notes-cog-ai-reasoning-myOntology

this isn't (yet) an ontology, but just some notes on parts i like/dislike from others' ontologies.

See also:


Distinctions i do like

set/type/(individual?)

Types are distinguished from sets because they can have intensional definitions.

Eg in UFO-A [4]; [5], the defn of type: "an entity that has an extension (...a set of entities that are instances of it) and an intension, (...applicability criterion for determining if an entity is an instance of it); and which is captured by means of an axiomatic specification, that is, a set of axioms that may involve a number of other entity types representing its essential features. A type is a space-time independent pattern of features, which can be realized in a number of different individuals".

i would drop the requirement for an axiomatic spec, because my feeling is that cognitive psychology and linguistics have shown that humans don't always use axiomatic specifications to define categories, which means that a number of common categories would not be 'types' under this definition; i would prefer to allow types defined by mixtures of definitions and prototype (see https://en.wikipedia.org/wiki/Prototype_theory ); and perhaps to have a subtype of 'type' which is a 'clearly-defined-type' which has an axiomatic spec. Note that for a type which is not a clearly-defined-type there may be 'edge cases' where different speakers disagree (or don't care) whether or not certain entities fall within the extension of the type.

Note that types can be instances of other types, even though types are not individuals.

It seems clear that individuals, of which there can only be one, are objectively distinct from sets and types (although i suppose that in a parallel world there could be another of you, but i would say that this is really a new type consisting of multiple individuals).

(Substance? Individual?)/(property? attribute? relation?)

It seems objective that 'Bob' is a different sort of thing from 'blue'; even though 'blue' can itself be a subject, and even though we can speak of an individual instance of blueness, eg Bob's blueness.

I guess Aristotle's distinction of things that cannot be predicated upon others is a good one here (to say that a person has the property of Bob-ness implies that that person is identical to Bob; in a pointer-equality sense, not just a structural equality or an equivalence sense).

atomic event vs physical object

A person might be a process, but a person is not an atomic event.

things that exist on their own vs. things that inhere

eg Bob's blueness can't exist without Bob; it inheres in Bob

forms vs individuals

As argued below ('a little later'), the distinction is that a single form (eg an equation) can have multiple simultaneous 'representations' (eg you write the equation twice on the chalkboard), whereas an individual (eg a person) cannot (if something that we count as an individual did seem to acquire a second representation, eg if you copied the person, then as soon as it acquired the possibility of a non-identical and non-unitary future, (eg since the copy might in the future receive different sensory input and acquire different memories, and be unable to synchronize these with the original), then we would consider it to be a distinct individual).

Types have a formal component (the intensional definition) and an extensive component which might well involve a set of individuals, so we may as well keep them as a separate thing (although their intensional definition component is itself a form).

Distinctions i am not sure about (later: i think i've resolved this)

'forms' vs non-form types

abstractions vs physical things

The idea of Science is pretty different from a person or an atomic event. Perhaps it is only a type of brain configurations though?

Pointer equality vs. structural equality/equivalence

This instance of a triangle and that instance of a triangle might have exactly the same properties as a shape, but they might be different instances. They do not have pointer equality, although they are equivalent under the homomorphism that discards all non-shape information.

my current thinking

Consider chemistry and physics. These are both instances of the type 'sciences'.

Now consider the equation "1 + 1 = 2". This is an individual 'thing' of some sort. I might write it on two places on a blackboard. But you would say that these are merely two different symbols (signs) that both refer to the same referent (the abstract equation). Similarly, i might say that "1 + 1 = 2" 'occurs' in the middle of proof A, and that it also 'occurs' in the middle of proof B; but you would say that, although in one sense there is a type of thing described by "1 + 1 = 2" of which the occurences in proofs A and B are instances, in another sense these are still the "same equation" and hence in this other sense there is only one "1 + 1 = 2".

Perhaps define a 'form' to be a type such that if there are two distinct instances of the type found in the world, we say that these are 'identical' in the sense that, with respect to the (substance? topic? set of relevant attributes? type?) of the form, they are completely determined and identical in every way.

This may or may not be the same thing that https://en.wikipedia.org/wiki/Type%E2%80%93token_distinction#Occurrences is talking about: in the sentence "A rose is a rose is a rose", there are three 'words' (a, rose, is) if 'words' means 'word types'; and there are 8 'words' if words means 'word occurrences'. (or is that just isa vs subtype?)

Some difficulties:

a little later: a think the key is to distinguish between 'representations' and 'instances'. An idea has a 'representation' in the configuration of your brain; the equation has a 'representation' on the chalkboard. But Chemistry and Physics are instances of the idea of Science. Based on this, we can indeed distinguish between individual things that could have multiple simultaneously existing representations (like an equation) and things that could have only one (like an individual person; if you made an exact living copy of a living person, you would have to start calling them two separate individuals, assuming their minds weren't linked).

Now, the intensional definitions of Types are necessarily abstract, and so can have various representations. But it's unclear if the Type as a whole can, since it has only one extension; and some Types' extensions may contain individual physical objects. But by the argument above, there is definitely a different between an abstract individual like an equation and a concrete one like a person.

Distinctions i dont like

perdurant/endurant

This distinction (eg in UFO-A [6]; [7]) seems somewhat subjective or at least context-dependent; at one time, one might view a person as an endurant, but then later in the same conversation one might view the same person as a process existing in time, a perdurant, similar to a storm. Similarly, one might take an endurant perspective on a storm. Perhaps this is only similar to the way that, in Aristotle's Categories, a particular person is a substance, which is distinct from a quality like a color, but one could grammatically and conceptually think of a color as the subject of a sentence, and ascribe predicates to it; in Aristotle, a color is not a substance just because it can be predicated upon subjects, but this does not imply that it cannot itself be a subject. Contrast to the distinction between an atomic event and a person; these things are objectively different, because a person exists (or at least, might exist) in more than one point in time.

agent

This seems to be context-dependent, eg is an AI robot an agent? Is a software program bargaining on your behalf an agent?


ontology

and possibly a Type (composite consisting of a set (its extension) and a form (its intentional definition))? but this seems very composite to me.

Relations i like

To be considered

catChIndex

the following are all listed on [10]

the random list of categories in https://en.wikipedia.org/wiki/Category_of_being#Categories_of_being and Aristotle's categories both seem pretty good:

umbel 32 'SuperType?'s in 9 clusters ('dimensions') seem okay:

http://techwiki.umbel.org/index.php/UMBEL_Specification#SuperType_Dimensions

SUMO

BFO

GIST

GFO

schema.org types

DOLCE

REA

opencyc

DOLCE+DnS? Ultralite (DLL)

DOLCE-Lite and GFO

ontology4 merge

aristotle's categories

hegel's categories

kant's categories

random list of categories in https://en.wikipedia.org/wiki/Category_of_being#Categories_of_being

Sowa's ontology