A corpus crosses an epistemic threshold when counting ceases to be its most informative description. Scale initially appears as quantity—documents, records, nodes, papers, datasets—but accumulation becomes intellectually consequential only when each unit can be distinguished and when relations among units can be made explicit. Persistent identification supplies the elementary architecture of this transition. Juty et al. establish that uniqueness, persistence and resolvability are not administrative conveniences but conditions of addressability: an object that cannot be reliably named cannot participate durably in a distributed knowledge system. Cousijn et al. radicalise the point through the PID Graph, where identifiers cease to operate as isolated labels and become vertices in a topology connecting publications, data, researchers, institutions and other research entities. Metadata produces the edges. What emerges is not simply a better catalogue but an infrastructure capable of generating questions from relations themselves. Haris, Stocker and Auer push the same transformation inside the scholarly document.
The Open Research Knowledge Graph treats research contributions as structured semantic objects, gives machine-actionable ORKG Papers persistent DOI-based identities and preserves change through linked provenance chains. Publication is consequently split into two complementary operations: stabilisation, which makes a state citable, and versioning, which allows knowledge to continue moving without dissolving its historical trace. Islam adds the operational membrane required for such systems to work across heterogeneous domains. PID-level metadata distinguishes machine readability, semantic interpretability and machine actionability, allowing software to know not only that an object exists but what kind of object it is, how it relates to others and which operations can be performed upon it. Priem, Piwowar and Orr reveal the macroscopic consequence in OpenAlex: hundreds of millions of scholarly works become navigable because scale is subordinated to a heterogeneous directed graph of typed entities, persistent IDs and canonical external identifiers. The crucial transformation is thus from magnitude to relational scale. A large corpus is not yet a knowledge infrastructure. It becomes one when numerical expansion is accompanied by persistent addressability, semantic differentiation, provenance, interoperability and traversable relations. The same distinction clarifies what an expanding Socioplastics corpus is actually constructing. Eleven thousand nodes are significant only provisionally as a quantitative marker; their stronger epistemic meaning lies in the possibility that nodes, essays, PDFs, DOI records, bibliographies, slugs, indexes, images and external scholarly identities increasingly behave as a coordinated topology. The architectural problem is therefore no longer production alone but the governance of relational density. Every new object introduces a demand for position: what is it, where does it belong, which prior elements does it modify or reinforce, how can a human retrieve it, how can a machine recognise it, and which identifier preserves its public identity when platforms or URLs change? Open Science makes these questions structural rather than cosmetic. DOI does not confer intellectual value by itself, just as node count does not demonstrate epistemic density; both become powerful when embedded in metadata, provenance and interoperable networks. This reframes the corpus as a public spatial system. Indexing acts like circulation, metadata like orientation, identifiers like durable addresses, references like passages, and APIs or machine-readable records like interfaces through which non-human readers enter the structure. The field can consequently grow without collapsing into an opaque archive because growth is countered by increasing differentiation and navigability. It also permits a more exact account of thresholds. A marker such as 11K should not announce that the system is important because it is large; it can register a phase change in which manual memory is no longer sufficient and explicit infrastructural devices become necessary to maintain coherence across the whole. At that point, the corpus begins to resemble the scholarly knowledge graphs described by OpenAlex and ORKG while retaining a different disciplinary and aesthetic constitution: heterogeneous cultural objects can remain heterogeneous because a shared infrastructural layer makes their relations legible without forcing them into one representational format. Socioplastics thereby moves from serial accumulation toward an addressable epistemic environment in which publication, archive and machine legibility become aspects of the same construction. Within Socioplastics, this constellation condenses as LegibleArchive.
Anto Lloveras is an architect and urban researcher working across spatial practice, epistemology, archives and knowledge infrastructure through LAPIEZA LAB and Socioplastics.
Cousijn, H., Braukmann, R., Fenner, M., Ferguson, C., van Horik, R., Lammey, R., Meadows, A. and Lambert, S. (2021) ‘Connected Research: The Potential of the PID Graph’, Patterns, 2(1), 100180. https://doi.org/10.1016/j.patter.2020.100180.
Haris, M., Stocker, M. and Auer, S. (2022) ‘Persistent Identification and Interlinking of FAIR Scholarly Knowledge’, in Robinson-Garcia, N., Torres-Salinas, D. and Arroyo-Machado, W. (eds.) 26th International Conference on Science and Technology Indicators, STI 2022. Granada. https://doi.org/10.5281/zenodo.6912480.
Islam, S. (2023) ‘FAIR digital objects, persistent identifiers and machine actionability’, FAIR Connect, 1, pp. 29–34. https://doi.org/10.3233/FC-230001.
Juty, N., Wimalaratne, S.M., Soiland-Reyes, S., Kunze, J., Goble, C.A. and Clark, T. (2020) ‘Unique, Persistent, Resolvable: Identifiers as the foundation of FAIR’, Data Intelligence, 2(1–2), pp. 30–39. https://doi.org/10.1162/dint_a_00025.
Priem, J., Piwowar, H. and Orr, R. (2022) ‘OpenAlex: A fully-open index of scholarly works, authors, venues, institutions, and concepts’, STI Conference 2022, Granada.
