In an ontology, concrete concepts such as "Dog" or "Person" stand for sets (classes) of invididual instances. Abstract concepts, such as "love", "hate" or "wisdom" don't have any instances, but if they are empty sets, they are all identical. If they are individuals on the other hand, they are not "concepts" in the first place (also, individuals of which class?).

Where is my error in reasoning and how would one model such abstract concepts in an ontology?

  • 2
    Why do you think that love and hate in particular do not have instances? We use them to denote instantiated emotional statuses, aren't we? They correspond neurophysiologic patterns. You seem to think that all that is not individual, particular object of perception does not exist (naive Realism). The fact that presupposing this, speaking of love, hate, morals, free will, etc. becomes totally meaningless (although it is the very core of our values and meaning in life) should indicate that this view lacks something. E.g. Pragmatism is able to circumvent these problems.
    – Philip Klöcking
    Oct 24, 2016 at 15:39
  • @PhilipKlöcking, not quite - there is no brain state which corresponds to love, hate, etc... much less the utterances of "I love you", "I hate you" or what have you. Furthermore, you need to distinguish when something is totally meaningless from "totally meaningless to you". Your appeal to core values is simply a canard and there is no good reason to doubt direct or naive realism is the case. This is philosophy, not the love of sentiment.
    – MmmHmm
    Oct 24, 2016 at 20:02
  • 1
    @Mr.Kennedy: sigh 1) The Amygdala is a part of the brain associated with emotions and shows, together with surrounding areas, very specific patterns of neural activity that correspond to certain feelings the proband experiences. 2) I just restated the very old problem of fatalism arising when we deny ontological being of what constitutes values, i.e. the outcome that they are mere chimera. 3) Naive or direct realism is challenged in many ways, not the least of them scientific realism, because e.g. the ontological status of "colour" as emergent property is highly problematic.
    – Philip Klöcking
    Oct 24, 2016 at 20:13
  • 1
    Neither "dog" nor "person" stand for classes of individuals, but (on the traditional interpretation of "meaning" at least) for (lists of) properties used to identify individuals as dogs or persons. "Love", "hate" or "wisdom" also have "individual instances", and the lists of properties that allow identifying them would form their concepts. "Brain state" version does not work because the same concepts can be instantiated by different states in different individuals or even in the same individual at different times, as Putnam pointed out to early physicalists.
    – Conifold
    Oct 24, 2016 at 22:33
  • @PhilipKlöcking 1) a "sigh"? The diaphragm is associated with emotional states - are we to take your statement of a breath type as demonstration of behavior indicating that you are romantically infatuated? 2) Not quite, again, you need to distinguish what is true from what is "true to you". Fatalism is not implied by stating the case that abstract ideas do not exist and they are only to be found in language. This is simply the ontological status of abstract ideas 3) Except for unilluminating appeals to skepticism, there is no good reason to doubt naive realism e.g. color is but prismatic range
    – MmmHmm
    Oct 25, 2016 at 1:24

4 Answers 4


You can do it either way. In 'traditional' first-order logic, the only individuals are objects, not properties. In first-order logic, one can say "there is an object that has the property F" or "all objects have the property F" but one cannot "quantify over" properties themselves.

However, in second-order logic this is possible. In fact, Leibniz's Law, which summarizes something important about identity is a second-order formula: It says, "for two objects x and y, [if x = y, then for every property F (Fx if and only if Fy)]."

  • In order that x = y, iff is over kill, because F(x) -> F(y) is sufficient: if x does not have a property, then y does not have that property either. F(x) -> F(y) guarantees that g(y)^~g(x) never happens. Oct 24, 2016 at 20:22
  • @GeorgeChen, I don't follow on how the plain material conditional rules out the G(y) and not G(x) case. I think you need the biconditional because you want to say intuitively that if x and y are the same then they have all and only the same properties. The plain material conditional doesn't rule out this case: two objects x and y, such that y has all the properties that x does, but more besides. Let G be such a property that y has and x lacks.
    – user5172
    Oct 24, 2016 at 23:58
  • Let F(x) = ~G(x), i.e. x has the property of ~G. If F(X) -> F(y), then ~G(y). That is, if x does not have the property G, y either. Here the predicate is "not having the property G." Thus material implication is sufficient to rule out y having any properties that x doesn't have. The key point is "does not have G" is also a property. Oct 25, 2016 at 0:53
  • @GeorgeChen Your symbolism here is presupposing that the negations of properties are also properties. I don't know enough about the standard semantics for second-order logics to know for sure, but that looks like a suspicious assumption to me. Just because G is a property does look to me like ~G automatically makes a new property (in the sense that properties are items in the domain of discourse). Fness could exist without not-Fness existing. To predicate not-Fness of something is just to say the thing lacks Fness. But I'm getting out of my depth here.
    – user5172
    Oct 25, 2016 at 1:16
  • One is either ill or not ill. Being not ill is the property everyone wants to have. Of course being not ill is synonymous to being healthy. Whether or not to use "not" to represent a property is non-essential. Oct 25, 2016 at 1:42

Long comment

We have in place two dichotomies :



universal/particular (or individual).

If we use the term "concept" not to denote a psychological entities (a mental representation) but as a component of an (objective) thought, it is an abstract and the basic relation is that of "falling under" :

Fido is a dog

is true because my dog Fido (an indivudual) falls under the concept "dog" (an universal).

But a concept can be predicated also of another concept :

prudence is a virtue.

Here we have an universal that is a "subset" of a more wide universal : both are abstracts.

Roughly speaking, particulars are objects and universals "curve out" classes i.e. collections of objects falling under the corresponding concepts.

One "big" issue is :

are there abstract objects ?

If we agree that e.g. numbers are abstract objects, we have that the number 2 is an individual falling under the concept "even".

  • Numbers are abstract in the sense that e.g. the number two is abstraction of all sets with two elements, e.g. "two" abstracts from "two women", "two trees" and so on. Some languages did not make this abstraction step and still have different numbers for different entities. On the other hand I would say a particular number is an individual of the class (concept) "number" so that e.g. "2" is neither abstract or concrete, because it is not a concept but an individual. Oct 28, 2016 at 13:34

The notion of a word as a reference to a set of objects fails for obvious reasons best described by Quine's discussion of Natural Kinds.

Late Wittgenstein proposes that instead of a set of objects, a definition is a token in a game that transfers power in the form of knowledge. The power that a definition gives is the ability to predict the behavior of the world around you, including other people.

The notion of love is no different from the notion of dog in this respect. Dogs behave a given way, they cause us to experience various things, and the people around them behave a given way because the dog is present. So does love, it 'behaves' in certain ways, increasing or decreasing and evolving in character, it causes us to experience various things and people behave in given ways when it is present.

So the right way to model both in an ontology is in terms of potential and expected effects.


Abstract ideas are universals - it is also referred to as concepts. A universal is a quality; a particular is a bundle of qualities.The best way to single out a quality is by pointing out what different particulars have in common - this process is called abstraction. All concepts are not equally abstract: the following examples show that there are hierarchies of concepts and that some concepts are extracted from other concepts.

Dog, red and person are not concrete; love, hate or wisdom are not as abstract as one thought.

Red is a class of sensations. No one has ever seen red alone; no one can imagine red independent of other visual properties such as area, shape, depth, brightness, etc, although it is not difficult to imagine a red thing, e.g., Clifford the Big Red Dog. Red is actually abstract - hard to believe, I know - it is the name of a universal. Names standing for particulars are called proper names, e.g., Clifford the Big Red Dog; a name that stands for nothing is a nonsense - from here we depart from the Platonist view. Whether the Platonist view is true or not I personally do not think it's worth considering at all.

Dog stands for a class of animals each of which we call a dog; no one can see or even imagine what dog looks like, although one can easily imagine an instance of dog - Clifford the Big Red Dog, for example. Dog is actually a universal or an "abstract idea" or a concept. On this account, the English language, which puts an indefinite article before a noun, cultivates a very good habit of mind: an indefinite article before a noun indicates that the noun stands for a universal - people have been thinking like this all the time, but it takes a Bertrand Russell to point it out. (English language is a manifestation of a superior mind: there is a famous Chinese paradox "white horse not horse," which is no paradox at all in English, owing to the English indefinite article; this example illustrates how language affects thinking.)

In Mauro Allegranza's answer, Fido, a proper name in the strictest sense, is used to stand for a particular. This is the kind of precision we should at least make an effort to practice in here.

In the case of a particular dog, everything a person knows about it he knows through his sensations in his own head. The dog in itself as an individual is not the origin of our knowledge about it; it is the sensations in our own head that are the origins of our knowledge about the dog. We believe the dog was the beginning of a chain of physical events that lead to our sensations, but everything we can say about the physical world is inferred from sensations; as a matter of fact, the causal chain between the dog and the canine patch of colour in the mind is open to doubt: there is no evidence to suppose we are not dreaming. It follows that in order for a proper name to have meanings, it is only necessary that it invokes mental images; this explains why Hamlet is considered a fiction while Louis XIX is considered meaningless: there is a whole play describing Hamlet in picturesque terms, whereas Louis XIX invokes no mental image.*

Like dog and red, love is also a universal or an abstract idea, but it is not detached from particulars: it is a class of particular mental states or feelings. An instance of love is not intrinsically different from an instance of seeing a dog: they are all mental occurrences, although we believe they both have physical causes.**

All universals derive their meanings ultimately from particulars: some are classes of particulars, some are class of classes of particulars, some others are class of classes of classes of particulars, etc; there are hierarchies of concepts - that is why Bertrand Russell said he discovered the theory of types because that is how people actually think. For example:

Red is the class of all red sensations; red is a colour, but "Clifford the Big Red Dog is a colour" is not true, therefore colour is a class of classes.

2 is the class of classes each of which has only two members; number is a class of numbers and is therefore a class of classes of classes.

A cleat is a member of a carefully selected pair; a pair of cleats is an instance of 2; 2 is an instance of number.

Regarding everything that exists only in thoughts as abstract ideas - many dictionaries are still defining abstract this way - has licensed many nonsenses disguised as "abstract concepts" and has given aura to a mysterious manner of speech that sounds like announcing aphorisms containing profound truths but is actually emitting nonsenses. If a word stands for neither a universal nor a particular, it is a nonsense, or an incomplete symbol at best.


  • Please do not mistake me for an idealist. Idealists believe the existence of things depends on the mind; what I'm talking about here is our knowledge: our knowledge of the external world - including our knowledge of the existence of things - begins from the mind, not the external things. If you are inside a submarine, everything you can say about the outside depends on what the sensors tell you. Knowing the true origins of our knowledge does not increase the credibility of the mind; on the contrary, it gives us more reason to doubt everything we firmly believe.

** Germans don't say "Ich liebe dich" very often because when they say it they mean it. Americans say "I love you" all the time as a compliment - whether or not this expresses a mental state does not matter - but this kind of lies are fully approved of by Shakespeare.

Source: Russell, Bertrand. "Knowledge by acquaintance And Knowledge by Description." The Problem of Philosophy.

  • 1
    I don't think this answers the main question of how one can model a concept. Stating that "Dog" is an abstract of which there exist instances does nothing to illuminate what (if anything) the content of that abstract is.
    – commando
    Oct 25, 2016 at 5:55
  • What people called abstract concepts are just hierarchies of universals, and people don't realize how ubiquitous they are in our language. But I don't want to sound too "abstract" in my answer. @Mauro Allegranza's answer is on point. Yes, there are controversies about the reality of "abstract objects," in which I do not have the slightest interest. Oct 25, 2016 at 12:02

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .