Ontology, Part I
The Architecture of Meaning
Intentional Arrangement is a reader-supported publication. Consider becoming a free or paid subscriber. And thank you for helping to make this writing journey possible ❤️ Here’s 20% off an annual subscription, in gratitude.
This is the first article in a four-part series, all about ontologies. Ontologies have emerged as a sought after brass ring in the AI race but they are also the most confusing aspect to many. There is confusion around the concept, the terminology, the technology stack — all of it. The confusion is worth noting because it has become a barrier to entry. Understanding the causes of this confusion is the first step toward working productively with ontologies.
Why the Confusion Exists
If you have spent any time in the semantic technology space, you have had some version of this experience. You read an article that says an ontology must be expressed in OWL. You read another that says any RDF model with classes and properties qualifies. A third says a taxonomy is a kind of ontology. A fourth says an ontology is a knowledge graph with extra steps. A vendor at a conference calls their product category tree an “ontology-powered AI solution.” Two textbooks contradict each other on where the boundary falls.
You are not confused because you are missing something. You are confused because the field is flooded with conflicted messaging and too many domain experts regaling audiences with different and partial versions of the truth. The mixed messaging has real consequences — for practitioners trying to build, for organizations trying to invest, and for AI systems that increasingly depend on formally structured knowledge to perform reliably.
This essay lays the foundation for the series. It traces the concept of ontology from philosophy, into computer science. It walks through the core building blocks of the Semantic Web, examining which standards qualify as ontologies and which do not (a surprisingly interesting question), and introduces the Ontology Pipeline® alongside the well-known ontology spectrum as two complementary frameworks for understanding where ontologies fit in the broader landscape of knowledge infrastructure.
Let’s first examine some of the common misperceptions around ontologies because the confusion around ontologies has identifiable sources.
