Language Agnostic Knowledge (LAK)
A New Ontology for Multilingual Knowledge, with LeraLink and ISSH (Iran Academia) as Its Living Laboratory
Introduction
For centuries, knowledge has been filtered, structured, and distributed through systems that privilege a single language, primarily English. This has created an invisible barrier between vast bodies of thought and the communities that produced them. The result is a double failure: societies cannot fully access locally relevant research, and the global scholarly community cannot benefit from the diversity of perspectives that already exist.
In the age of Artificial Intelligence and interconnected infrastructures, this imbalance is no longer a necessity—it is an error. The time has come for what I call Language Agnostic Knowledge (LAK): an epistemology, a framework, and a practical infrastructure for building a knowledge system that is not bound to a single language but instead operates across all of them.
The Problem We Face
Today’s education and scholarly communication systems are failing faster than ever. They are unable to adapt to the speed of technological change, the social need for inclusivity, and the growing recognition that language itself is a form of structural inequality. Research shows that linguistic bias in scholarly communication leads to invisibility, lower citations, and reduced career opportunities for non-English speakers . At the societal level, it prevents communities from benefiting from research produced in their own contexts.
Global infrastructures like Crossref, ORCID, DataCite, and OpenAIRE are aware of these issues, but the underlying problem persists: metadata schemas and evaluation systems are still overwhelmingly designed with English as the default. Initiatives like the Helsinki Initiative on Multilingualism and the UNESCO Recommendation on Open Science have pushed the conversation forward, yet implementation lags.
This is not simply a matter of accessibility. It is a matter of equity, epistemic justice, and the survival of cultural and intellectual diversity.
The Origin of LAK
The roots of LAK go back much further than its first public articulation. Since childhood, I have been obsessed with the problem of how knowledge is fragmented, distorted, or lost across languages. By 2020, I had begun formulating this obsession into a structured project, and from 2023—when the AI boom accelerated—it became my professional focus. I have to thank Nina (my wife) for being the first person supporting my crazy ideas.
What emerged was a newly formulated knowledge ontology designed to support a robust, language-agnostic system of knowledge. This ontology is not an abstract schema but a living foundation for scalable, AI-native infrastructures.
I first shared LAK in an articulated form with a group of highly educated and expert people in various fields including AI and Data engineering, scholarly writing, design, system thinking, teaching, knowledge management, etc., then at the Crossref Sprint in Madrid (2025), and continued refining it in Oslo (2025) at the PKP Sprint . What a great opportunity to meet so many great people. There, multilingualism emerged as the unavoidable center of every discussion, whether the focus was on workflows, publishing tools, or culture. It was clear: the current infrastructure cannot deliver the inclusivity we need without a fundamental rethinking. The title of the session I was thankful to lead was “How to get rid of MS Word”. I felt it encapsulates many things and functions as a meme. It did, and it worked.
Ramazan Turgut, a fellow polyglot whom I first met in Madrid, immediately recognized the significance of LAK. Together, we embarked on an advocacy journey initiated by him. Over time, we established connections between LAK and PKP’s infrastructure, particularly OJS, OMP, and OPS. I believe these are the most pragmatic mediators between today’s scholarly publishing and tomorrow’s open, language-inclusive knowledge systems. Ramazan was highly skilled and experienced, and I gained numerous new insights. I want to acknowledge that he not only enriched LAK with his inputs and expertise but also encouraged me, recognized my value, and motivated me to pursue this more in the academic realm. I hope more individuals like him will join and support the LAK initiative. LAK will serve as a blueprint, inspiring the creation of numerous applications like LeraLink..
ISSH: The Home and Testbed
None of this work would have been possible without the Institute for Social Sciences and Humanities -
(AKA Iran Academia) in The Hague. From the very beginning, ISSH has been the home where these ideas could be tested, refined, and given real-world shape. ISSH is the first fully online, tuition-free, multilingual graduate school, and it has provided not only the institutional foundation but also the trust and support to pursue LAK.I owe particular gratitude to Ali Reza Kazemi, the founder and executive director of ISSH Foundation, who placed his faith in this project and in me and encouraged me and supported me to carry it forward. ISSH is more than a host and partner—it is the environment where multilingual knowledge infrastructures are not only theorized but lived, every day, in practice.
My current role as an educational development and innovation strategist at ISSH is to revamp our entire publication infrastructure, utilizing PKP products such as OJS, OMP, and OPS. I envision this transformation as a pioneering example of multilingual publication within our existing infrastructure. This initiative marks a significant step towards the future of knowledge, encompassing not only scholarly production but also other forms of knowledge supported by ISSH.
In a recent and exciting development, we have also begun laying the technical foundations of LeraLink in collaboration with Mehdi Mohseni (in his free time), a highly talented and experienced system architect and senior developer, and more importantly, a compatriot and friend with whom I experienced the hardship far beyond building LeraLink. His expertise is helping transform the LAK ontology into a functioning platform. We will share more about this progress once the time is right, but the groundwork is being set.
The Foundations of LAK
LAK rests on several principles:
Language Agnosticism: Knowledge should not be trapped within the boundaries of a single language. Metadata, identifiers, and systems must be designed to operate independently of linguistic constraints.
Co-Creation of Knowledge: Knowledge is not owned by publishers, platforms, or elites. It emerges through collaborative production, translation, annotation, design, proofreading, editing, supervising, reviewing and exchange.
Integration with Persistent Identifiers (PIDs): ORCID for people, ROR for organizations, DOIs for works. These identifiers must support multilingual fields, cross-lingual discoverability, and transparent provenance.
AI-Native and Ontology-Driven: LAK leverages AI not as a patchwork translation tool but as a native element of the system—extracting, enriching, and aligning knowledge across languages through ontologies and semantic frameworks.
Social Learning and Cognitive Efficiency: Building on cognitive load theory, LAK aims for systems that are not only rigorous but also usable, exchangeable, and trackable. Learning should be as engaging as Netflix, as social as TikTok, and as rigorous and contextually rich as the best academic curricula..
From Concept to Application: LeraLink
LAK is not only a theoretical vision—it is embodied in LeraLink, the unified knowledge system I have been developing together with many others. I would like to thank Sadik and Bujar Bakiu, Polina Cico, Armando Demaj, Ziad Moubayed, Mike,
and Marjan de Jong, along with many others whose names cannot all be listed here, for their invaluable support at different phases of this journey.LeraLink is designed with an atomic architecture, or what I prefer to call a fractal architecture. The smallest unit, known as a Piece, can represent any knowledge element anywhere multiple times while maintaining its single point of truth.
In short, LeraLink is a working example of how LAK principles can be applied in practice.
Rising Beyond the Legacy
LAK is not a solitary work. It builds upon the pioneering work of infrastructures and initiatives that have shaped today’s knowledge ecosystem:
Crossref and DataCite for persistent identifiers and metadata networks.
PKP for democratizing open publishing infrastructures.
OpenAIRE and EOSC for aggregating and interlinking research outputs.
I am grateful to be a member of the Crossref Metadata Advisory Group, and to have witnessed first-hand the potential and limitations of these systems. LAK is not a replacement—it seeks to align them toward a common horizon.
Fellowship in Italy: Extending the Journey
As part of this ongoing journey, (of course, with Ramazan’s usual encouragement) I was recently awarded a fellowship with the H2IOSC–OPERAS Transnational Access Programme at CNR ILIESI (National Research Council of Italy) in Rome. This opportunity allows me to deepen the work on Language Agnostic Knowledge (LAK) in collaboration with Ramazan Turgut and to situate our ideas within one of Europe’s most respected centers for ontology and multilingual scholarship. Our aim is to work closely with the institution’s experts and researchers, to refine LAK’s ontology, to test its edge cases, and hopefully, to co-create conceptual models that integrate seamlessly into FAIR and multilingual infrastructures. Beyond the technical dimension, I hope this fellowship becomes a place where I can meet new colleagues willing to join us in building the foundations of equitable, language-agnostic knowledge systems.
Allow me to clarify: While terminology may vary across different fields—such as decolonizing knowledge, socio-linguistic bridging, or global North-South collaboration—the underlying objective remains consistent. The focus is on enhancing the availability, accessibility, and diversity of knowledge, fostering meaningful exchanges that transcend socio-linguistic barriers.
Beyond One Essay
This is not a manifesto, but a beginning. The work of LAK is ambitious: rethinking the architecture of knowledge in ways that dismantle linguistic monopolies, redesign ownership, contribution, and incentive models, and integrate AI responsibly into the heart of education and research.
It is also deeply personal. My journey as a life-long learner, project manager, researcher, and educator of methodology and knowledge management has always been shadowed by the failures of existing systems. ISSH was a home for experimentation, while LeraLink is becoming the embodiment of these principles. At VU Amsterdam too, I am grounding this vision in academic research while preparing my trajectory for a PhD-level inquiry.
Ontology-Driven Multilingual Publishing Pipelines for Epistemic Justice in AI-Era Knowledge Infrastructures
AI, Multilingualism, and Knowledge Infrastructures: Toward Epistemically Just and Equitable Learning Ecosystems.
I investigate how AI-native, multilingual publishing pipelines in LeraLink can operationalize epistemic justice by enforcing language-aware metadata, PID integrity, and linked-data interoperability at design time. The work formalizes a five-tier knowledge ontology—optimized for atomic knowledge units and cross-language versioning—mapped to established bibliographic/LD models (conceptually consonant with FRBR/FRBR-LRM and BIBFRAME) while remaining implementation-specific to our stack (tiers: Piece, Source, Volume, Series, Collections). Multilingual rigor follows COAR’s repository recommendations (ISO-639 language coding, UTF-8, multilingual keywords, OAI-PMH exposure) and BCP 47 tagging requirements used on the web and in Schema.org (inLanguage), ensuring machine-actionable discovery and UI parity. I also examine Crossref practices for translations and multilingual records (separate DOIs plus hasTranslation relationships; language metadata in deposits), treating them as constraints on deposit/harvesting logic, with PKP’s multilingual publishing guidance informing workflow ergonomics and metadata capture. The result is envisioned as an AI-assisted, ontology-driven infrastructure that resists digital colonialism by making non-English scholarship first-class, traceable, and interoperable across identifiers, catalogs, and web search surfaces.
Peer-Review Reviewed
In developing LeraLink, we inevitably have to revisit the processes of peer review and the broader question of quality metrics, especially in relation to qualitative scholarship. Conventional models of evaluation, often biased toward quantitative indicators or English-language outputs, are insufficient for an infrastructure that seeks to be language-agnostic, inclusive, and epistemically just. Reimagining peer review in this context means treating it not only as a gatekeeping mechanism but as a collaborative, multilingual, and traceable exchange that acknowledges diverse epistemic traditions. This requires new forms of quality measurement—dynamic, context-aware, and transparent—aligned with the ontology of knowledge we are building. I will continue developing these ideas in dedicated discussions with like-minded colleagues as part of the LeraLink framework.
Learning Is a Knowledge Infrastructure
Is learning—and by extension online education—truly a separate category, or is it simply a particular way of structuring, accessing, exchanging and integrating knowledge? Once we move into the domain of knowledge ontology, the distinction between “learning infra” and “knowledge infra” begins to collapse. Educational programs, especially online ones, are not more than specialized treatments of knowledge flows and communication with knowledge and particular ways of exchanging it. But knowledge is not only something to be linked. It is the foundation of wisdom and synthesis, the raw material out of which we form original thought, actual thinking, and genuine learning. If we build infrastructures that are less flawed, more inclusive, pluralistic, and just, we lay the foundation for an entirely new learning era. That conviction is precisely why Sina and I called this project Lera.Link—because learning is not separate from knowledge, but the very process of knowledge becoming alive, connected, and transformed into thought.
Will You Join?
If LAK resonates with you, this can be an invitation. Not to adopt a ready-made solution, but to join a growing conversation about how we build the next knowledge infrastructure—rigorous, multilingual, equitable, AI-native, and social. A new knowledge economy beyond the status quo.
The horizon has never been brighter. But it will not wait.