofia case resists this structure.</p><p>First, the monument was already being politically neutralised, mocked and repurposed long before its physical dismantling.</p><p>Second, some forms of re-signification did not simply oppose communism. They addressed commercial culture, contemporary Russia, Ukraine, Czech memory and local urban freedoms.</p><p>Third, the 2023 decision did not propose pure destruction. It proposed preservation of sculptural material under altered spatial and interpretive conditions.</p><p>Fourth, the legal and administrative disputes demonstrate that democratic governance can slow or complicate symbolic action precisely because public authority is constrained by procedure.</p><p>What may appear politically as “failure to finish the transition” may therefore sometimes be partly produced by features of post-authoritarian legality itself: plural jurisdiction, property rules, rights of appeal and institutional competence.</p><p>This does not make all delay normatively desirable. It means delay requires explanation.</p><h3>Reconsidering vandalism</h3><p>A second conclusion concerns terminology.</p><p>Many interventions on the monument were illegal alterations and may properly be described in legal or conservation terms as vandalism.</p><p>But “vandalism” alone is analytically insufficient.</p><p>It classifies an action by its relationship to property or conservation rules but does not explain its communicative mechanism.</p><p>Painting Soviet soldiers as superheroes, colouring the relief pink in reference to Prague 1968, or using Ukrainian colours after Russian military intervention were not interchangeable acts of random damage.</p><p>They relied on the monument’s existing semantic structure.</p><p>The analytically useful question is therefore not whether an intervention was vandalism <strong>or</strong> political speech. It can be legally the first and communicatively the second.</p><h3>Reconsidering preservation</h3><p>Preservation is equally complex.</p><p>To preserve a monument materially does not necessarily mean preserving its original ideology.</p><p>Dimitrova’s account of everyday appropriation shows that a site can remain physically intact while acquiring uses largely unrelated to its founding commemorative programme. Decheva’s discussion similarly identifies desacralisation and redefinition as alternatives to physical destruction. (<a href="https://www.ceeol.com/search/article-detail?id=423986&amp;utm_source=chatgpt.com">CEEOL</a>)</p><p>The opposite is also true.</p><p>Relocation can preserve physical objects while deliberately ending their original spatial function.</p><p>The 2023 proposal thus complicates the preservation/removal binary. It represents <strong>material preservation combined with commemorative displacement</strong>.</p><h3>Space as part of meaning</h3><p>The case ultimately demonstrates that monuments cannot be analysed separately from location.</p><p>The same bronze figures in Knyazheska Garden, dominating a central urban axis, do not communicate in the same way when presented as objects inside a museum.</p><p>Location is not a container for meaning. It is part of meaning.</p><p>The central city grants unavoidable visibility and symbolic competition with surrounding institutions and urban routes.</p><p>A museum places the object inside a curatorial system.</p><p>A storage facility suspends public communication almost completely.</p><p>A peripheral memorial park would generate another relationship.</p><p>The 2023 conflict was therefore as much about <strong>where history speaks</strong> as about <strong>what history says</strong>.</p><h3>Conclusion</h3><p>This thesis asked how public re-signification and institutional action transformed the political status of Sofia’s Monument to the Soviet Army between 2011 and 2023.</p><p>The answer is that the monument underwent neither simple ideological continuity nor straightforward symbolic erasure.</p><p>It moved through a prolonged transformation in the relationship among material form, public interpretation and institutional authority.</p><p>The first important transformation was visible by 2011. The monument’s reliefs became repeatedly writable political surfaces. “In Pace with the Times” demonstrated that the original heroic iconography could be retained materially while radically altered semiotically. Subsequent interventions connected the site to Pussy Riot, the Prague Spring, Ukraine, Katyn and contemporary Russian military power. These interventions accumulated rather than forming one coherent counter-narrative.</p><p>The second transformation concerned the social life of the site. Research on urban appropriation shows that the surrounding space supported everyday activities and civic uses that weakened the assumption that the monument’s ideological programme determined all interactions with the site. The monument remained politically available, but it also became urban infrastructure.</p><p>The third transformation followed the new geopolitical environment after February 2022. Russia’s invasion of Ukraine made the monument’s Soviet military iconography newly relevant to present political conflict. Older campaigns for removal could now operate in a substantially altered symbolic environment.</p><p>Sofia Municipal Council Decision №168 of March 2023 marked the decisive institutional shift.</p><p>The decision did more than demand that a controversial object be taken away. It proposed a different regime of heritage. Sculptural elements were to be removed from their central commemorative location and exhibited in a museum or another suitable state space, while the original site would receive a newly authorised historical interpretation.</p><p>The process may therefore be described as an attempted movement from <strong>monumental commemoration to contextualised heritage</strong>.</p><p>December 2023 demonstrated the limits of interpreting this shift as simple political completion.</p><p>Physical dismantling triggered questions of law, ownership, conservation, storage, finance and institutional competence. Court proceedings revealed the importance of administrative documentation and standing. Subsequent higher-court decisions complicated the legal meaning of the initial orders.</p><p>The monument’s history consequently exposes three distinct processes that are too often collapsed into one.</p><p>A society can delegitimise a monument’s message.</p><p>It can establish political support for relocation.</p><p>And it can possess the lawful, technical and administrative capacity to carry out relocation.</p><p>These processes need not occur simultaneously.</p><p>The central hypothesis of this thesis can therefore be retained in modified and more precise form.</p><p>Between 2011 and 2023, the Monument to the Soviet Army did not simply undergo progressive decommunisation. It underwent a redistribution of interpretive authority. Repeated public interventions deprived the original commemorative programme of its exclusive control over the site’s meaning. Everyday appropriation pluralised the site’s social functions. The municipal decision of 2023 attempted to replace this unresolved plurality with an institutional framework based on relocation and explicit historical contextualisation. The legal and administrative disputes surrounding dismantling demonstrated that even a politically delegitimised monument remained embedded in material and institutional structures that limited the speed and form of transformation.</p><p>The case also supports a broader conclusion about post-socialist memory.</p><p>Political transition does not conclude when a statue comes down.</p><p>A removed monument can continue to organise political discourse through photographs, stored fragments, empty pedestals, museum displays, court records and arguments about what should occupy the space that remains.</p><p>Removal is therefore not the opposite of memory.</p><p>It is one of the ways memory is reorganised.</p><p>For Sofia’s Monument to the Soviet Army, December 2023 should consequently be understood not as the end of the monument’s political life but as the beginning of another phase: the struggle over how a formerly dominant public object will be preserved, contextualised, displayed or remembered after its monumental authority has been materially interrupted.</p><h3>Notes</h3><h3>Bibliography</h3><p>Administrative Court Sofia-City. “АССГ се произнесе по искания за спирането на демонтажа на ‘Паметник на Съветската армия.’” December 19, 2023.</p><p>Debruyne, Nina, and Georgeta Nazarska. “Contentious Heritage Spaces in Post-Communist Bulgaria: Contesting Two Monuments in Sofia.” <em>Journal of Historical Geography</em> 83 (2024): 176–190. <a href="https://doi.org/10.1016/j.jhg.2023.09.003">https://doi.org/10.1016/j.jhg.2023.09.003</a>.</p><p>Decheva, Daniela. “Inconveniences of Memory: The Monument to the Soviet Army and Georgi Dimitrov’s Mausoleum in Sofia after 1989.” <em>Acta Poloniae Historica</em> 126 (2022): 89–100. <a href="https://doi.org/10.12775/APH.2022.126.05">https://doi.org/10.12775/APH.2022.126.05</a>.</p><p>Dimitrova, Kristina. “Appropriations of Urban Space as Resistance: The Soviet Army Monument in Sofia.” <em>Critique &amp; Humanism</em> 46, no. 2 (2016).</p><p>Genova, Neda. “Temporal Politics of the Surface: Keeping Pace with the Monument to the Soviet Army in Sofia.” <em>New Formations</em> 106 (2022): 25–43. <a href="https://doi.org/10.3898/NEWF:106.02.2022">https://doi.org/10.3898/NEWF:106.02.2022</a>.</p><p>Ivanova, Mina. “The Bulgarian Monument to the Soviet Army: Visual Burlesque, Epic, and the Emergence of Comic Subjectivity.” <em>Quarterly Journal of Speech</em> 100, no. 3 (2014): 273–302. <a href="https://doi.org/10.1080/00335630.2014.982159">https://doi.org/10.1080/00335630.2014.982159</a>.</p><p>Koleva, Roberta. “‘The Bulgarian Berlin Wall Is Going Away Step by Step’: ‘Standby Transition’ and the Quest for ‘Europeanness’ and ‘Democracy’ in the Discourses Around the Dismantling of the Monument to the Soviet Army in Sofia.” <em>Anthropology Journal</em> 11, no. 2 (2024).</p><p>Macdonald, Sharon. <em>Memorylands: Heritage and Identity in Europe Today</em>. Abingdon: Routledge, 2013.</p><p>Nora, Pierre. “Between Memory and History: Les Lieux de Mémoire.” <em>Representations</em> 26 (1989): 7–24. <a href="https://doi.org/10.2307/2928520">https://doi.org/10.2307/2928520</a>.</p><p>Smith, Laurajane. <em>Uses of Heritage</em>. Abingdon: Routledge, 2006.</p><p>Sofia Municipal Council. Decision №168. March 9, 2023.</p><p>Supreme Administrative Court of the Republic of Bulgaria. Decisions concerning Administrative Court Sofia-City proceedings on the dismantling of the Monument to the Soviet Army. April 2024.</p><p>Tunbridge, J. E., and G. J. Ashworth. <em>Dissonant Heritage: The Management of the Past as a Resource in Conflict</em>. Chichester: Wiley, 1996.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c9b673fe3f98" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Bosna Tarihi]]></title>
            <link>https://medium.com/@batuhanayrba/bosna-tarihi-8dffd009c85f?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/8dffd009c85f</guid>
            <category><![CDATA[bosna]]></category>
            <category><![CDATA[tarih]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Tue, 28 Jul 2026 07:09:08 GMT</pubDate>
            <atom:updated>2026-09-03T16:42:28.545Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*71D8T_CLimRTLrRRJty56A.jpeg" /></figure><p>Bosna tarihi, yalnızca Balkanlar’ın yerel geçmişi olarak okunamaz. Bu tarih, Roma mirası, Orta Çağ Bosna devleti, Osmanlı idaresi, Avusturya-Macaristan yönetimi, Yugoslavya deneyimi, 1992-1995 savaşı ve Dayton sonrası siyasal düzenin birbirine eklemlendiği uzun bir süreklilik taşır. Bosna-Hersek’in bugünkü toplumsal yapısı da bu tarihsel katmanların sonucudur. Bu nedenle Bosna’yı anlamak, etnik ya da dini kimlikleri donmuş kategoriler gibi ele almak yerine, bu kimliklerin farklı imparatorluklar, sınırlar, hukuk düzenleri ve savaş deneyimleri içinde nasıl biçimlendiğini incelemeyi gerektirir.</p><h3><strong>Orta Çağ Bosnası ve Yerel Siyasi Kimliğin Doğuşu</strong></h3><p>Bosna, Orta Çağ’da Bizans, Macar ve yerel Balkan güçleri arasında kalan bir sınır bölgesi olarak önem kazandı. Noel Malcolm’un Bosnia: A Short History adlı çalışması, Bosna tarihinin yalnızca dış egemenlikler üzerinden değil, yerel siyasi oluşumlar ve dini çoğulluk üzerinden de okunması gerektiğini gösterir. Eserin bölüm yapısı, Bosna’nın 1180-1463 arasındaki Orta Çağ devleti, Bosna Kilisesi, Osmanlı fethi ve İslamlaşma süreçlerini ayrı tarihsel sorunlar olarak ele alır. Bu yaklaşım, Bosna’nın tarihini tek bir ulusal anlatıya indirgemeyi zorlaştırır. Çünkü bölgenin Orta Çağ geçmişi, hem Katolik hem Ortodoks dünyayla temas eden, aynı zamanda Bosna Kilisesi gibi yerel ve tartışmalı bir dini yapının varlığıyla ayrışan karmaşık bir zemine sahiptir.</p><p>Bu dönemde Bosna’nın kimliği, bugünkü anlamıyla modern ulus kavramına dayanmaz. Daha çok hanedan, bölgesel sadakat, dini aidiyet ve dış baskılara karşı yerel özerklik arayışı üzerinden şekillenir. Bu nedenle Orta Çağ Bosnası, sonraki yüzyıllarda geliştirilen milliyetçi tarih yorumlarına doğrudan malzeme edilse de, kendi döneminin siyasi ve toplumsal mantığı içinde değerlendirilmelidir.</p><h3><strong>Osmanlı Dönemi: İdare, Şehirleşme ve Dini Dönüşüm</strong></h3><p>Osmanlıların Bosna’yı 15. yüzyılda fethetmesi, bölgenin siyasi tarihinde köklü bir kırılma yarattı. Ancak bu kırılma yalnızca askeri egemenlik değişimi değildi. Osmanlı dönemi, Bosna’da şehirleşmenin, vakıf kurumlarının, İslam kültürünün ve yeni idari düzenin güçlendiği uzun bir dönem oldu. Sarajevo ve Mostar gibi şehirlerin gelişimi bu dönüşümün en görünür sonuçları arasındadır.</p><p>Bosna’da İslamlaşma konusu, tarih yazımında sıkça tartışılmıştır. Daha dengeli çalışmalar, bu süreci basitçe “zorla din değiştirme” anlatısıyla açıklamanın yetersiz olduğunu gösterir. Dönüşüm; vergi düzeni, toplumsal yükselme imkanları, yerel dini yapıların zayıflığı, Osmanlı idaresinin sağladığı yeni statüler ve bölgenin sınır karakteri gibi birçok unsurla ilişkilidir. Bu nedenle Bosna’daki Müslüman kimliğin ortaya çıkışı, dışarıdan getirilen bir nüfusun sonucu olmaktan çok, yerel Slav nüfusun uzun süreli toplumsal dönüşümü olarak görülmelidir.</p><p>Osmanlı dönemi, Bosna’yı imparatorluğun batı sınırlarından biri haline getirdi. Bu sınır konumu, bölgeyi hem askeri açıdan önemli kıldı hem de Bosna toplumunu Habsburg, Venedik ve Osmanlı rekabetinin etkilerine açık hale getirdi. Böylece Bosna, yalnızca Osmanlı merkezinin bir taşrası değil, Avrupa ile Osmanlı dünyası arasındaki temas alanlarından biri oldu.</p><h3>Avusturya-Macaristan Yönetimi ve Modernleşme</h3><p>1878 Berlin Kongresi’nden sonra Bosna-Hersek fiilen Avusturya-Macaristan yönetimine geçti. Bu dönem, Osmanlı idaresinden Habsburg bürokrasisine geçiş anlamına geliyordu. Bosna toplumu için bu geçiş, yalnızca yönetici değişikliği değil, hukuk, eğitim, altyapı, şehir planlaması ve bürokratik denetim alanlarında yeni bir modernleşme biçimiydi. Bosna-Hersek Parlamentosu’nun tarihsel özetinde de 13 Temmuz 1878 Berlin Kongresi sonrasında bölgenin geçici Avusturya-Macaristan işgaline bırakıldığı belirtilir.</p><p>1908’deki resmî ilhak ise Avrupa siyasetinde daha geniş sonuçlar doğurdu. Bosna Krizi, özellikle Sırp milliyetçiliği ve Güney Slav siyasetleri açısından gerginliği artırdı. 1914’te Saraybosna’da Avusturya-Macaristan veliahdı Franz Ferdinand’ın öldürülmesi, Birinci Dünya Savaşı’nın tetikleyici olaylarından biri oldu. Bu bakımdan Bosna, 20. yüzyıl Avrupa tarihinin başlangıç kırılmalarından birinin merkezinde yer aldı.</p><h3>Yugoslavya Deneyimi ve Kimliklerin Siyasi Çerçevesi</h3><p>Birinci Dünya Savaşı sonrasında Bosna, Güney Slav devlet oluşumları içinde yer aldı. İkinci Dünya Savaşı’nın ardından sosyalist Yugoslavya’nın cumhuriyetlerinden biri oldu. Josip Broz Tito döneminde Yugoslavya, etnik farklılıkları federal bir yapı içinde dengelemeye çalıştı. Ancak bu denge, özellikle 1980’de Tito’nun ölümünden sonra zayıfladı. Ekonomik kriz, cumhuriyetler arası rekabet ve milliyetçi siyasetlerin yükselişi, Yugoslavya’nın çözülmesini hızlandırdı.</p><p>Bosna-Hersek’in durumu diğer cumhuriyetlerden daha hassastı. Çünkü Bosna’da Boşnak, Sırp ve Hırvat nüfuslar iç içe geçmişti. Congressional Research Service raporu, Bosna’daki yerleşim düzeninin “bölünmeyi” zorlaştıran dağınık bir yapı taşıdığına dikkat çeker. Bu iç içelik, savaş öncesi siyasal krizin neden bu kadar yıkıcı hale geldiğini açıklayan temel unsurlardan biridir.</p><h3>1992-1995 Savaşı, Srebrenitsa ve Dayton Düzeni</h3><p>Bosna-Hersek 1992’de bağımsızlık sürecine girdiğinde, Yugoslavya’nın parçalanması artık savaşla iç içe geçmişti. 1992-1995 Bosna Savaşı, Avrupa’nın İkinci Dünya Savaşı sonrasındaki en ağır insani felaketlerinden biri oldu. Sarajevo kuşatması, etnik temizlik politikaları, zorla yerinden edilmeler ve toplu katliamlar savaşın hafızasında merkezi yer tuttu.</p><p>Srebrenitsa, bu hafızanın en ağır noktasıdır. Eski Yugoslavya Uluslararası Ceza Mahkemesi, Temmuz 1995’te Srebrenitsa’da Bosnalı Müslüman erkek ve erkek çocukların kitlesel infazının soykırım oluşturduğunu tespit etti. ICTY’nin aktardığına göre Srebrenitsa 1993’te BM koruması altında “güvenli bölge” ilan edilmişti, fakat Temmuz 1995’te Bosnalı Sırp güçler bölgeyi ele geçirdi ve 7.000’den fazla erkek ve çocuğun öldürülmesiyle sonuçlanan planlı şiddet yaşandı.</p><p>Savaş, 1995 Dayton Barış Antlaşması ile sona erdi. Dayton düzeni Bosna-Hersek’in uluslararası sınırlarını korudu, fakat devleti iki ana entiteye ayırdı: Bosna-Hersek Federasyonu ve Republika Srpska. Office of the High Representative tarafından yayımlanan Anayasa metni, Bosna-Hersek’in uluslararası hukuk bakımından devlet olarak devam ettiğini, fakat iç yapısının değiştirildiğini belirtir. Aynı metin, ülkenin iki entiteden oluştuğunu ve başkentinin Sarajevo olduğunu kaydeder.</p><p>Dayton, savaşı durdurdu. Fakat kurduğu sistem, barışı yönetilebilir kılarken siyaseti etnik veto mekanizmalarına bağımlı hale getirdi. BTI 2026 ülke raporu da Bosna-Hersek’in kurumlarının hâlâ 1990’lar savaşının mirasıyla şekillendiğini, Dayton sonrası yapının zayıf merkezi kurumlar ve güçlü alt birimler ürettiğini vurgular.</p><h3>Sonuç</h3><p>Bosna tarihi, tek çizgili bir “kurbanlık”, “çatışma” ya da “çok kültürlülük” anlatısına sığmaz. Bölge, yüzyıllar boyunca imparatorluk sınırı, dini geçiş alanı, yerel devlet geleneği, modernleşme laboratuvarı, Yugoslav federalizminin parçası ve savaş sonrası karmaşık bir anayasal düzen olarak var oldu. Bugünkü Bosna-Hersek’i anlamak için bu katmanların hepsini birlikte düşünmek gerekir.</p><p>Bosna’nın tarihsel önemi de buradadır. Ülke, Balkanlar’ın yalnızca acı dolu bir hafızası değil, Avrupa siyasetinin kimlik, egemenlik, çoğulluk ve adalet meselelerini açık biçimde gösteren bir örneğidir. Dayton sonrası düzen hâlâ tartışmalıdır, fakat Bosna’nın geleceği yalnızca anayasal mühendislik meselesi değildir. Aynı zamanda tarihsel hafızanın nasıl işlendiği, savaş suçlarıyla nasıl yüzleşildiği ve çok kimlikli bir toplumun ortak siyasal hayat kurup kuramayacağı sorusuna bağlıdır.</p><h3>Kaynakça</h3><p>Malcolm, Noel. Bosnia: A Short History. Internet Archive kaydı: <a href="https://archive.org/details/bosniashorthisto0000malc_i7r9">https://archive.org/details/bosniashorthisto0000malc_i7r9</a></p><p>Office of the High Representative. “Annex 4: Constitution of Bosnia and Herzegovina.” <a href="https://www.ohr.int/dayton-peace-agreement/annex-4/">https://www.ohr.int/dayton-peace-agreement/annex-4/</a></p><p>International Criminal Tribunal for the former Yugoslavia. “Bridging the Gap in Srebrenica.” <a href="https://www.icty.org/en/outreach/bridging-the-gap-with-local-communities/srebrenica">https://www.icty.org/en/outreach/bridging-the-gap-with-local-communities/srebrenica</a></p><p>United States Holocaust Memorial Museum. “Prosecuting the Crime of Genocide.” <a href="https://www.ushmm.org/genocide-prevention/countries/bosnia-herzegovina/prosecuting-the-crime-of-genocide">https://www.ushmm.org/genocide-prevention/countries/bosnia-herzegovina/prosecuting-the-crime-of-genocide</a></p><p>Congressional Research Service. “Bosnia and Herzegovina: Background on U.S. Policy Concerns.” <a href="https://www.everycrsreport.com/reports/RL30906.html">https://www.everycrsreport.com/reports/RL30906.html</a></p><p>Bertelsmann Stiftung. “BTI 2026 Bosnia and Herzegovina Country Report.” <a href="https://bti-project.org/en/reports/country-report/BIH">https://bti-project.org/en/reports/country-report/BIH</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8dffd009c85f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[İkinci Bulgar İmparatorluğu (2. Bulgar Hanlığı)]]></title>
            <link>https://medium.com/@batuhanayrba/i%CC%87kinci-bulgar-i%CC%87mparatorlu%C4%9Fu-2-bulgar-hanl%C4%B1%C4%9F%C4%B1-53890f18abcc?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/53890f18abcc</guid>
            <category><![CDATA[bulgaristan]]></category>
            <category><![CDATA[tarih]]></category>
            <category><![CDATA[han]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Fri, 03 Jul 2026 21:11:22 GMT</pubDate>
            <atom:updated>2026-07-03T21:11:22.204Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/563/1*7xWNY8dsN52XnituT5HZXA.png" /></figure><p>İkinci Bulgar İmparatorluğu (İkinci Bulgar Devleti veya 2. Bulgar Hanlığı olarak da bilinir), 1185 yılında Tırnovo’da başlayan bir ayaklanma sonucu kurulmuş ve 1396 yılında Osmanlı İmparatorluğu tarafından fethedilene kadar varlığını sürdürmüş bir Orta Çağ Bulgar devletidir. Birinci Bulgar İmparatorluğu’nun 1018&#39;de Bizans tarafından fethedilmesinin ardından geçen 167 yıllık Bizans egemenliği döneminden sonra, Bulgarlar bağımsızlıklarını yeniden kazanmışlardır <strong>[1]</strong>.</p><h3>Kuruluş ve Asen Hanedanı</h3><p>İkinci Bulgar İmparatorluğu’nun kuruluşu, Asen (Asenevtsi) ailesine mensup kardeşler Teodor (daha sonra IV. Petăr olarak anılacaktır) ve İvan Asen-Belgun önderliğinde gerçekleşmiştir. 1185 sonbaharında Tırnovo’da patlak veren ayaklanma, Bizans İmparatorluğu’na karşı başlatılmıştır. Ayaklanmanın ideolojik temelini oluşturan en önemli unsur, Selanik’ten Tırnovo’ya getirilen Aziz Dimitrios’un mucizevi ikonası olmuştur. Aziz Dimitrios, Asen ailesinin ve Bulgar halkının koruyucu azizi ilan edilmiş, böylece ayaklanmanın Tanrı’nın takdiriyle gerçekleştiği fikri halka aşılanmıştır <strong>[3]</strong>.</p><p>Kardeşlerden büyüğü olan Teodor, taç giyme töreninden önce adını değiştirerek Petăr adını almıştır. Bu isim, Birinci Bulgar İmparatorluğu döneminde hüküm sürmüş ve aziz ilan edilmiş olan Çar I. Petăr’ın (927–969) anısına seçilmiştir. Bu şekilde Asen ailesi, kendilerini Birinci Bulgar İmparatorluğu’nun meşru mirasçıları olarak konumlandırmış ve halk nezdinde meşruiyet kazanmıştır. Ayaklanmanın başarıya ulaşmasında Kuman savaşçılarının desteği de kritik rol oynamıştır <strong>[4]</strong>.</p><h3>Siyasi Gelişmeler ve Genişleme</h3><p>İkinci Bulgar İmparatorluğu’nun başkenti, ayaklanmanın merkezi olan Tırnovo olmuştur. Yantra Nehri’nin derin vadisi içinde, üç tarafı sarp kayalıklarla çevrili Tsarevets ve Trapezitsa tepeleri üzerine kurulu olan Tırnovo, doğal savunma avantajları sayesinde stratejik bir konuma sahipti. Şehir, kısa sürede sadece siyasi değil, aynı zamanda dini ve kültürel bir merkez haline gelmiştir <strong>[7]</strong>.</p><p>İmparatorluğun genişlemesi, Çar Kaloan (1197–1207) döneminde hız kazanmıştır. Kaloan, Bizans, Macaristan ve 1204&#39;te kurulan Konstantinopolis Latin İmparatorluğu’na karşı önemli zaferler kazanmıştır. 1205 yılında Edirne Muharebesi’nde Latin İmparatoru I. Baudouin’i esir alarak büyük bir başarı elde etmiştir. Kaloan, aynı zamanda Roma Kilisesi ile birlik (unia) anlaşması imzalayarak (1204) uluslararası tanınırlık kazanmış ve papalıktan kraliyet tacı almıştır <strong>[1,3]</strong>.</p><h3>İvan Asen II Dönemi ve Zirve Dönemi</h3><p>İkinci Bulgar İmparatorluğu en parlak dönemini Çar II. İvan Asen (1218–1241) döneminde yaşamıştır. 1230 yılında Klokotnitsa Muharebesi’nde Epir Despotluğu’nu mağlup eden İvan Asen, Bulgaristan’ı Balkanlar’ın en güçlü devleti haline getirmiştir. Sınırları Karadeniz’den Adriyatik Denizi’ne, Ege Denizi’nden Tuna Nehri’ne kadar uzanmıştır <strong>[1]</strong>.</p><p>Bu dönemin en önemli başarılarından biri, 1235 yılında Bulgar Patrikhanesi’nin yeniden kurulması olmuştur. İznik İmparatorluğu’nda toplanan bir konsilde, tüm Doğu patriklerinin onayıyla Bulgar patrikhanesi restore edilmiş, böylece Bulgar devletinin tam bağımsızlığı ve egemenliği pekiştirilmiştir. II. İvan Asen’in unvanı “İsa Mesih’te Tanrı’ya sadık, tüm Bulgarların ve Yunanların çarı ve otokratı” şeklinde değişmiş, bu da Bizans mirasına yönelik iddiaları yansıtmıştır <strong>[1,3]</strong>.</p><h3>İdari Yapı ve Bizans Etkisi</h3><p>İkinci Bulgar İmparatorluğu’nun idari sistemi, büyük ölçüde Bizans modelini takip etmiştir. Yapılan araştırmalar, devlet görevlilerinin unvanlarının yaklaşık %74&#39;ünün Grekçe kökenli olduğunu göstermektedir. Bunların %57,4&#39;ü doğrudan Grekçeden çeviri yapılmadan (transliterasyon) alınırken, %16,6&#39;sı birebir çeviri (kalk) yoluyla Bulgarcaya uyarlanmıştır. Slavca kökenli terimler %20,3 oranında kalırken, Türkçe kökenli terimler ise yalnızca %5,5 oranındadır <strong>[2]</strong>.</p><p>Bu terminolojik durum, İkinci Bulgar İmparatorluğu’nun Birinci Bulgar İmparatorluğu’ndan farklı olarak, Avrasya bozkır geleneklerini terk ettiğini ve tam anlamıyla Bizans tipi bir devlet yapısına büründüğünü göstermektedir. Saray unvanları arasında despot, sevastokrator ve kaisar/kesar gibi Bizans kökenli yüksek soyluluk unvanları bulunmaktaydı. Merkezi yönetimde protovestiyarios (hazine sorumlusu), megas logothetes (başbakan/şansölye), protostrator (ordu komutanı) gibi görevliler yer almıştır <strong>[2]</strong>.</p><p>Ülke, “hora” adı verilen büyük askeri-idari bölgelere ayrılmış, bu bölgeler “kefalya” adı verilen valiler tarafından yönetilmiştir. Yerel yönetimde “duka”, “katepano”, “kastrophylaks” (kale muhafızı) gibi Bizans kökenli unvanlar kullanılmıştır. Vergi toplama sistemi de Bizans modeline uygun şekilde örgütlenmişti <strong>[2]</strong>.</p><h3>Kuman ve Tatar Etkisi</h3><p>Kumanlar, İkinci Bulgar İmparatorluğu’nun kuruluşunda ve gelişmesinde önemli bir rol oynamıştır. Asen hanedanı, Terter hanedanı ve Şişman hanedanı gibi Bulgar hükümdar ailelerinin Kuman kökenli olduğu belirtilmektedir. Ayrıca Eflak’ın Basarabid hanedanı da Kuman kökenlidir <strong>[4]</strong>.</p><p>Tatarlar ise 1242&#39;den itibaren Balkanlar’da etkili olmaya başlamıştır. Altın Orda’nın güçlü emiri Nogay, 13. yüzyıl sonlarında Don Nehri’nden Tuna’ya ve kuzey Bulgaristan’a kadar kontrolünü genişletmiştir. Bulgaristan, 1242&#39;den itibaren Tatarlara haraç ödemek zorunda kalmış, bu durum 13. yüzyıl sonunda vasallık ilişkisine dönüşmüştür <strong>[1,4]</strong>.</p><h3>Kültürel Gelişmeler</h3><p>İkinci Bulgar İmparatorluğu döneminde Tırnovo, “Yeni Konstantinopolis” veya “Üçüncü Roma” olarak anılmaya başlanmıştır. Bu fikir, özellikle Kaloan döneminde ortaya çıkmış ve Pandeh adlı bir yazarın kehanet metninde ifade edilmiştir. Pandeh, “Bulgar gençtir” diyerek Bulgar İmparatorluğu’nun Bizans’ın mirasçısı olduğunu ileri sürmüştür <strong>[3]</strong>.</p><ol><li>yüzyılda hüküm süren Çar İvan Aleksandr (1331–1371), Bizans kültürel mirasını sahiplenme konusunda en iddialı hükümdarlardan biri olmuştur. Londra’da bulunan dört İncil el yazması (Tetraevangelion, 1356), İvan Aleksandr’ın himayesinde üretilmiş en görkemli eserlerdendir. El yazmasında çar, Bizans imparatoru gibi giyinmiş, kırmızı çizmeleri üzerinde Palaiologos hanedanının sembolü olan çift başlı kartal bulunmaktadır. Unvanı “tüm Bulgarların ve Yunanların çarı ve otokratı” olarak kaydedilmiştir <strong>[5]</strong>.</li></ol><p>Boyana Kilisesi’ndeki (Sofya yakınları) 1259 tarihli freskler, bu dönemin sanatsal başarısının önemli bir örneğidir. Sebastokrator Kaloian tarafından yaptırılan kilisedeki resimler, Bizans ikonografik geleneğine bağlı kalmakla birlikte yerel Bulgar unsurlarını da barındırmaktadır. Dönemin Bulgar Çarı Konstantin Tih (1257–1277) ve eşi İrini’nin portreleri de burada yer almaktadır <strong>[6]</strong>.</p><h3>Gerileme ve Osmanlı Fethi</h3><ol><li>yüzyılın ortalarından itibaren İkinci Bulgar İmparatorluğu, feodal parçalanma sürecine girmiştir. Çar İvan Aleksandr’ın ölümünden sonra ülke dört temel feodal bölgeye ayrılmıştır: Tırnovo Krallığı, Vidin Krallığı, Velbujd (Köstendil) Despotluğu ve Dobruca Despotluğu. Bu parçalanma, Osmanlı ilerleyişi karşısında birleşik bir direniş gösterilememesine yol açmıştır <strong>[1]</strong>.</li></ol><p>Osmanlı Türkleri, 1350&#39;lerde Balkanlar’da toprak kazanmaya başlamış, 1371&#39;deki Çirmen Muharebesi’nden sonra Bulgar hükümdarlarının çoğu Osmanlı vasallığını kabul etmiştir. 1393&#39;te Tırnovo’nun, 1396&#39;da ise Vidin’in düşmesiyle İkinci Bulgar İmparatorluğu tamamen ortadan kalkmıştır. Vidin Kralı İvan Sratsimir’in 1396&#39;da Niğbolu Muharebesi’nde Haçlılar safında savaşması, Vidin’in fethiyle sonuçlanmıştır <strong>[1]</strong>.</p><h3>Sonuç</h3><p>İkinci Bulgar İmparatorluğu, yaklaşık iki yüzyıl boyunca Balkanlar’ın en önemli siyasi güçlerinden biri olmuştur. Bizans medeniyetinin bir parçası olarak, Slav-Bizans kültür geleneğinin sürdürülmesinde ve geliştirilmesinde önemli rol oynamıştır. Devletin idari yapısı, unvan sistemi ve kurumları Bizans modeline dayanmakla birlikte, Bulgar halkının kimliğinin korunması ve Slav edebiyatının gelişmesi açısından da kritik bir dönem olmuştur <strong>[1,2]</strong>.</p><h3>Kaynakça</h3><p><strong>[1]</strong>Gjuzelev, V. (1981). The Bulgarian Medieval State: Seventh To Fourteenth Centuries. Southeastern Europe, 8(1), 45–85<br>DOI: 10.1163/187633381x00037<strong>[2]</strong>Biliarsky, I. (2001). Some observations on the administrative terminology of the second Bulgarian empire (13th-14th centuries). Byzantine and Modern Greek Studies, 25(1), 69–89<br>DOI: 10.1179/byz.2001.25.1.69<strong>[3]</strong>Kaymakamova, M. (2010). Power and history in Bulgaria at the end of the 12th and during the 13th century. Zbornik radova Vizantološkog instituta, (47), 215–245<br>DOI: 10.2298/zrvi1047215k<strong>[4]</strong>Vásáry, I. (2005). Cumans and Tatars: Oriental Military in the Pre-Ottoman Balkans, 1185–1365. History: Reviews of New Books, 34(1), 3–3<br>DOI: 10.1080/03612750509602053<strong>[5]</strong>Schroeder, R. (2007). Displacing Byzantium, Disgracing Convention: The Manuscript Patronage of Tsar Ivan Alexander of Bulgaria. Manuscripta, 51(2), 181–208<br>DOI: 10.1484/j.mss.1.100104<strong>[6]</strong>Schroeder, R. (2016). Disappearing Byzantine Heritage: The Case of the Medieval Church at Boiana. Journal of Modern Greek Studies, 34(2), 249–278<br>DOI: 10.1353/mgs.2016.0006<strong>[7]</strong>(1946). Turning of Tirnovo. Economic Geography, 22(1), 42–54<br>DOI: 10.2307/141698</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=53890f18abcc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The History of Unix: From Bell Labs Experiment to Global Computing Infrastructure]]></title>
            <link>https://medium.com/@batuhanayrba/the-history-of-unix-from-bell-labs-experiment-to-global-computing-infrastructure-639be727fe0d?source=rss-6b60319cf5a7------2</link>
            <guid isPermaLink="false">https://medium.com/p/639be727fe0d</guid>
            <category><![CDATA[history]]></category>
            <category><![CDATA[unix]]></category>
            <category><![CDATA[computer-science]]></category>
            <dc:creator><![CDATA[Batuhan Ayrıbaş]]></dc:creator>
            <pubDate>Fri, 12 Jun 2026 19:49:45 GMT</pubDate>
            <atom:updated>2026-06-12T19:49:45.485Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*He1eaQ5OE9NkkMlwDdaDJw.png" /></figure><p><strong>Author:</strong> Batuhan Ayrıbaş<br><strong>Author note:</strong> Batuhan Ayrıbaş is an engineer and software developer focused on systems, algorithms, and practical research, according to his personal website, <a href="https://batuhanayribas.com/">batuhanayribas.com</a>.</p><h3>Abstract</h3><p>Unix is much more than an operating system family. In fact, Unix can be considered one of the most innovative architectures in computing’s intellectual history. Born out of the failure of the ambitious Multics program, which was designed as an experimental time-sharing system in the late 1960s, Unix quickly became a powerful and compact multi-user and multi-processing operating system. Many of its technical innovations, such as its hierarchical filesystem design, abstraction of hardware devices, piping, process management techniques, shell scripting, and systems programming using the C language, helped define the current generation of computing. This article traces the evolution of Unix throughout its long history, from its creation at Bell Labs to its widespread use as an academic tool, its commercial success, its standardization, and its continuing legacy in today’s Unix-based systems.</p><h3>Introduction</h3><p>The development of Unix represents a unique case for several reasons. Unlike most of the other great inventions in the world of software, Unix has had a significant influence on programming beyond the walls of its institutional creator. Most software systems developed out of large-scale efforts sponsored by corporations or governmental organizations. In contrast, Unix came out of the mind and hands of a few researchers from Bell Labs who wanted a convenient and intellectually satisfying computer system to work with. It had much more to do with constraints rather than abundance: little hardware, rejected proposals for funding, minimal institutional resources, and necessity.</p><p>However, this constraint became a strength in itself, allowing developers to create a flexible system based not on rigid architecture but on composability. The key features that emerged during the development of Unix were small programs, a textual interface, hierarchies, and the concept of viewing devices and files as part of a greater computational environment. All of these principles would be crucial in the future of software engineering.</p><p>According to Dennis Ritchie, the historical development of Unix dates back to 1969 despite being formally described in 1974. Ritchie talks about Unix not only as an interesting piece of technology but also as a product of culture and collaborative communication. Ritchie, Bell Labs/Nokia.</p><h3>From Batch Processing to Time-Sharing and Unix — A Short History of Multics</h3><p>Multics is the system that predated Unix. It originated when the researchers at the Massachusetts Institute of Technology (MIT), General Electric, and Bell Labs decided to develop the next generation of timesharing computers. They envisioned a system that would implement the then new ideas related to security, files, dynamic linking, and interactive computing Ritchie. Nevertheless, for Bell Labs team members, Multics came to represent all things that took too long and turned out to be too complicated Ritchie. The Bell Lab researchers left the project in 1969. They continued to share the dream of an interactive system, however. Their goal was to recreate the time-sharing system but in a more manageable and more realistic way. That’s why they created Unix. One can say that the original Unix was a compact recreation of ideas and goals of Multics rather than a completely new approach.</p><h3>2. Creating the First Edition of Unix</h3><p>Unix started off in 1969 on a little-used DEC PDP-7 at Bell Labs Ritchie. The idea to build the system occurred in part to Ken Thompson who needed a better environment for his space-travel simulator game Ritchie. In this context, the group led by Ritchie and Thompson began working on a file system, a kernel, and other useful utilities. According to Ritchie, the file system became central in the new OS with Thompson doing most of the work and adding such concepts as device files Ritchie.</p><p>The name “UNIX” was a play on the “MULTICS” acronym. Initially it meant “UNics” Ritchie. The pun reflected an attitude towards simplicity and minimalism characteristic of Unix. Rather than trying to solve everything with a single solution, Unix offered a coherent set of abstractions that could be freely combined to solve problems.</p><p>By 1971, Unix already moved to the PDP-11 and saw regular use for text processing and patent-related activities at Bell Labs. 1971 can be considered the release of Unix’s “First Edition” (with assembler, a file system, fork(), roff, and ed) as documented in The Open Group’s historical timeline The Open Group. Note the practicality of the first use of the system in real-world activities.</p><h3>3. Key Technical Concepts in Unix: Files, Processes, and Shell</h3><p>Ritchie and Thompson published a key paper “The UNIX Time-Sharing System” in 1974, introducing the general public to this system. Ritchie and Thompson, 1974 highlight some important features of Unix: hierarchical file system, uniform handling of ordinary files and special files (such as files for devices), asynchronous processes, selectable command languages, and numerous utilities Ritchie and Thompson, 1974.</p><p>Among the key ideas that were implemented in Unix, perhaps the file abstractions stand out as the most fundamental. The file hierarchy allowed one to see all elements of the system as files, whether regular or device, thus simplifying interactions and making the system easy to expand. Special files in turn allowed one to interact with devices through the well-known read/write interface Ritchie.</p><p>The ability to easily manage processes was another powerful feature. The fork model made creation of processes easy while shells allowed for programmable use of the operating system. In particular, the users could create complex chains of actions based on commands, redirect output and input of processes, and even develop complex automated pipelines without any additional coding Ritchie and Thompson, 1974.</p><p>The pipe functionality was added to the system by Doug McIlroy’s suggestion in 1972. A pipe allowed one to take the standard output of one process and connect it to another process, thus feeding data from one program to another Ritchie. This led to the development of a huge number of small utilities designed for narrow tasks and capable of being combined into powerful pipelines. Such culture was probably one of the greatest contributions of Unix to software development.</p><h3>4. Unix Written in C and Portability</h3><p>The key moment in Unix evolution is probably the re-writing of the system in C language. Originally, Unix was programmed in assembler language and thus remained tied to a particular hardware platform. In 1973 the 4th edition of the OS was programmed in C — a language invented by Dennis Ritchie at Bell Labs. This marked the transition of Unix to other platforms. Thus, according to The Open Group’s timeline, the 4th Edition “makes OS portable, changing the course of OS history” The Open Group.</p><p>C made possible porting Unix from PDP-11 to other platforms. Although this type of portability does not correspond to modern standards of complete abstraction from hardware, it marked a breakthrough in the world of operating systems nevertheless. Moreover, it promoted the development of C as a language of systems programming. Thus, Unix and C complemented each other — the former became a testing ground for a language, and the latter became a means of moving beyond the limitations of hardware Ritchie. This symbiosis had a profound impact on later computer systems.</p><h3>5. Academic Diffusion and the Rise of BSD</h3><p>Unix spread through universities partly because AT&amp;T, constrained by telecommunications regulation, did not initially market Unix as a conventional commercial product. Instead, source code licenses were made available to academic institutions. Students and researchers could study, modify, and extend the system. This created a community of Unix-literate programmers and helped establish Unix as an educational operating system.</p><p>The University of California, Berkeley became especially important. The Berkeley Software Distribution, or BSD, began as a collection of extensions to AT&amp;T Research Unix. FreeBSD’s documentation describes BSD as source distributions from UC Berkeley that were originally extensions to AT&amp;T’s Research Unix <a href="https://docs.freebsd.org/en/articles/explaining-bsd/">FreeBSD Documentation</a>. BSD contributed major tools and technologies, including vi, the C shell, virtual memory work, and influential networking code.</p><p>BSD’s role in Internet history is particularly significant. The FreeBSD documentation notes that 4.2BSD included a widely distributed implementation of TCP/IP in 1982 <a href="https://docs.freebsd.org/en/articles/explaining-bsd/">FreeBSD Documentation</a>. This helped connect Unix culture to the growth of networked computing. Unix systems became natural hosts for research networking, and many Internet services evolved in Unix environments.</p><p>BSD also illustrates the legal and institutional complexity of Unix history. Because early BSD included AT&amp;T code, it required Unix source licenses. Later efforts such as Net/2 and 4.4BSD-Lite attempted to separate freely redistributable BSD code from proprietary AT&amp;T code. This history shaped the later development of FreeBSD, NetBSD, OpenBSD, and other BSD descendants.</p><h3>6. Commercial Unix and the Unix Wars</h3><p>By the late 1970s and 1980s, Unix had moved from research and academia into commercial computing. AT&amp;T’s System III and System V became major branches of commercial Unix. Hardware vendors also created their own Unix variants, including SunOS, HP-UX, AIX, Xenix, and others. This produced innovation but also fragmentation.</p><p>The “Unix wars” of the 1980s and early 1990s were struggles over standards, market control, and compatibility. Vendors wanted Unix’s reputation and technical ecosystem, but each had incentives to differentiate its own version. The Open Group’s historical account describes how AT&amp;T and Sun attempted to unify the market around their version in 1987, prompting other vendors to form the Open Software Foundation <a href="https://www.unix.org/unix_history.html">The Open Group</a>.</p><p>This fragmentation created a paradox. Unix had become powerful because it offered a portable environment and shared programming model, yet its commercial success produced incompatible variants. Developers and customers needed standard interfaces. This need led to POSIX and the Single UNIX Specification.</p><h3>7. Standardization: POSIX and the Single UNIX Specification</h3><p>Unix eventually became not just a codebase but a standard. This distinction is essential. “UNIX” as a trademark came to refer to systems certified as conforming to specifications, not simply systems descended from AT&amp;T source code. The Open Group became the custodian of the UNIX trademark and specification in the 1990s. ISO notes that The Open Group has been custodian of the UNIX specification and trademark since 1993 <a href="https://www.iso.org/news/2002/11/Ref837.html">ISO</a>.</p><p>The Single UNIX Specification helped define a common application programming environment. It allowed systems from different vendors to claim compatibility if they conformed to a shared standard. The Open Group’s timeline records the introduction of UNIX 95 branding, later UNIX 98, UNIX 03, and UNIX V7 certification <a href="https://www.unix.org/unix_history.html">The Open Group</a>. macOS, Solaris, AIX, and other systems have participated in this certification history.</p><p>Standardization preserved Unix’s influence even as the original code lineage became less central. A system could be Unix-like in design, POSIX-compatible in interfaces, or formally UNIX-certified by trademark. These categories overlap but are not identical. Linux, for example, is Unix-like and broadly POSIX-oriented, but it is not generally described as a certified UNIX system. macOS, by contrast, has had certified UNIX versions.</p><h3>8. GNU, Linux, and the Unix-Like Explosion</h3><p>The 1980s and 1990s brought another transformation: the rise of free software and open-source Unix-like systems. Richard Stallman announced the GNU Project in 1983 as an effort to build a free Unix-like operating system. The Free Software Foundation describes GNU as a Unix-like system created for user freedom and notes that GNU was usually combined with the Linux kernel to form a complete operating system <a href="https://www.fsf.org/news/forty-years-of-gnu-and-the-free-software-movement">FSF</a>.</p><p>Linux began in 1991 when Linus Torvalds started developing a Unix-like kernel for personal computers. Its success depended not only on the kernel but also on the surrounding GNU tools, compilers, libraries, shells, and utilities. Together, GNU and Linux made Unix-like computing widely available on commodity hardware.</p><p>This changed the social meaning of Unix. Earlier Unix culture had circulated through Bell Labs, universities, and commercial vendors. GNU/Linux extended Unix-like ideas into global volunteer communities, later joined by corporations, foundations, cloud providers, and embedded-device manufacturers. The result was a massive expansion of Unix principles into servers, smartphones, routers, supercomputers, and development environments.</p><h3>9. The Unix Philosophy</h3><p>Unix is often associated with a design philosophy: build small tools, make them composable, treat text as a universal interface where practical, and prefer simple mechanisms that can be combined. This philosophy was never a single formal doctrine in the early years, but it emerged from the system’s structure and community practices.</p><p>The pipe is the clearest expression of this philosophy. A command such as cat file | grep pattern | sort | uniq represents more than syntax. It reflects an architecture of cooperation among programs. Each tool does one task; the shell and pipe mechanism connect them. This approach influenced software beyond Unix, including scripting languages, data-processing pipelines, microservices thinking, and command-line tooling in modern development.</p><p>The philosophy also had limits. Unix security was simple by design but later required more sophisticated extensions. Text streams are elegant but not always sufficient for complex structured data. Process-based composition can be inefficient for some workloads. Still, Unix’s durability comes partly from the fact that its philosophy was pragmatic rather than absolute. It offered defaults that worked well for a wide range of computing tasks.</p><h3>10. Unix in Contemporary Computing</h3><p>Today, Unix survives in multiple forms. Some systems are formally UNIX-certified. Others are Unix-like but not certified. BSD descendants remain important in networking, security, storage, and operating-system research. Linux dominates servers, cloud infrastructure, containers, Android-based systems, and much of high-performance computing. macOS and iOS inherit important Unix and BSD foundations. Developer tools across platforms continue to reflect Unix conventions.</p><p>Even Windows, historically separate from Unix, has increasingly accommodated Unix-like workflows through tools such as Windows Subsystem for Linux. This does not make Windows a Unix system, but it shows the strength of Unix’s programming model. Developers expect shells, package managers, scripting tools, SSH, Git, C libraries, and POSIX-like environments because Unix shaped the grammar of modern computing.</p><p>The most enduring Unix legacy may be conceptual. It taught the computing world that operating systems could be portable, interactive, multi-user, networked, programmable, and built around composable tools. It also demonstrated that software ecosystems can grow through documentation, source sharing, academic exchange, standards, and communities of practice.</p><h3>Conclusion</h3><p>Unix began as a modest Bell Labs response to the failure and complexity of Multics. Its creators did not set out to dominate global computing. They wanted a good environment for programming, communication, and experimentation. Yet the system’s technical choices proved extraordinarily fertile: hierarchical files, device abstraction, process control, pipes, shell scripting, and implementation in C gave Unix a portability and conceptual clarity that few systems matched.</p><p>Its history is therefore not a simple sequence from invention to success. It is a layered story: Bell Labs research, academic diffusion, BSD innovation, commercial fragmentation, standardization, free software, Linux, and contemporary Unix-like infrastructure. Unix became influential because it was both a system and a style of thought. It offered not only code, but a way of organizing computational work.</p><p>More than fifty years after its creation, Unix remains visible wherever developers compose tools, administer servers, write portable C programs, use terminals, deploy containers, run macOS or Linux, or rely on Internet infrastructure shaped by Unix traditions. Its historical importance lies in this continuity: Unix did not merely survive as an operating system family; it became part of the intellectual foundation of modern computing.</p><h3>References</h3><p>Ritchie, D. M. “The Evolution of the Unix Time-sharing System.” Bell Labs/Nokia. <a href="https://www.nokia.com/bell-labs/about/dennis-m-ritchie/hist.html">https://www.nokia.com/bell-labs/about/dennis-m-ritchie/hist.html</a></p><p>Ritchie, D. M., &amp; Thompson, K. “The UNIX Time-Sharing System.” Communications of the ACM, 1974. <a href="https://dsf.berkeley.edu/cs262/unix.pdf">https://dsf.berkeley.edu/cs262/unix.pdf</a></p><p>The Open Group. “History of UNIX.” <a href="https://www.unix.org/unix_history.html">https://www.unix.org/unix_history.html</a></p><p>ISO. “ISO/IEC and The Open Group announce international approval of the joint revision to POSIX and the Single UNIX Specification.” <a href="https://www.iso.org/news/2002/11/Ref837.html">https://www.iso.org/news/2002/11/Ref837.html</a></p><p>FreeBSD Documentation Project. “Explaining BSD.” <a href="https://docs.freebsd.org/en/articles/explaining-bsd/">https://docs.freebsd.org/en/articles/explaining-bsd/</a></p><p>Free Software Foundation. “Forty years of GNU and the free software movement.” <a href="https://www.fsf.org/news/forty-years-of-gnu-and-the-free-software-movement">https://www.fsf.org/news/forty-years-of-gnu-and-the-free-software-movement</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=639be727fe0d" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>