{"id":29957,"date":"2026-10-16T08:00:06","date_gmt":"2026-10-16T06:00:06","guid":{"rendered":"https:\/\/voelkerrechtsblog.org\/?p=29957"},"modified":"2026-10-16T17:57:11","modified_gmt":"2026-10-16T15:57:11","slug":"against-the-suspicion-of-machines","status":"publish","type":"post","link":"https:\/\/voelkerrechtsblog.org\/de\/against-the-suspicion-of-machines\/","title":{"rendered":"Against the Suspicion of Machines"},"content":{"rendered":"

Fifteen years ago, on EJIL:Talk!<\/a>, I mulled over what was then an emerging but hugely growing practice: academic expert blogging. I called it, only half-ironically, a \u201chazardous tool\u201d. The hazard, at the time, was that blogging seemed to threaten the slow, deliberative, peer-reviewed economy of scholarship. Blogging was then said to be too fast, too casual, too public, too exposed. Serious scholars, we were told, did not blog. Blogging was for the impatient, the self-promotional, the intellectually unserious, or, worse, for those who had not yet earned the right to speak in the sacred pages of a proper refereed journal. Fifteen years on, blogging is uncontroversial, and one struggles to think that it was once contested. It is, in fact, part of the ordinary architecture of academic life.<\/p>\n

I want to reflect today on another \u2013 this time genuinely hazardous \u2013 transformation in the way scholarship is carved and produced: the now constantly discussed arrival of generative artificial intelligence (hereafter AI) in the writing, reading, and circulation of scholarship. While blogs undoubtedly reshaped how quickly ideas could travel and who could participate in scholarly debate, AI is reinventing how texts are produced, organized, revised, proof-read, referenced, indexed, circulated, and evaluated. Interestingly, just as with blogging and previous changes in how scholars manufacture their texts, the initial reaction of academia today has been suspicion, defensiveness, and a certain panic dressed up in the language of \u201cintellectual integrity\u201d. And once again, just as fifteen years ago, I believe that the current academic suspicion, defensiveness, and panic \u2013 along with the editorial policies they have produced \u2013 are misplaced, and framed in the wrong terms.<\/p>\n

Before going any further, an important disclaimer is in order. What follows should not, under any circumstances, be read as the expression of some na\u00efve euphoria at the arrival of an unprecedented technology in academic practice. We all know perfectly well what sort of political, environmental, anthropological, cultural, intellectual, social, and financial dystopia the generalization of AI risks ushering in if it is not properly framed, tamed, and regulated. My point is a different one. It is simply that the current academic suspicion towards the use of machines in scholarly writing rests on a mistaken premise, in so far as it seeks to preserve the romantic and mythical figure of the single-minded sovereign author, and with it an old regime of inscription that was never as unmediated as its defenders now presuppose. Having misframed the debate in this way, contemporary academia would be on far firmer ground invoking the intolerable environmental cost of these technologies, the unprecedented concentration of capital they enable, and the new technologies of governance of the mind they make possible, to justify its current conservative, repressive, and defensive practices. All these objections are momentous and pressing. But these are not the objections that drive the current academic suspicion towards the use of AI.<\/p>\n

The Permanent Sniff Test<\/strong><\/p>\n

What kind of suspicious attitude am I speaking about here? I have come to see that scholarly manuscripts today are increasingly read through a single anxious lens: has this been generated by AI? Indeed, just like in the publishing industry<\/a> in general, a growing number of scholarly articles and books are nowadays being delayed, rejected, or disqualified because they \u201csmell\u201d of the work of the machine. A paragraph sounds too polished. A transition is too smooth. A sentence is too well-balanced. A stylistic tic, the famous em-dash, the tricolon, the tidy conclusion, the misplaced quotation mark, etc. become prima facie evidence of algorithmic contamination. In this context, many editors and reviewers, in the name of a holy \u2013 but very enigmatic \u2013 principle of \u201cintellectual integrity\u201d, now perform a role of amateur forensics, hunting for the telltale rhythms of the machine instead of looking for intellectual quality or field-shaking originality.<\/p>\n

It is true that, since the advent of modern thought<\/a>, being critical has been synonymous with a systematic suspicious attitude<\/a>, the latter still being the dominant mindset in the humanities today. Yet, I want to argue here that the current generalized suspicion towards the possible use of AI in the production of scholarship and the generalization of editorial practices zeroed in on the detection of the use of AI are deeply misguided. They miss the point entirely. To put it bluntly, I think most of the current policies being adopted by academic actors (see e.g. here<\/a> and here<\/a>) simply approach the matter in the wrong terms. The relevant question is not whether AI has contributed to the production of a given text meant to become scholarship. The relevant question is whether the text thinks, whether there is a proper intellectual movement in it, whether it argues, narrates, provokes, connects, disrupts, illuminates, and eventually makes readers think. To put it differently, the question is not whether inscription has been delegated to a machine \u2013 which is fine as I will argue here \u2013 but whether thought has. That distinction between delegation of inscription and delegation of thought, as I will claim here, is the decisive compass for navigating the controversies related to the use of AI in the production of scholarship. I am of the opinion that it is academia\u2019s inability to grasp such distinction that is producing a witch-hunt that is not merely futile but actively corrosive of the intellectual creativity it claims to defend. Not to mention the practices of cancellation of disruptive voices such witch-hunt facilitates and to which I will come back later.<\/p>\n

Our Modes of Inscription Have Always Been Changing<\/strong><\/p>\n

Let us begin with a brief historical observation. The modes by which humans, and especially those we call scholars today, have produced inscription, i.e. the material act of putting thought into durable form, have evolved continuously across millennia. Unsurprisingly, each transformation in our modes of rendering thought into durable form was greeted with roughly the same suspicion as that we now witness in relation to the use of AI.<\/p>\n

For instance, it is well-known that, in the Phaedrus, Plato has Socrates recount the myth of Theuth, inventor of writing, who presents his discovery to King Thamus as an aid to memory and wisdom. Thamus replies that writing will instead encourage forgetfulness, because people will rely on external marks rather than internal recollection, and that it will produce only the appearance of wisdom in those who have not truly learned. Socrates adds that written words are like paintings: they seem to speak, but cannot answer questions, defend themselves, or adapt to their audience once sent out into the world. The irony, of course, is that this critique of writing survived because Plato wrote it down, making it one of the classic examples of early resistance towards the evolution of human modes of inscription.<\/p>\n

The consolidation and generalization of the printing press<\/a> offers another illustration of how a new mode of inscription is commonly met with defensiveness. As is well-known, the new printing press was immediately denounced by many sixteenth-century clerics as a machine of vulgarization that would flood Christendom with unvetted opinion in vernaculars with a lower degree of sophistication than Latin. Even Erasmus, who owed his European reputation to print, worried about the torrent of pestilent little books the press was making possible.<\/p>\n

In the same vein, some of us may remember that, not so long ago, following the rise of computational and digital technologies, questions arose whether the word-processing software, the citation manager, the translation software, the dictation tools were going to destroy the practice of careful composition and slow writing. Yet, nobody today inquires whether an author used Google, Google Scholar, Zotero, Word, or DeepL (and requires that such uses be divulged). These are tools of inscription, and they have all quietly been absorbed into the ordinary practice of scholarly writing.<\/p>\n

My point here is that the use of AI in scholarly writing belongs to this historical continuum. From the perspective of the history of scholarly literature, AI is a technology of inscription, a rather extensive one admittedly, but just a technology of inscription. In that sense, the current suspicion towards the use of AI witnessed in academia can possibly be construed as a mere transitional phenomenon. I am convinced that, in half a decade, asking whether a text was inscribed with the assistance of AI will seem as preposterous as asking whether a manuscript was typed with word-processing software rather than handwritten. And, for the same reason, I believe that the editorial policies that the main actors in the field are hastily putting in place are soon going to prove remarkably short-lived, if not an embarrassment in the eyes of next generations.<\/p>\n

The Single-Minded Author Has Always Been a Mythical Figure<\/strong><\/p>\n

There is yet another reason why the current mainstream defensive reaction of academia against AI is without proper foundations. More often than not, the current panic over AI is not really about AI. It is about authorship, that modern construction on which the distribution of academic authority still rests. Yet, the author, as we learned from post-structuralist philosophy, is never the sovereign origin of the text. The author is an effect of the text, produced retrospectively as a way of organizing, legitimating, and disciplining meaning (in that regard I have never understood why Roland Barthes\u2019 The Death of the Author<\/a> was not made a compulsory component of any university curriculum in the humanities). The author, in other words, has always been a mythical, composite, and intertextual figure hovering above the text and meant to perform various functions<\/a>: stabilizing, authorizing, and making the text legible within certain institutional practices, while also allowing repression when a text overcomes the boundaries of what is socially or institutionally admissible. The figure of the author is even more mythical when it comes to scholarly texts which have always been the product of prior texts, inherited forms, disciplinary conventions, technologies of writing, editorial interventions, research assistants, anonymous reviewers, copyeditors, colleagues\u2019 comments, academic debates, etc. The lone genius authoring a manuscript ex nihilo is a romantic fiction that has never reflected how scholarship actually gets made.<\/p>\n

There is even something comical in the fact that the humanities, of all places, should still be clinging to the figure of the single-minded author. Indeed, it has been more than a century since psychoanalysis dismantled the very idea that a human being thinks, speaks, or writes as a unified sovereign agent. From the perspective of psychoanalysis, the author as a self-identical, self-transparent, sovereign source of meaning has always been an impossibility well before any machine entered the scene.<\/p>\n

If the foregoing were not sufficient to make my point, let\u2019s remember that, until not so long ago, we lived without the modern cult of the author. Medieval intellectual and artistic practices routinely produced works, manuscripts, songs, folk traditions, theological commentaries, without attributing them to a singular named author. Scribes copied, glossed, and often silently rewrote the texts they transmitted. Aquinas famously dictated to several secretaries at once, moving between arguments while his assistants inscribed what he said, the great scholastic summae ending up being assembled through an elaborate network of collective dictation, copying, glossing, and correction. Nobody at the time thought this compromised the authority or integrity of the works concerned. Texts were collective artifacts without causing any intellectual impoverishment.<\/p>\n

The argument here is not, of course, that we should return to anonymous scholarship (although this has at times been the fantasy of some scholars<\/a> who wish books could be judged for what they do rather than for who wrote them). It is rather that AI exposes that authorship has always been mediated, distributed, composed, and constructed. In that sense, one of the main transformations brought about by AI-generated text is that it reveals that the author has always been an after-effect of the text, a figure that we conventionally construct and interrogate once we have already engaged with the text with a view to seeking authority for our interpretation of that text. In that regard, AI could be seen as doing us a considerable favour. It is helping us desacralize the author by bringing down this mythical figure from the pedestal on which academic writing has long placed it.<\/p>\n

AI-Hunting as the New Cancellation Practice in the Humanities<\/strong><\/p>\n

There is yet a further dimension of the generalized academic suspicion towards AI that deserves to be mentioned here because it is beginning to shape a new and disturbing form of everyday politics of the humanities, and of legal scholarship in particular. The hunt for AI use runs the risk of turning into a new mechanism of academic cancellation. Accusing a colleague of having used AI, and therefore \u2013 so the reasoning goes \u2013 of having produced work devoid of intellectual authority, can become a remarkably efficient way to discredit a scholar without ever engaging with the content, originality, or novelty of what they actually argue. The accusation disqualifies before it discusses. And, most conveniently for those who deploy it, it requires no evidence beyond the usual stylistic patterns associated with machine-made texts or inaccurate referencing.<\/p>\n

What makes this development particularly troubling is that such practice could someday (if this is not already the case) be used as a tool to stifle the disruptive or subaltern voice, to brush aside the uncomfortable questions, to bring into disrepute works that unsettle doctrinal orthodoxies or that refuse the ideological or political consensus of the field. This possibility should alarm anyone who cares about the humanities. The Global North has entered, more or less everywhere, a new era of authoritarian politics, in which the disqualification of speech through all kinds of gimmicks and tactics has once again become a routine instrument of governance. Universities, which used to imagine themselves as counterweights to such authoritarian tendencies, could ironically reproduce them internally through a generalization of the AI accusation in the name of \u201cintellectual integrity\u201d.<\/p>\n

Cultivating Self-Discipline<\/strong><\/p>\n

That the current witch-hunt against the use of AI in academia is misplaced, if not dangerous, as I have tried to demonstrate in the previous paragraphs, does not mean that anything should go when it comes to using AI in the production of academic scholarship. Hardly anyone would disagree that intellectual and creative processes that give rise to scholarship should be preserved. But what does that platitude mean concretely? In my view, if the argument, the intuition, the connections, the imagination, the critique, the disruptive story, the risks taken, the anger, the rage, the enchantment, the love, or the provocations remain engineered by the scholar\u2019s mind (in a non-romanticized sense) then the delegation to a machine of certain acts of inscription, that is the rendering of thought into text, should never be, in itself, problematic. As a tool for inscription, i.e. as a tool of mise en texte, AI should not be subject to the current generalized suspicion that has overtaken the editorial boards of academic journals and book series. AI-assisted writing becomes troubling when it replaces thought, not when it serves as one of thought\u2019s mode of inscription. In my view, and as was said earlier, the relevant dichotomy in this debate, is not between human writing and machine writing but between the delegation of inscription and the delegation of the entire thinking process to a machine.<\/p>\n

It is true that the distinction between thinking and writing has long been porous. The distinction I draw here between thinking and inscription should similarly be approached as a fluid binary that each scholar negotiates differently. Indeed, I am convinced that everyone practices a different kind of thought-forming process. Sometimes the form coexists with the formation of the idea; sometimes the idea precedes the form; sometimes the form precedes the idea. The frontier between a thought and its rendering into form is therefore not drawn in the same way by any two scholars, nor, indeed, in the same way by the same scholar from one piece of writing to the next. This is why I do not feel the need to spell out here what, in my view, qualifies as inscription and what counts as thought. More importantly, this is the reason why academic actors, while using the distinction between the delegation of inscription and the delegation of thinking as a guiding principle, should resist the temptation to formalize it into a list of permitted and prohibited acts.<\/p>\n

The foregoing explains why I believe that the survival of thought in scholarship will not depend on general prohibitions, elaborate AI-detection software (which, incidentally, cannot be fully trusted), obligations of disclosure, disciplining practices, and the drawing of lists of permitted and prohibited delegations. What does it depend on then? In my humble view, it will depend solely on the self-discipline of the scholar who must constantly reevaluate the breaking point at which the delegation of the rendering of thought into form turns into the delegation of the thought-forming process itself. It must be acknowledged that such self-discipline may prove extremely demanding because it cannot be formalized into a code or a convention. Indeed, as was said earlier, the frontier between the thought-forming process and its inscription is not drawn in the same place for everyone. That frontier is not fixed and sits in a different place for each scholar since for some the form coexists with the formation of the idea, for others the idea precedes the form, and for others still the form precedes the idea. Self-discipline, in this context, cannot be reduced to the application of a settled rule to oneself. It is the ongoing practice of heeding where, in a particular piece of writing, on a particular day, with a particular prompt, the delegation of inscription has quietly begun to carry out operations that belong to the thought-forming process. It is a practice of constant self-scrutiny that has to be relearned and rehearsed with each new project, each new intellectual inquiry, and each new piece of scholarship.<\/p>\n

It is my belief that universities should be one of the places where that sort of self-discipline is encouraged, cultivated, and, above all, passed on to the next generations. Precisely because the frontier between inscription and the thought-forming process cannot be codified, it cannot be taught through regulations or conventions alone. It has to be transmitted the way scholarly craft has always been transmitted: through apprenticeship, seminars, supervision, etc. But universities cannot be the only such places. The self-discipline necessary to a thought-preserving use of AI must not become the exclusive privilege of those who happen to have passed through university education, for it would then begin to function as a new social marker. The fracture to be feared is not a fracture in access to the machines but a fracture in the capacity to use them in a disciplined manner. That is a far more consequential inequality, and one that will have to be addressed well beyond the walls of the university, in schools, professional bodies, and the many other sites in which minds are trained.<\/p>\n

A final observation is warranted at this stage. The self-discipline that ought to be cultivated at universities and beyond may well help us preserve the practice of thinking, but it will not prevent our ways of thinking from being altered by the use of AI. Surely, prompting changes how we think. The prompt is not merely an input command. It reorganizes the very architecture of our intellectual work: how we formulate a problem in the first place, how we select which questions are worth pursuing and prompting about, how we sequence and stage an argument, how we anticipate counter-arguments, how we judge when an idea is complete, and how we endure the long, uncomfortable intervals in which nothing yet coheres, etc. In particular, prompting compresses many of these operations, outsources others, and quietly displaces the rest. This new cognitive attitude has already been given a name: prompt thinking<\/a>. Such transformations are not necessarily something to be wary of, since the way we think, like the way we inscribe, has always shifted across the centuries. One need only recall how word-processing software reshaped the thought-forming process for many of us. Yet, even if it is too early to assess what prompting does to cognition and thought, and even if the effect is not harmful in itself, it is something that will need to be kept under watch in the years and decades to come. This, however, is not a question that ought to be addressed here for it is not what drives the current academic suspicion towards the use of machines.<\/p>\n

The Sad But Inevitable End of the Positivist Commentary<\/strong><\/p>\n

There is a specific practice whose sustainability legal scholars should be worried about. To appreciate such worry, let\u2019s remember that legal scholarship (just like most of the humanities<\/a>) is a coded discourse. It operates through structured vocabularies, formal categories, recognizable argumentative moves, a well-delineated repertoire of authorized references, and predictable rhetorical patterns. It is, in the most literal sense, a language of codes: statutory codes, doctrinal codes, jurisprudential codes, stylistic codes, critical codes, referential codes, bibliographical codes, etc. And here is the uncomfortable truth of the AI era we have entered: when it comes to coding, decoding, and recoding, AI will always beat us<\/a>. This is why AI is astonishingly good at producing traditional legal scholarship precisely because law and legal scholarship are astonishingly patterned and coded. A large language model trained on the corpus of judicial decisions, treatises, and law review articles can produce a highly competent case note, a well-structured and compelling doctrinal journal article<\/a>, and a convincing summary of the state of the art, often faster and more thoroughly than a scholar. This means that a whole category of legal scholarship, for instance the positivist commentary on a judicial decision or a given practice or the descriptive or policy-oriented account of a regulatory regime, can now be delegated to a machine without much loss of quality. This bears a major social and epistemological consequence in my view. The legal scholar (as I once was) who spends their career coding an already-coded practice, restating doctrine in slightly more elegant prose, or producing policy recommendations has now been outclassed by the machine and must accordingly reinvent themselves completely.<\/p>\n

In These Circumstances, How to Keep Out-Thinking AI? <\/strong><\/p>\n

Contrary to what the anthropomorphic notion of \u201cartificial intelligence\u201d may imply, the machine does not think<\/a>. AI tools are statistical models trained to predict the most probable continuation of an utterance based on its context. This is why the scholar is not made obsolete completely as long as they uphold thinking. Not restating, not summarizing, not synthesizing, but thinking. That means producing arguments, stories, narratives, images that are not already latent in the training data, making connections the machine cannot make because they are not yet inscribed anywhere for it to extract. That also means seeking the unexpected juxtaposition, the productive misreading, the overinterpretation, the argument that reframes a debate, the intuition that runs against the doctrinal habit, etc.<\/p>\n

Out-thinking AI also requires a strong commitment to interdisciplinarity. Indeed, AI is ultra-efficient within a discipline (a discipline is, after all, a code) but much weaker across disciplines, where the codes do not align, where translation is required, where the frictions between vocabularies and intelligibility frameworks produce insight rather than statistical linearity. I am convinced that only interdisciplinary scholarship will survive the current algorithmic revolution.<\/p>\n

To out-think AI, the scholar must simultaneously increase their reliance on non-Western literature, traditions of thought and intellectual practices. As is well-known, the training data on which contemporary AI relies is overwhelmingly Anglophone, overwhelmingly Western, overwhelmingly rooted in the intellectual corpus of the Global North. By virtue of such epistemic imperialism, non-Western and non-modern practices, philosophical traditions, and scholarly literatures are drastically under-represented in what the machine relies on. Yet, such practices, traditions and literatures are where thought has the most room to move, blossom, innovate, and reinvent itself. This explains why they must come to constitute a much larger source of our insights than they do today if we are to out-think AI. And the day AI will have ingested all rare books<\/a> from such non-Western and non-modern traditions, the intellectually ambitious AI-using scholar will have to move, once more, to unfamiliar and uncharted territories in order to stay ahead of the machine, without such constant quest for new literary materials having any end in sight. Seeking a constant intellectual exile will be key in upholding the relevance of scholarship in the humanities.<\/p>\n

At the same time, the scholar who uses AI, but who does not want to be outclassed by it, has an absolute obligation to be a far better and more extensive reader than their predecessors ever were. This point cannot be overstated, and it cuts against the lazy assumption that AI somehow relieves us of the burden of reading. It is the opposite. Precisely because the machine can now perform, with astonishing fluency and efficiency, the operations of summary, synthesis, and doctrinal restatement, the scholar\u2019s value has to be entirely projected to the terrain the machine cannot occupy: the deep, wide, slow, patient, cultivated reading across the humanities. Never in the history of scholarship have books been as important as they are now. And I am speaking here of a specific type of books: sustained, challenging, long-form works of philosophy, history, anthropology, literature, linguistics, narratology, theology, political theory, aesthetics, etc., and the countless other fields whose vocabularies do not yet coincide with the field one identifies oneself with. Such books are precisely what an algorithm cannot give us, namely a space in which to think, at length, without being told where the argument must go. This is the reason why, in the age of AI, we need more books, not fewer. We need more time with books, more space for them, more eagerness to be slowed down by them.<\/p>\n

If reading is now the condition for scholarly survival, then academia must protect and reward it in ways it has largely ceased to do. An academy that rewards slow reading, patient thought, and genuine intellectual range is one that does not surrender to the machine. In the end, the future of thinking in scholarship will not be salvaged by detection software, disclosure mandates, disciplining practices, or the paranoid reading of em-dashes. It will be secured by scholars who continue to read voraciously and widely as well as universities that encourage them to do so.<\/p>\n

Where to Go From Here?<\/strong><\/p>\n

This short essay has tried to show that the contemporary witch-hunt against AI-using authors and the editorial policies that formalize the current suspicion towards the resort to AI in scholarly writing are misplaced. Academic actors should accordingly cease to inquire, in the name of \u201cintellectual integrity\u201d, whether AI was used in the production of a piece of scholarship. Such suspicion has been vain from the beginning, not because it is bound to fail to catch the use of AI, but because its premise has never been the right one. The only germane question for editors and reviewers is whether the text thinks, whether the argument has depth, whether the intellectual movement is novel or nourished by new thoughts and intellectual traditions. Whether the scholar has delegated inscription to a machine is beside the point and so is the elaborate apparatus of policing and disclosure that academia has built around that non-question.<\/p>\n

Editorial note: In a Call for Contributions<\/a>, V\u00f6lkerrechtsblog invites reflections on the changing conditions under which legal and academic knowledge is produced.<\/em><\/p>\n

 <\/p>\n","protected":false},"excerpt":{"rendered":"

Fifteen years ago, on EJIL:Talk!, I mulled over what was then an emerging but hugely growing practice: academic expert blogging. I called it, only half-ironically, a \u201chazardous tool\u201d. The hazard, at the time, was that blogging seemed to threaten the slow, deliberative, peer-reviewed economy of scholarship. Blogging was then said to be too fast, too […]<\/p>\n","protected":false},"author":35,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[6639],"tags":[],"authors":[7245],"article-categories":[6000],"doi":[],"class_list":["post-29957","post","type-post","status-publish","format-standard","hentry","category-uncategorized","authors-jean-daspremont","article-categories-article"],"acf":{"subline":"Why Academia Should Stop the Witch-Hunt"},"meta_box":{"doi":""},"_links":{"self":[{"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/posts\/29957","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/users\/35"}],"replies":[{"embeddable":true,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/comments?post=29957"}],"version-history":[{"count":10,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/posts\/29957\/revisions"}],"predecessor-version":[{"id":30018,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/posts\/29957\/revisions\/30018"}],"wp:attachment":[{"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/media?parent=29957"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/categories?post=29957"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/tags?post=29957"},{"taxonomy":"authors","embeddable":true,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/authors?post=29957"},{"taxonomy":"article-categories","embeddable":true,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/article-categories?post=29957"},{"taxonomy":"doi","embeddable":true,"href":"https:\/\/voelkerrechtsblog.org\/de\/wp-json\/wp\/v2\/doi?post=29957"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}