I Drafted an (unfinished) essay about classification on Wikidata including metamodeling here : https://www.wikidata.org/wiki/Help:Item_classification

Metamodeling seems to me (and has proven to be) convenient in Wikidata. Think that there is a lot of classes, and that regrouping classes with some similarity in metaclasses can have a lot of applications, inlcuding: * tag classes in different ontology with some ''instance of'' <class of this ontology> metaclass to help query the class by intology * regroups class of the same kinds : classes of chemical elements can be regrouped because they all class atom with respect to their atomic numbers. This is what we call a chemical element and is consistent with the definition of chemical element in the corresponding french Wikipedia article. * regroup classes by definition of their instances : for example the laws of a country actually defined implicitely a lot of classes. Those classes can be regrouped using metaclasses like <class defined by french law> * ...

2014-09-25 20:07 GMT+02:00 Benjamin Good ben.mcgee.good@gmail.com:

Based on the CHEBI ontology perspective, alcohol is a class with subclasses like 'aromatic alcohol' which has subclasses like 'benzyl alcohols' which has subclasses like 'methylbenzyl alcohol' and so on.

These relationships seem worth capturing and subclass seems like a reasonable way to do it. If not, would another property be better?

Check out CHEBI here: http://www.ebi.ac.uk/chebi/advancedSearchFT.do?searchString=alcohol&quer...

-Ben

On Thu, Sep 25, 2014 at 10:52 AM, Alain Cuvillier < alain.cuvillier@gmail.com> wrote:

...

Hi all,

why are both the *subclass of* and *instance of* properties set for the ethanol (showcase) item? For me ethanol is a single concrete alcohol and it is not a class. There is only one ethanol, with a single chemical formula and structure, so only the *instance of* property is right for this item ?

https://www.wikidata.org/wiki/Q153

Thank you Alain

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l

Oh, It was written "ethanol" and I read "hydrogen" .... my bad /o\

Anyway, metaclasses are a way to solve this kind of paradoxes.

Lets say we got a class <alcohol types>.

This class would have, amongst its instances, the <ethanol> class.

But a basic principle of ontologies (philosophically) is the token/type distinction ( https://en.wikipedia.org/wiki/Type%E2%80%93token_distinction ), description logic are based on it, so also the OWL ontologies. Like Daniel said, this distinction discriminates token, real world object or events, and classes of this kind of objects. Abstract objects like ethanol are classes in this view. Its token are all the concrete ethanol molecules there exists.

To be consistent with your model we also can see the class of classes <alcohol types>, then <methanol> is an instance of it.

The beauty of things here is that we can do both :) If we accept that item can have both instance of and subclass of claims.

2014-09-25 21:32 GMT+02:00 Thomas Douillard thomas.douillard@gmail.com:

This is about chemical compounds, molecules. Here it's about atoms, I think. Correct me if I have wrong information, but afaik CheBi has no chemical elements class, so there is no real answer to the original question.

But yes, those classes if they are not already on Wikidata seems OK to have, and subclass of is appropriate as its semantics is the same in Chebi and in Wikidata.

2014-09-25 20:07 GMT+02:00 Benjamin Good ben.mcgee.good@gmail.com:

...

Based on the CHEBI ontology perspective, alcohol is a class with subclasses like 'aromatic alcohol' which has subclasses like 'benzyl alcohols' which has subclasses like 'methylbenzyl alcohol' and so on.

These relationships seem worth capturing and subclass seems like a reasonable way to do it. If not, would another property be better?

Check out CHEBI here: http://www.ebi.ac.uk/chebi/advancedSearchFT.do?searchString=alcohol&quer...

-Ben

On Thu, Sep 25, 2014 at 10:52 AM, Alain Cuvillier < alain.cuvillier@gmail.com> wrote:

...

Hi all,

why are both the *subclass of* and *instance of* properties set for the ethanol (showcase) item? For me ethanol is a single concrete alcohol and it is not a class. There is only one ethanol, with a single chemical formula and structure, so only the *instance of* property is right for this item ?

https://www.wikidata.org/wiki/Q153

Thank you Alain

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l

Hi,

I fully agree with Thomas and the other replies given here. Let me give some other views on these topics (partly overlapping with what was said before). It's important to understand these things to get the subclass of/instance of thing right -- and it would be extremely useful if we could get this right in our data :-)

What is a class and what is an item is often a matter of perspective, and it is certainly accepted in the ontology modelling community that one thing may need to be both.

The important thing is that "subclass of" is a relation between *similar* things (usually of the same type):

* "sports car" subclass of "car" * "Porsche Carrera" subclass of "sports car" * "Porsche 356" subclass of "Porsche Carrera"

Use "A subclass of B" if it makes sense to say "all A's are also B's" as in "all Porsche Carreras are sports cars".

In contrast, "instance of" is between things that are very *different* in nature:

* "Douglas Adams" instance of "human" * "human" instance of "species"

Subclass naturally forms chains, like in my example. You can leave out some part of the chain and the result is still meaningful:

* "Porsche Carrera" subclass of "car" [makes sense]

For instance of, this does not work:

* "Douglas Adams" instance of "species" [bogus]

So if you want to organise things in a hierarchy (specific to general), then you need "subclass of". If you just describe the type of one thing, then you need "instance of". It is perfectly possible that one thing participates in both types of relationships.

In addition to these general guidelines, I would say that a well-modelled ontology should be organised in "levels": whenever you use instance of, you go to a higher level; if you use subclass of, you stay on your current level. Each thing should belong to only one level. Here is an example where this is violated:

* "Porsche Carrera" subclass of "sports car" * "Porsche 356" subclass of "Porsche Carrera" * "Porsche 356" instance of "sports car"

Each of these makes sense individually, but the combination is weird. We should make up our mind if we want to treat Porsche 356 as a class (on the same level as sports car) or as an instance (on a lower level than sports car), but not do both at the same time. I think "subclass of" usually should be preferred in such a case (because if it is possible to use subclass of, then it is usually also quite likely that more specific items occur later [Porsche 356 v1 or whatever], and we really will need subclass of to build a hierarchy then).

Cheers,

Markus

On 25.09.2014 20:10, Thomas Douillard wrote:

Hi, this is a long discussion :) Is is allowed by OWL2 notion called "Punning".

The rationale is that Hydrogen is a chemical elements, and that the chemical element is not a subclass of atom. Rather a chemical elements is a type of atom, so chemical elements is a metaclass : a class of class of atoms. ​

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l

The statement "ethanol *instance of* chemical compound" is ontologically incorrect. Importantly, it is also incompatible with ChEBI, the most widely-used chemistry ontology.

The matter of how to apply *instance of* (P31, rdf:type) and *subclass of* (P279, rdfs:subClassOf) on Wikidata in relation to chemical entities has been, as Thomas puts it, a long discussion [1-5]. Hopefully with a wider audience and experts like Markus Krötzsch and Denny Vrandečić now interested, we can come to a resolution at least in the particular domain of chemical compounds. Since it concerns interoperability with another large Semantic Web project, I have copied Janna Hastings and Alan Ruttenberg on this discussion. Janna coordinates ChEBI. Alan coordinates BFO, the upper ontology used by ChEBI and many other major ontologies in the natural sciences, like Gene Ontology and Disease Ontology.

Denny indicates how the statement "Porsche 356 *instance of *car" would be incorrect in Wikidata even though "Porsche 356 *is a* car" is acceptable in everyday speech. Similarly, "ethanol *instance of* chemical compound" is incorrect in Wikidata even though "ethanol *is a* chemical compound" is acceptable in less formal contexts.

A key difference between talk about cars and talk about chemicals is that, with cars, we have familiar terms like "car model" that distinguish concrete instances (that *particular* car you see on the street) from abstract "instances" (i.e. metaclasses, classes that are also instances, the *kind* of car that you see on the street). We do not have a well-known term like "chemical model" or "chemical compound type" to distinguish classes (types) of chemicals and instances (tokens) of chemicals. When one speaks of the properties of ethanol or hydrogen, it is understood that the subject is *all concrete, particular, spatiotemporal tokens, i.e. instances *of ethanol and hydrogen -- not just a specific ethanol molecule floating in that container before you on a Saturday with friends, but all molecules that we label "ethanol" everywhere.

Thus, in order to formally classify ethanol itself as opposed to some particular ethanol molecule, we must say for an item like http://www.wikidata.org/wiki/Q153: "ethanol *subclass of* chemical compound" and not "ethanol *instance of* chemical compound". (On Wikidata, the statement is more precisely "ethanol *subclass of *alcohol", but it is entailed from the statements "alcohol *subclass of* organic compound" and "organic compound *subclass of* chemical compound" that "ethanol *subclass of* chemical compound".)

A common defense of statements like "ethanol *instance of* chemical compound" is that Wikidata will never have items about any concrete molecules of ethanol, so, since ethanol is a "leaf node" in our concept taxonomy, it makes sense to state that ethanol is an instance. That interpretation of "instance" is short-sighted. It precludes us from ever talking about particular tokens of ethanol, or particular aggregates of such objects, without overhauling our chemistry ontology. Excluding consideration of metaclasses like "chemical compound type", the fact that an entity is a leaf node in a concept hierarchy is a necessary but not sufficient condition for using *instance of*.

Another common suggestion is that we should state something like "ethanol *instance of* chemical compound type" and "ethanol *subclass of* chemical compound".

To see where that gets us, try wrapping your head around this: https://commons.wikimedia.org/wiki/File:Atom_classes.svg. Really, take a look. If we want Wikidata's concept hierarchy to be seen as of dauntingly complex, pervasively applying that kind of three-layer classification scheme will do.

The kind of explicit metamodeling seen when punning things like cars and car models, ships and ship classes, biological taxa and organisms, etc. works reasonably well in certain domains. But, while we hold that hammer in one hand, we should be careful not to see everything as a nail. Outside domains that have established vocabulary for metaclasses, imposing explicit metamodeling with statements like "ethanol *instance of* chemical compound type" or "hydrogen *instance of* atom type" will strike users as unduly complex.

Without such metamodeling, though, querying for a list of chemical compounds becomes murkier. Surely we would want to return "ethanol" and not "organic compound" in such a list. How about "alcohol"? Relatedly, if we don't state "oxygen *instance of *chemical element", then how can we easily query for all the elements in the Periodic Table of Elements without including in the results of any potential subclasses of oxygen (e.g., isotopes of oxygen like oxygen-16, oxygen-17, etc.)?

There are ways to achieve that in SPARQL using rdfs:subClassOf / P279 / *subclass of*, but they require adhering to certain conventions. When faced with requiring many potential query users to learn some Wikidata MetaObject Protocol, though, I'm inclined to make some sacrifices for simplicity, ontological correctness, and consistency with major existing ontologies.

In summary, this ball has punted for over a year now. Because of the impasse in how to classify chemical entities, we now have showcase items that have obvious problems like entailing that something is both a class and an instance of chemical compound. We need input from a wider group of people knowledgeable about ontology or chemistry, ideally both. Hopefully with a Wikimedian in Residence at the Royal Society of Chemistry [6] we'll get some more focused resources on this. All major scientific ontologies use *subclass of* (rdfs:subClassOf), not *instance of* (rdf:type), to classify such things. In my opinion, Wikidata should maintain technical and philosophical compatibility with ontologies like ChEBI and remove statements like "ethanol *instance of* chemical compound". This would improve interoperability between Wikidata and the rest of the Semantic Web.

Thanks, Eric

https://www.wikidata.org/wiki/User:Emw

1. https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2014/07#Forth_an... 2. https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2014/05#chemical... . 3. https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Chemistry#Germanium_....

4. https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2014/07#Subclass... 5. https://www.wikidata.org/wiki/Help_talk:Basic_membership_properties#Proposit... 6. http://pigsonthewing.org.uk/wikimedian-residence-royal-society-chemistry/

Hi David,

How does your treatise relate to the fact that https://www.wikidata.org/wiki/Q153 has statements that entail the following? [1]

ethanol *instance of* chemical compound *subclass of* chemical compound

How would it resolve that specific problem?

Such statements make Wikidata incompatible with ChEBI and other major ontologies, like Gene Ontology and Disease Ontology, which use *instance of* (i.e. rdf:type, P31) and *subclass of* (i.e. rdfs:subClassOf, P279, is_a) as recommended in the Relation Ontology (RO) [2] and the Basic Formal Ontology (BFO).

For Wikidata to be interoperable with other major ontologies in the Semantic Web, we cannot "scrap everything and start again from the beginning". We must stand on the shoulders of giants. Doing away with "entity" [3] as a the top of the *subclass of* hierarchy and introducing a raft of idiosyncratic ontological constructs like "identifiables", "cognizables", "negentropy" and "system survival-oriented direction" to Wikidata is probably not the way to go.

Regarding your concerns about the ability of existing approaches to represent emergent properties and natural language, I recommend perusing [4], an influential paper that informs DOLCE and BFO -- particularly the section on "The Role of Identity Criteria". I also highly recommend lectures 3 and 1 in http://ontology.buffalo.edu/smith/IntroOntology_Course.html for gaining a perspective on how BFO maintainers think about things like distinguishing objects and representations, and that upper ontology's philosophical roots. Given your interest in phenomenology and Husserl, you may also be interested in what BFO maintainers have written on those subjects in e.g. [5] with regard to ontology.

Best, Eric

[1] https://www.wikidata.org/wiki/Q153 currently states "ethanol *subclass of* alcohol", but given "alcohol *subclass of* organic compound" and "organic compound *subclass of* chemical compound", it is entailed that "ethanol *subclass of* chemical compound". [2] Barry Smith et al. (2005). *Relations in Biomedical Ontologies*. http://genomebiology.com/2005/6/5/r46 [3] "Entity" item on Wikidata. https://www.wikidata.org/wiki/Q35120. Mapped to owl:Thing. [4] Nicola Guarino (1998). *Some Ontological Principles for Designing Upper Level Lexical Resources*. http://arxiv.org/pdf/cmp-lg/9809002v1 [5] Barry Smith (1989). *Husserl: Logic and Formal Ontology*. http://ontology.buffalo.edu/smith/articles/lfo.html

Sorry Micru, I read your document, but I can't see a relation beetween ethanol, negenthropy, and Wikidata's survival :)

2014-10-05 22:48 GMT+02:00 David Cuenca dacuetu@gmail.com:

Hi Eric,

The idea is to separate in those edge cases the perceptual model (what is recognized) from the name given to it. As such you can consider "ethanol" and "chemical compound" metaclasses (names) pointing to a label-less class that represents the model.

In practical terms what it would entail is:

• remove the labels from Q153
• create one item for the label "ethanol" and one item for the label

"chemical compound"

• link Q153 with those names using "has name"
• "ethanol" is no longer an instance, but a class that can take different

names, "ethanol" being more specific

Do you realize that after this "scrap everything and start again from the beginning" the resulting structure is more comprehensive and encompasses previous efforts? "Entity" stays as it is, but now it can be examined by its qualities and they can be taken apart if needed. The term "cognizable" is 1:1 compatible with the subclass/instance model, but it emphasizes the necessity of having an observer for it to be meaningful. Classes do not exist in isolation, it is in fact very naive to keep the notion of objectivity when dealing with observation. Even logic needs a system to be executed, and based on which reality models? How were they produced? And how are those models and the models based on them verified? The problem with logicians is that they think themselves isolated from the world, however even they were born from a womb.

And have you seen who introduced the term "negentropy"? If you check the names behind it you will see that those ideas are in fact standing on the shoulders of giants. It is hard to model life without understanding first that life itself is survival-oriented.

Those papers are interesting but fail to address basic questions like: who decides identity? How does the observer interact with it? Where does information come from? And without tackling those questions, and by extension, emergent processes, logic seems like a deux ex machina that appears from nothingness and acts in nothingness.

Best, Micru

On Sun, Oct 5, 2014 at 6:49 PM, Emw emw.wiki@gmail.com wrote:

...

Hi David,

How does your treatise relate to the fact that https://www.wikidata.org/wiki/Q153 has statements that entail the following? [1]

ethanol *instance of* chemical compound *subclass of* chemical compound

How would it resolve that specific problem?

Such statements make Wikidata incompatible with ChEBI and other major ontologies, like Gene Ontology and Disease Ontology, which use *instance of* (i.e. rdf:type, P31) and *subclass of* (i.e. rdfs:subClassOf, P279, is_a) as recommended in the Relation Ontology (RO) [2] and the Basic Formal Ontology (BFO).

For Wikidata to be interoperable with other major ontologies in the Semantic Web, we cannot "scrap everything and start again from the beginning". We must stand on the shoulders of giants. Doing away with "entity" [3] as a the top of the *subclass of* hierarchy and introducing a raft of idiosyncratic ontological constructs like "identifiables", "cognizables", "negentropy" and "system survival-oriented direction" to Wikidata is probably not the way to go.

Regarding your concerns about the ability of existing approaches to represent emergent properties and natural language, I recommend perusing [4], an influential paper that informs DOLCE and BFO -- particularly the section on "The Role of Identity Criteria". I also highly recommend lectures 3 and 1 in http://ontology.buffalo.edu/smith/IntroOntology_Course.html for gaining a perspective on how BFO maintainers think about things like distinguishing objects and representations, and that upper ontology's philosophical roots. Given your interest in phenomenology and Husserl, you may also be interested in what BFO maintainers have written on those subjects in e.g. [5] with regard to ontology.

Best, Eric

[1] https://www.wikidata.org/wiki/Q153 currently states "ethanol *subclass of* alcohol", but given "alcohol *subclass of* organic compound" and "organic compound *subclass of* chemical compound", it is entailed that "ethanol *subclass of* chemical compound". [2] Barry Smith et al. (2005). *Relations in Biomedical Ontologies*. http://genomebiology.com/2005/6/5/r46 [3] "Entity" item on Wikidata. https://www.wikidata.org/wiki/Q35120. Mapped to owl:Thing. [4] Nicola Guarino (1998). *Some Ontological Principles for Designing Upper Level Lexical Resources*. http://arxiv.org/pdf/cmp-lg/9809002v1 [5] Barry Smith (1989). *Husserl: Logic and Formal Ontology*. http://ontology.buffalo.edu/smith/articles/lfo.html

-- Etiamsi omnes, ego non

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l

I have removed the statement *"instance of* chemical compound" from ethanol (Q153) [1].

A few proposals have been made in this thread about how -- or whether -- to use *instance of* (i.e. rdf:type, P31) to classify 'ethanol' and other chemical compounds, but there seems to be consensus that "*instance of* chemical compound" is not the way to do it.

Summary of proposals:

1. *Do not use instance of for chemical compounds*. Such statements make Wikidata incompatible with many major scientific ontologies, like ChEBI, Gene Ontology and Disease Ontology, which use *instance of* as defined in the Relation Ontology (RO) [2]. Note that RO defines instances as particular things that have a unique location in space and time, whereas classes are universal, general entities which have particular instances. Instances and classes are thus disjoint, so RO-based ontologies cannot have entities that have both *instance of* (rdf:type, P31) and *subclass of* (rdfs:subClassOf, P279) statements as is possible in OWL 2 DL via punning.

2. *Use statements like "instance of type of chemical compound" for chemical compounds*. Doing so makes it easier to generate lists of chemical compounds, and is valid in OWL 2 DL -- it is metamodeling via punning.

Let's build consensus for how (or whether) we want to use *instance of* for chemical compounds before any mass edits to remove or replace the 14969 other "*instance of* chemical compound" claims [3] or adding statements like "*instance of *type of chemical compound" to ethanol.

Micru has a different proposal for how to model items, which incidentally does not represent "ethanol" as an instance [4]. However, that proposal is clearly a more radical vision for Wikidata, and probably warrants a separate thread for discussion.

Eric

https://www.wikidata.org/wiki/User:Emw [1] Removal of "*instance of* chemical compound" from ethanol: https://www.wikidata.org/w/index.php?title=Q153&diff=162563849&oldid... [2] Barry Smith et al. (2005). *Relations in Biomedical Ontologies*. http://genomebiology.com/2005/6/5/r46 [3] All "*instance of* chemical compound" claims on Wikidata. http://tools.wmflabs.org/wikidata-todo/autolist.html?q=claim%5B31:11173] [4] "'ethanol' is no longer an instance, but a class". https://lists.wikimedia.org/pipermail/wikidata-l/2014-October/004691.html

So it is clear, instance_of, when translated to OWL, has generally been written as the predicate rdf:type. There is no specific instance_of relation defined as a property in OWL versions.

There is still a difference between rdf:type and instance_of, which is that it is ternary temporally indexed relation as defined in BFO *Alan* instance_of* Researcher* in *2014*. Unfortunately there is no clear way to translate this into OWL. Fixing that is among the work being done for BFO2

On Wed, Oct 8, 2014 at 9:25 PM, David Cuenca dacuetu@gmail.com wrote:

I agree with using "instance of" as RO prescribes, also because it would clarify its use.

Regarding #2, what is the difference between stating "<ethanol> instance of <type of chemical compound>" or "<ethanol> type of <chemical compound>? We have some antecedents using ad-hoc typing properties, that perhaps could be merged into a more generic property: https://www.wikidata.org/w/index.php?search=p%3Atype&title=Special%3ASea...

Cheers, Micru

On Wed, Oct 8, 2014 at 3:51 AM, Emw emw.wiki@gmail.com wrote:

...

I have removed the statement *"instance of* chemical compound" from ethanol (Q153) [1].

A few proposals have been made in this thread about how -- or whether -- to use *instance of* (i.e. rdf:type, P31) to classify 'ethanol' and other chemical compounds, but there seems to be consensus that "*instance of* chemical compound" is not the way to do it.

Summary of proposals:

1. *Do not use instance of for chemical compounds*. Such statements

make Wikidata incompatible with many major scientific ontologies, like ChEBI, Gene Ontology and Disease Ontology, which use *instance of* as defined in the Relation Ontology (RO) [2]. Note that RO defines instances as particular things that have a unique location in space and time, whereas classes are universal, general entities which have particular instances. Instances and classes are thus disjoint, so RO-based ontologies cannot have entities that have both *instance of* (rdf:type, P31) and *subclass of* (rdfs:subClassOf, P279) statements as is possible in OWL 2 DL via punning.

2. *Use statements like "instance of type of chemical compound" for

chemical compounds*. Doing so makes it easier to generate lists of chemical compounds, and is valid in OWL 2 DL -- it is metamodeling via punning.

Let's build consensus for how (or whether) we want to use *instance of* for chemical compounds before any mass edits to remove or replace the 14969 other "*instance of* chemical compound" claims [3] or adding statements like "*instance of *type of chemical compound" to ethanol.

Micru has a different proposal for how to model items, which incidentally does not represent "ethanol" as an instance [4]. However, that proposal is clearly a more radical vision for Wikidata, and probably warrants a separate thread for discussion.

Eric

https://www.wikidata.org/wiki/User:Emw [1] Removal of "*instance of* chemical compound" from ethanol: https://www.wikidata.org/w/index.php?title=Q153&diff=162563849&oldid... [2] Barry Smith et al. (2005). *Relations in Biomedical Ontologies*. http://genomebiology.com/2005/6/5/r46 [3] All "*instance of* chemical compound" claims on Wikidata. http://tools.wmflabs.org/wikidata-todo/autolist.html?q=claim%5B31:11173] [4] "'ethanol' is no longer an instance, but a class". https://lists.wikimedia.org/pipermail/wikidata-l/2014-October/004691.html

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l

-- Etiamsi omnes, ego non

Fully agree with Markus' beautifully written explanation, although I am not completely convinced of the level theory - but it seems to work in the given examples.

Note that "Porsche 356" could very much be an instance of "car model" - but not of "car". All the rules that Markus has mentioned would stay intact in this case. We often don't make the difference between "car" and "car model" in our day to day speech, which is a common source of confusion (i.e. "the Porsche 356 is a beautiful car" vs "the Porsche 356 is a beautiful car model" - both would be acceptable in natural language, Hi,

I fully agree with Thomas and the other replies given here. Let me give some other views on these topics (partly overlapping with what was said before). It's important to understand these things to get the subclass of/instance of thing right -- and it would be extremely useful if we could get this right in our data :-)

What is a class and what is an item is often a matter of perspective, and it is certainly accepted in the ontology modelling community that one thing may need to be both.

The important thing is that "subclass of" is a relation between *similar* things (usually of the same type):

* "sports car" subclass of "car" * "Porsche Carrera" subclass of "sports car" * "Porsche 356" subclass of "Porsche Carrera"

Use "A subclass of B" if it makes sense to say "all A's are also B's" as in "all Porsche Carreras are sports cars".

In contrast, "instance of" is between things that are very *different* in nature:

* "Douglas Adams" instance of "human" * "human" instance of "species"

Subclass naturally forms chains, like in my example. You can leave out some part of the chain and the result is still meaningful:

* "Porsche Carrera" subclass of "car" [makes sense]

For instance of, this does not work:

* "Douglas Adams" instance of "species" [bogus]

So if you want to organise things in a hierarchy (specific to general), then you need "subclass of". If you just describe the type of one thing, then you need "instance of". It is perfectly possible that one thing participates in both types of relationships.

In addition to these general guidelines, I would say that a well-modelled ontology should be organised in "levels": whenever you use instance of, you go to a higher level; if you use subclass of, you stay on your current level. Each thing should belong to only one level. Here is an example where this is violated:

* "Porsche Carrera" subclass of "sports car" * "Porsche 356" subclass of "Porsche Carrera" * "Porsche 356" instance of "sports car"

Each of these makes sense individually, but the combination is weird. We should make up our mind if we want to treat Porsche 356 as a class (on the same level as sports car) or as an instance (on a lower level than sports car), but not do both at the same time. I think "subclass of" usually should be preferred in such a case (because if it is possible to use subclass of, then it is usually also quite likely that more specific items occur later [Porsche 356 v1 or whatever], and we really will need subclass of to build a hierarchy then).

Cheers,

Markus

On 25.09.2014 20:10, Thomas Douillard wrote:

Hi, this is a long discussion :) Is is allowed by OWL2 notion called "Punning".

The rationale is that Hydrogen is a chemical elements, and that the chemical element is not a subclass of atom. Rather a chemical elements is a type of atom, so chemical elements is a metaclass : a class of class of atoms. ​

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l

_______________________________________________ Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l