Once we begin finding the semantics of a system, we need to consider and organize these semantics according to their "scope."
Is the semantic term or statement "universally true?" Few things are universally true. Those that are, we are trying to work into our semantic primitives. For example, we could say that an object that recorded an event that has already occurred should have an occurred-on date, and that date should be in the past. It should have an occurred-at place, but that place may be fairly imprecise (Illinois) or extremely precise (latitude 41.872833, longitude −087.62441; street address: 720 S. Michigan Avenue, Chicago, IL).
We might uncover a subtype of historical event that is not a physical, real-world event, such as a Web search or a chat session, and only be able to say that it occurred at a place but not know where that place was.
Some assertions are true in some places and not in others. In Singapore, vehicle tax (which is some 250% of the value of a new vehicle) can be drastically reduced by retiring another, older car from the island population.
In many western states, property owners are obligated to build fences to keep animals (primarily cattle) out of their property; in most eastern states the obligation is on the animal owner to keep them in.
In the United States, a sport utility vehicle (SUV) is considered to be a small truck and therefore is exempt from certain mileage and safety requirements.
There is a much larger set of facts, terms, axioms, and so on that are true for an entire industry. If nothing else, the semantics of most of the vocabulary will belong to the industry.
For example, in the mining industry a claim is the deed to mineral rights on a given unit of land, whereas in the health care industry a claim is a bill for services. A mineral claim must have a precisely defined geographic boundary, and to be enforceable (a subtype or status of claim) it must be registered with the county or counties within which the claim is located.
Although it may seem that many items are company specific, on closer examination, we see that few of them are. Most items reflect policy, "the way we do things here," identification of products, channels, subclassing relationships with customers and vendors, and so on.
Rules relative to categorizing a customer as a "cash only" customer or a carrier as a "preferred carrier," or definitions of what constitutes a "full load," are examples of items that have a company-wide scope.
There is a branch of philosophy that holds that each person lives in his or her own world, created by his or her perceptions and categorizations. This seems to be an extreme view, and if true it invalidates most of this book, to say nothing of 2000 years of science and philosophy.
However, there are things that are true only for an individual or a small group. To make this manageable it needs to be an extension of what is true for everyone. Preferences fall into this category, especially preferences for user interface display. However, many items that have worked their way into a company-wide categorization actually belong in only a few individual's semantics. For example, rules about categorizing tasks for the purpose of reestimating or calculating reserves can be restricted to a small number of people.