What's the SINGLE most common reason teams choose NOT to import an upper ontology?

Steep learning curve and import bloat — every BFO term imported pulls dozens of axioms the team doesn't understand

You're modelling clinical processes. Which combination is the standard choice?

BFO as the upper + OBO Foundry domain ontologies (e.g. OGMS, GO) that already extend BFO consistently

No upper ontology — it's overkill for clinical data

Upper Ontologies — BFO, DOLCE, SUMO

Anchor your domain ontology in a foundational one — or accept that you'll re-invent its mistakes.

0/4 done

Overview

Upper Ontologies — BFO, DOLCE, SUMO

Anchor your domain ontology in a foundational one — or accept that you'll re-invent its mistakes.

Why it matters

Upper ontologies pre-answer the hard philosophical questions (continuants vs occurrents, objects vs processes, roles vs types). Importing one is cheaper than re-deriving the answers.

Going deeper

Pick-list, with what each upper is good at and the price you pay:

Upper	Sweet spot	Cost
BFO (Basic Formal Ontology)	Biomedical, scientific data; lean (~40 classes)	Spartan: you'll need extensions for many domains
DOLCE	Cognitive / linguistic / cultural domains	Heavier, more philosophical commitment
SUMO	General knowledge, IT, enterprise	Large; full import is unwieldy
gist (Semantic Arts)	Business / enterprise modelling	Smaller community, less academic backing

Two rules that keep upper-ontology adoption sane:

Import a module, not the whole upper. All the major uppers ship MIREOT- style fragments — just the terms you actually subclass from.
Document why each upper class was chosen as the parent of your domain class. Future you (and reviewers) need to read it.

Analogy

An upper ontology is the foundation slab a house is built on.

You can build a house directly on dirt. It'll stand up for a while. Then the first winter cracks the walls, the second year your back door doesn't close, and by year five you're paying twice as much in repairs as the slab would have cost.

BFO, DOLCE, and SUMO are pre-poured slabs. They've already settled the load-bearing decisions: is the role of CEO the same kind of thing as Tim Cook the person? (No.) Is Q4 2024 a thing or an event? (Event.) Is the temperature of this room a property or its own entity? (DOLCE has an opinion.) Skipping them means you'll re-litigate every question, badly, in design reviews — forever.

Tools & resources

BFO — Basic Formal Ontology source + tutorials: https://basic-formal-ontology.org/ · GitHub: https://github.com/BFO-ontology/BFO-2020
DOLCE+DnS Ultralite — the most-used DOLCE module: http://www.ontologydesignpatterns.org/ont/dul/DUL.owl
SUMO — Suggested Upper Merged Ontology browser: https://www.ontologyportal.org/
gist (Semantic Arts) — enterprise-flavoured minimalist upper: https://github.com/semanticarts/gist
MIREOT explanation (Minimum Information to Reference an External Ontology Term) — how to import a module rather than the whole upper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3104852/

Make it stick

Use the prompts below to anchor upper ontologies — bfo, dolce, sumo to a real ontology you care about.

›What philosophical question does your domain quietly repeat in every design review (role vs type? event vs object? container vs content)?
›Which upper ontology already answers that question — and what would the smallest pilot of importing it look like?
›What's the cost in your team of having *no* shared upper today: how often do model decisions get reversed?

Reading in progress · 0 of 4 activities done

Overview

Upper Ontologies — BFO, DOLCE, SUMO

Why it matters

Going deeper

Analogy

Tools & resources

Tools & resources

💭Make it stick

Make it stick