Wikidata talk:WikiProject Taxonomy/Archive/2014/05

This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

Qualifier nomenclatural status (P1135) with nomen illegitimum (Q1093954)

We created nomenclatural status (P1135) to mark some nomenclatorical issues including a nomen illegitimum (Q1093954). Brya has reverted this in 'Bridgesia' [Backeb. (1934)] (Q15710816) multiple times. Eriosyce villosa (Q15316109) is linked to 'Bridgesia' [Backeb. (1934)] (Q15710816). So what should we do? --Succu (talk) 06:39, 3 May 2014 (UTC)

To me twice is not "multiple times". As I pointed out twice, Bridgesia is not a validly published name (it does not exist), and is not a taxon name. It certainly is not an illegitimate name. Bridgesia polyraphis is not a taxon name either; it is not possible to make a combination with a non-existing generic name. BTW: if Bridgesia were a nomen illegitimum then Bridgesia polyraphis would not be available as a taxon name either. - Brya (talk) 06:53, 3 May 2014 (UTC)

@Brya: Please pardon me as I know nothing to the field, but the name must come from somewhere. What about NPOV ? strict application of the taxonomic rules can imply a loss on more relaxed point of views. You must find a solution and be constructive about that, how to state a controversy without ignoring that ? TomT0m ^(talk) 11:26, 3 May 2014 (UTC)

Obviously, NPoV is very important, but this means that each position should be represented accurately. I did not propose to delete anything: there is not much point including a not-validly-published Bridgesia in Wikidata, but it does not hurt either, provided it is represented accurately. It would be a stretch to put it in "common name", but that would be less bad than putting it in "taxon name".

        It is not a perfect world so I guess we should allow clade names in "taxon name", even although this gives all kinds of constraint problems. Perhaps even provisional names ("Cereus spec. A"), although this is going to cause all kinds of problems (as would formae speciales), but not known pieces of junk. - Brya (talk) 06:48, 4 May 2014 (UTC)

Usage of taxon name (P225)

Nowhere is stated that taxon name (P225) is restricted to valid (=according to rules) names. In my opinion it is simply a placeholder for any name. I think we should clarify this. --Succu (talk) 07:27, 3 May 2014 (UTC)

Well, if you can put in any bit of junk in "taxon name", what exactly is the point of having a property "taxon name"? How is the reader going to find out what name is to be used? Not by looking in "taxon name". - Brya (talk) 07:33, 3 May 2014 (UTC)

This is no junk, it is cited information. And yes, in each and every publication out there do errors exist. Nonetheless, I think no user here is adding a publication as reference to Wikidata if he/she thinks this publication bears no interesting information. If among the valuable information, the authors (!, not the Wikidatian) made an error, we should keep it and add a qualifier why it is wrong. I think this is similar to [sic!] in quotations.  — Felix Reimann (talk) 13:41, 4 May 2014 (UTC)

Cited information? In the case of Bridgesia there is cited information (ING, IPNI, Tropicos) that it is not a validly published name, in other words, that it is "junk". As a scientific name, Bridgesia is not wrong, it does not exist at all. The fact that it is junk may be "valuable information" that we should store, but only if we have a place where we can do that without any risk of Bridgesia being confused with a scientific name.

        FelixReimann, you do not seems to be internally consistent. On the one hand you repeatedly claim that the aim is to help the user find the correct name for a taxon, while on the other hand you resist developing structures to include information on things that are wrong and "why it is wrong". You just want to dump every bit of garbage you can find into "taxon name" thus giving the user no chance at all of finding out what the correct name for a taxon is. - Brya (talk) 05:30, 5 May 2014 (UTC)

So IPNI, Tropicos provide "junk"? So it's better not to explain (via qualifier) why this name is not vailid an to supress this information? So it's better to wait until someone links Eriosyce villosa (Q15316109) to Bridgesia (Q2714076). --Succu (talk) 06:26, 5 May 2014 (UTC)

No, IPNI, Tropicos say it is junk. So you want to put Bridgesia in "taxon name" and then add a qualifier "not-a-taxon-name"? What is next? Put the Eifeltower into "ship" and then add the qualifier "not-a-ship"?

        I am dubious if Eriosyce villosa (Q15316109) is notable enough for inclusion (an invented taxon name, as it where), but if it is included it should be clearly marked as not being a taxon name and not having a parent taxon. - Brya (talk) 16:53, 5 May 2014 (UTC)

Brya, please calm down. All I want is to clarify if we should restrict taxon name (P225) to valid names (=published according to the rules). You have a strict opinion which says yes, I have another one. Your wording "junk" is not helpful. Both databases say, this name exists. It was intended to name a genus belonging to the familiy Cactaceae, but is not valid (The whole story is a little bit more complicate). So how can we provide this information? BTW: only PhyloCode (Q1189395) knows the term taxon name. --Succu (talk) 18:47, 5 May 2014 (UTC)

It is true that we are dealing with some grey areas. "Taxon name" is not strictly defined, and it is not easy to strictly define it, the more so since the term is mostly made-up. What it is intended first and foremost, is for "taxon name" to hold the correct name of a taxon (that is, correct according to the provided reference). As far as I am concerned this may also be a clade name, especially if this is widely used.

        I don't see a reason to put anything in taxon name other than the correct name of a taxon, again: correct according to the provided reference (this automatically requires it is Code-compliant). Anything else will just hopelessly confuse the readers. There may be a lot of other information that Wikidata could provide, but this should be done elsewhere, with separate properties (see the proposed properties) or separate qualifiers. Getting this organized may not going to be easy, but it is the only way forward that offers hope of providing real information (just scraping together everything that can be found and dumping it in "taxon name" is not helping anybody).

        And no, it is not true that "[b]oth databases say, this name exists." All three databases say, this is not a name: it does not exist. It cannot be used as a scientific name, nor does it affect scientific names. It has the same status as, say, "John Travolta". The technical term for it is a "designation". - Brya (talk) 05:44, 6 May 2014 (UTC)

Sure they express this. Otherwise they won't record this „name”. But they have a different (internal) database name for this fact and an annotation thats says „not vailidy published”. --Succu (talk) 18:58, 7 May 2014 (UTC)

I am not sure what you are saying here. They record such things to reduce confusion: WARNING, DO NOT MISTAKE THIS FOR A SCIENTIFIC NAME! - Brya (talk) 05:46, 8 May 2014 (UTC)

So do we. We add it to P225, set it to deprecated (=DO NOT MISTAKE THIS FOR A SCIENTIFIC NAME) and even add a qualifier why this name is deprecated. I totally comply with Succu here.  — Felix Reimann (talk) 08:41, 8 May 2014 (UTC)

No, you propose to do just the opposite. First you propose to add it to P225 (without the least need for it this could be placed anywhere), promoting it to a scientific name, then you propose to add a deprecated flag that the vast majority of the readers and all of the sofware deployed so far are not going to see "Haha! Fooled you. Made you think this is a scientific name, but not only may this not be used for as a scientific name for any taxon, but this is not a scientific name at all. It is just a bit of junk we found somewhere".

        Besides that if you propose to put all this junk into "taxon name", you are going to need a new property for the correct scientific name. Otherwise you can forget about the reader being able to find the name to use in the swamp that P225 is going to be. - Brya (talk) 16:36, 8 May 2014 (UTC)

clade

I see that Succu is deleting the rank of clade. That is deleting highly valable information. Also, I see that the Constraint violations/P105-page is reporting "clade" as an error. This does not make sense whatsoever. I do not see where this could be adjusted. The Property talk:P105 refers to a "latest list", which goes to a 208.80.153.172 which is off the web. In today's systematics "clade" is an important part of any hierarchy, perhaps as important as species. If all the world uses clades in hierarchies why is Wikidata going in the opposite direction? - Brya (talk) 05:41, 5 May 2014 (UTC)

I'm not deleting this information. I'm simply setting no value. I found a lot of invented clade names too. --Succu (talk) 06:29, 5 May 2014 (UTC) PS: I fixed the link. --Succu (talk) 16:29, 5 May 2014 (UTC)

Well, first it held the information "taxon rank = clade" and now it gives "taxon rank = no value"; this is as classic an example of deleting information as I can think of. What new property are you going to introduce to repair this?

        I don't know what you mean by "invented clade names", but certainly a clade name should be referenced, not made up by a Wikipedian.

        PS. the link now goes somewhere, but it does not appear to be to anyplace useful? - Brya (talk) 16:47, 5 May 2014 (UTC)

• On my question "what was the purpose of it" Succu answered:

  taxon rank (P105) = no value means „without rank”. Some explanation you'll find here. --Succu (talk) 12:53, 5 May 2014 (UTC)

  So continuing, If I understand correctly: clade is always monophyletic, some taxa are not, so they are not clades. But some taxa still can be clades, and even if they have not usual rank. Why can't we say that Panpulmonata (Q870378), for example, is taxon and clade simultaneously? And if so what it's rank? --Infovarius (talk) 14:47, 5 May 2014 (UTC)

Well, clades can be taxa, or really pretty much are taxa. A taxon is a taxonomic group, no matter if it is given a rank or not, and no matter if it is given a name or not. To a taxon it is irrelevant if it is a clade or not. It is a matter of perspective; if a taxonomist think a group belongs together, then it is a taxon. A different taxonomist may disagree, and to him it is not a taxon. If a cladistic analysis points to a group as being monophyletic then it is a clade. A different cladistic analysis may disagree, and from that perspective it is not a clade. It is all in the eye of a beholder. For both a taxon and a clade it is always important to include the viewpoint (a reference or bunch of references).

        The bit of propaganda on the en-wiki that Succu pointed to is not very accurate. It is not absolutely necessary to accept clade as a rank, but it goes with accepting a clade as a "taxon name"; if clade names were put in a new property "clade name" then it is not important to assign them a rank. But if clade names are "taxon names" then their rank must be "clade". - Brya (talk) 16:35, 5 May 2014 (UTC)

So, if I'm right, <clades> are subsets of taxons ? It's enough then to mark the corresponding clades as

⟨ animals ⟩ instance of (P31) ⟨ clade ⟩

and

⟨ clade ⟩ subclass of (P279) ⟨ taxon ⟩

then. And you don't have a rank problem. TomT0m ^(talk) 17:35, 5 May 2014 (UTC)

No, both "clade" and "taxon" are each the result of a way of thinking / looking / working. Both a "clade" and a "taxon" is a set of organisms. A given clade may be the same set as a given taxon, or one may be a subset of the other (for example, in APG III, the order Caryophyllalles is a subset of the clade eudicots). All permutations are possible.

        The way we (and the Wikipedia's) have been working is to make taxoboxes with for components both clades and taxa (interchangeably), following what real systems in the real world are doing. This means that for the purposes of Wikidata, "clade" is a rank (clades don't have ranks, they are ranks, or are "treated as a rank"), something like the Joker in a set of cards.

        If a clade is not a rank, we would have to make new properties like "clade name" and "parent clade". Strictly speaking this might be more accurate, but it would be a lot of trouble for very little gain. - Brya (talk) 05:11, 6 May 2014 (UTC)

I'm not sure I follow you, but I'll try to make a proposition:

We create a <(Biological) Organism group> class. We make both <taxon> and <clade> a subclass of it (as both taxon and clades are special kinds of biological organism groups) in claims :

⟨ organims group ⟩ subclass of (P279) ⟨ taxon ⟩

and

⟨ organims group ⟩ subclass of (P279) ⟨ clade ⟩

You mark clades as

⟨ rabbits ⟩ instance of (P31) ⟨ clade ⟩

and taxon as

⟨ reptile ⟩ instance of (P31) ⟨ taxon ⟩

. This is accurate, a taxobox can retrieve the information.

If you want to go further, we can also tags the <clade> and <reptile> as instance of <scientific classification unit>. And you have no weirdness and headeachs due to shoehorning. TomT0m ^(talk) 09:22, 6 May 2014 (UTC)

It rather sounds as if you have one answer for everything, no matter what the problem is. - Brya (talk) 17:49, 6 May 2014 (UTC)

@TomT0m: I think your first approach will do the job. So we can express Caryophyllales (Q21808) is a ranked clade (taxon rank (P105)=order (Q36602) and instance of (P31)=clade (Q713623)) with the unranked clade eudicots (Q165468) (taxon rank (P105)=novalue and instance of (P31)=clade (Q713623)) as parent. --Succu (talk) 21:00, 6 May 2014 (UTC)

"rank = order" is a property of the taxon Caryophyllales, not of a clade. Clades, as clades, do not have ranks, but are ranks ("are treated as ranks"), for the purpose of Wikidata hierarchies. - Brya (talk) 05:21, 7 May 2014 (UTC)

This approch enables us to say Caryophyllales (Q21808) are monophyletic. If you don't wish to express this (or it's not true), than simply set instance of (P31) to taxon (Q16521)). --Succu (talk) 06:28, 7 May 2014 (UTC)

If you want to replace "instance of taxon" by "instance of clade", well, I never quite understood why we have "instance of taxon" anyway. But a few points to consider:

• being monophyletic is relative, so it needs to be referenced. Also, different references may disagree on a group being monophyletic, so this can get pretty complicated.
• for items that are well-referenced anyway, this won't be adding information. If a taxon is accepted by APG III it is monophyletic (according to APG III), and everybody knows this already.
• for clades in higher ranks this is possible, but if you are going to do this universally you are going to hit theoretical boundaries.

- Brya (talk) 11:26, 7 May 2014 (UTC)

instance of (P31) statements can be sourced like the others, no problem with that. TomT0m ^(talk) 11:54, 7 May 2014 (UTC)

It's refinement, not a replacement. A machine knows nothing about APGIII, so we have to be more explicit. And you can add both statements

• ⟨ X ⟩ instance of (P31) ⟨ clade ⟩
• ⟨ X ⟩ instance of (P31) ⟨ taxon ⟩

and reference them accordingly. --Succu (talk) 14:19, 7 May 2014 (UTC)

It can be added, and if it is referenced I have no serious objections, although I don't see it as especially beneficial. - Brya (talk) 16:34, 7 May 2014 (UTC)

• @Succu:, why "clade" can't be value of "taxonomic rank"? --Infovarius (talk) 11:57, 8 May 2014 (UTC)

  @Infovarius: I don't know for Succu, but I can explain why I made this proposal. For me taxon ranks (order, class, ...) where defined a long time ago as levels in the classification tree of the species. Now taxonomy has evolved and they make less sense than they used to as the trees built with genetic engeniering techniques can have an arbitrary number of depth. And the basic unit of scientific classification, in those taxonomy, has become the clade. So I think it's better to use modern classification modeling tools as those who are defined in the semantic web and ontologies world as they both are generic and common in the Wikidata project, and can handle modern taxonomy and arbitrary classification trees without shoehorning into old rank properties that aren't that meaningful in modern science. TomT0m ^(talk) 14:27, 8 May 2014 (UTC)

I am with Infovarius here. An important function (long term) of Wikidata is to allow generating taxoboxes, and for that purpose it is handy to treat clades as taxa (putting clade names in "taxon name", in "parent taxon" and giving clades the "taxon rank" of "clade"). Strictly speaking this is not accurate, but the alternative (see above) is more complicated. - Brya (talk) 05:40, 9 May 2014 (UTC)

I beg to differ. A taxobox in nothing but a bunch of code who can treat any type of data, it can use the instance of (P31) information as easily as the rank information. So it's absolutely not a problem. I'll go further : shoehorning information in places they do not belong will get the code of the infobox more complicated and will be less robust to change. When a new concept in taxonomy will appear, you will complicate the model, and make it even harder to extract will strange rules that are here only for historical reasons. A big change like this one is the perfect occasion to get rid of the history and try to do things well. If the taxobox is not used, it's easy to write one. If the tabobox is used, it get a lot harder. TomT0m ^(talk) 12:47, 9 May 2014 (UTC)

Yes, arguably a taxobox is a pretty silly thing, but there is not much chance of convincing Wikipedians of that. Taxoboxes are here to stay. - Brya (talk) 06:34, 10 May 2014 (UTC)

@Infovarius: Simply because this statement is wrong. @Brya: the ability to create a correct taxobox is not affected. At german wikipedia we do not use the terms clade or node in our taxobox model. --Succu (talk) 15:01, 9 May 2014 (UTC)

Wrong is a big word. Strictly speaking it is also "wrong" to put a clade name in "taxon name" and a clade name in "parent taxon". It is not ideal (see above), but is not alone in that: "forma specialis" is not a rank either. The main function of "taxon rank" is to provide data that can be read for a taxobox. So, the German Wikipedia does not use "clade" in its taxobox, fine. Their choice, and they can determine that for themselves. This is not a problem: if, one day, we are going to have a working mechanism that makes taxoboxes, this mechanism can be set to read in "taxon rank = clade" and convert that to produce whatever the German Wikipedia wants to use (apparently "      " ?). However, other Wikipedia's do want to use "clade" in their taxobox; among the first few I checked I counted three Wikipedia's using clade (one would be enough). The information must be provided or, obviously, it cannot be read in by the mechanism. An additional disadvantage of "taxon rank=no value" is that it can be read in by bots to produce really silly results like this. - Brya (talk) 17:05, 9 May 2014 (UTC)

BoulaurBot was blocked because of generating garbage. The statement a clade is a rank is false. No matter what you complain. A program has simply to read out taxon rank (P105). If the value is no value it has to query instance of (P31) too. All is fine and correctly modeled. And extendeable to 'node' or whatever may be correct. BTW wé are talking about less then 300 usages. --Succu (talk) 17:46, 9 May 2014 (UTC)

That BoulaurBot was 'over-enthusiasticly copying stuff' and was therefore blocked may be so, but 'enthusiasticly copying stuff' is a very common property among Wikipedians. It may well prove in the future that this is not an isolated incident (I did not go looking for it, but this turned up in a cursory check).

        The work-around you are suggesting may be workable (although it looks very much like something TomT0m has designed), but it is pretty complicated. It is going to be harder and harder to design a working taxobox mechanism.

        The statement "a clade is a rank" may well be false (depending on your frame of reference), but then equally the statement "forma specialis is a rank" is false, and probably more such cases. If you are going to be strict about it, we also need to remove clade names from "taxon name" and "parent taxon". I am well aware that we are not dealing with many cases, which seems to me all the more reason to be pragmatic about it. - Brya (talk) 06:29, 10 May 2014 (UTC)

No, you would just have to rename taxon name to scientific name :) And for parent taxon, subclass of (P279) will do the trick really well. And taxoboxes will deal with this without any problem. And you'll stop to use a model which has so much history and shoehorning that it's unreadable. TomT0m ^(talk) 17:20, 10 May 2014 (UTC)

Add Label = taxon name (P225)

Can be useful/correct to add label = taxon name (P225) in every language (at least for languages that use Latin script (Q8229))? In detail: for every item that has taxon name (P225) and don't has label and don't has sitelink I add label=taxon name (P225), if item don't has label but has sitelink I add label=sitelink without disambiguation. --ValterVB (talk) 13:01, 10 May 2014 (UTC)

Sounds about right. Also for languages that use non-Latin scripts. - Brya (talk) 17:10, 10 May 2014 (UTC)

Some bot owners like Thieol refuse to do so, but add a description. My own bot sets labels for de. There are other bots who nurse some selected languages. Problems could arise if someone merges two items, because all labels will be overwritten. Somewhere you'll find an old voting which languages should be included. Can't remeber where. --Succu (talk) 18:07, 10 May 2014 (UTC)

Hello all. I am not against the fact to do so. The reason why i do not is because I do not know if best is to set the "popular name" or the "scientific label". if the communauty is ok, no matter for me ! :-)

A possible alternative would be to set it if no wikipedia link exists. In that case, it could be a possible alias. However, if a link was added, label would have to be changed and scientific name set as an alias. What do you think of that proposal ? The old proposal was here Automatic_labelling Regards.--Thieol (talk) 18:17, 10 May 2014 (UTC)

Yes, I said that if item has site link I use it for add label (if it don't exist). --ValterVB (talk) 19:32, 10 May 2014 (UTC)

I do not like this, as it will drastically increase the item JSON size for little benefit, which adds network traffic, which slows down bot activity, and makes gadgets freeze.
If the goal is to make it easier for systems to obtain the taxon name, why not put it into the Latin label. The Wikibase software could also smartly implicitly language fallback to Latin for all languages if the P31 is taxon.
If the problem is just autocomplete search, a better approach is for the Wikibase software to automatically recognise that the value of P225 (and others) is an alias in every language. John Vandenberg (talk) 02:40, 15 May 2014 (UTC)

@John Vandenberg: The wikibase software won't do anything in that direction itself. It aims at staying generic and won't embed any domain knowledge datas. Which means he can't include specific case in its code like if property = 225 then ... handle differently. Tools such as Reasonator or gadgets like autolabel don't have this constraint. But the Wikibase software do not do that and will nether, it's a strong requirement. TomT0m ^(talk) 11:20, 15 May 2014 (UTC)

@TomT0m:, it does not need to be domain specific. Some multi-language string properties are labels for all languages. This should become an aspect of the property. That said, it isnt necessary for my suggestion above which is limited to autocomplete results - another way to implement it is for all property values to be indexed, which is already done by mw:Extension:Wikibase Query. John Vandenberg (talk) 11:27, 15 May 2014 (UTC)

Seems not so easy to handle, as there is language with latin alphabet and other not. Typically a Chinese name is unreadable to a french, whereas a german and a mexican can more or less use the same string for a person. This mean multilanguage string must has languages fallback, which make it really close to the labels of an item. TomT0m ^(talk) 11:33, 15 May 2014 (UTC)

It would be more elegant if the software would read P225 as the default label for all languages. This would also be so for movies, songs, books, cars, ships and probably more properties. - Brya (talk) 16:35, 15 May 2014 (UTC)

Cercidocerus and Dryophthoridae/Dryophthorinae

Is Cercidocerus (Q16858739) a dup of Cercidocerus (Q14923055)?

en:Cercidocerus says it is a genus of the family Curculionidae (Q7415384) and subfamily Dryophthorinae (Q1261625).

sv:Cercidocerus says is in the family Dryophthoridae (Q15645604)

The list of species of each is very similar.

Note I have found problems with the author of the en.wp article in the past; see en:User_talk:Hectonichus#Cassida_duplicates. John Vandenberg (talk) 02:57, 15 May 2014 (UTC)

Yes, by the look of things, a difference in taxonomy. It is commonplace in taxonomy that there exist several different ways to classify a taxon. In this case, the genus is placed, respectively:

• in family DRYOPTERIDAE and superfamily CURCULIONOIDEA
• in subfamily DRYOPTERINAE and family CURCULIONIDAE

a difference in the rank assigned to taxa. The only weird thing is that both claim to be using the exact same source, so the en-wiki page would appear to be based on a copying error. However, Wikispecies uses the same taxonomy (source not apparent). Anyway, these two items can be merged without immediate problems. - Brya (talk) 05:29, 15 May 2014 (UTC)

Thanks! Would you do the honours of merging them? I dont feel competent enough for taxon with multiple classifications. John Vandenberg (talk) 05:38, 15 May 2014 (UTC)

OK. As a rule of thumb, detailed taxonomies can be ignored unless properly referenced (ITIS is not a real reference), and it is a good idea to assign taxa only as to family (nobody takes family assignations all that serious). - Brya (talk) 05:49, 15 May 2014 (UTC)

Genus change

Hello, how could we manage this situation? Relicanthus daphneae (Q16822617) and Boloceroides daphneae (Q2603657) refer to the same species. Should be merge both items? I am not really familiar with this situation so I'll leave the matter in your hands. Pamputt (talk) 09:05, 24 May 2014 (UTC)

I was about to ask exactly the same thing, if the way I did so far was the correct one. For the bot created articles I usually merged them, especially if the item here does not contain any statements yet. In some cases, the bot created an article for both of the synonyms, thus the two cannot be merged - in these cases I have used instance of (P31) set to synonym (Q1040689) (or synonym (Q7345499) for animals?), and added a said to be the same as (P460) to link to the current taxon. But now for the synonym Opuntia diademata (Q10859311) and the current taxon Tephrocactus articulatus (Q309937) both items here have lots of statements. How to merge them, should the synonym taxon information be added with rank deprecated into the current one? Or leave them separate, and link the two by some other properties? In some way a synonym is just a different name for the same thing, but can also have a different scope - one taxon made out of two. Adding all that into one item might get difficult, and not necessarily all Wikipedias would follow those taxonomists who propose the merge. Ahoerstemeier (talk) 16:57, 25 May 2014 (UTC)

Quite. Synonyms are not interchangeable names. In general merging heterotypic synonyms is very likely to cause problems, but even the merging of homotypic synonyms (like Relicanthus daphneae (Q16822617) and Boloceroides daphneae (Q2603657)) may cause problems. - Brya (talk) 17:50, 25 May 2014 (UTC)

Unfortunately we are not able to model taxonomic concepts by now. There is nothing like a „current taxon”. There are authors like Anderson/Eggli who are of the opinion to put Opuntia diademata (Q10859311) into synonymy of Tephrocactus articulatus (Q309937). Others could have a different taxonomic opinion. So allways references should be involved. BTW: there a some property proposals around this topic to evaluate. --Succu (talk) 19:13, 25 May 2014 (UTC)

Succu is quite right, there is not necessarily a right taxonomy; there may easily be two or more different (and plausible) views.

        It is also good of Succu to point out the properties that are proposed and discussed. In the context of this topic especially the "synonyms"-property is important; among other things that would help to make connections between items. Probably it would also help to have the other-side-of-the-coin property "correct name" / "accepted name" (or this could be a qualifier of "synonym"). An item could then not only have P225 but also have one or more synonyms and one or more "correct names" (properly referenced). This is how Tropicos does it. - Brya (talk) 04:32, 26 May 2014 (UTC)

 Question we got a mechanism to deprecate statements. If I understand the discussion well, synonyms are used to implement different taxonomy POV which are legitimate at some point in time, but might later be sorted out and one hypothesis may prove more plausible than others. How are in taxonomy these synonyms deprecated, and can we use the Wikidata deprecation mechanism to express this fact ? TomT0m ^(talk) 16:57, 26 May 2014 (UTC)

A main process for scientific hypotheses is that their popularity varies with the people who were taught it in school; as these people age and die, so does a hypothesis decrease in popularity. However, for taxa it is not uncommon that a hypothesis which was out of fashion for a century suddenly comes back in vogue. - Brya (talk) 05:54, 27 May 2014 (UTC)

What to do with subgenus and other sub-ranks?

Subj -- Alexander Vasenin (talk) 19:32, 29 May 2014 (UTC)

@Alex.vasenin: Mind to ask a real question? --Succu (talk) 19:56, 29 May 2014 (UTC)

@Succu: Sorry, I just thought it's self-explanatory. Some species have sub-genuses. It would be nice if Wikidata could keep that kind of data somewhere. Should we keep subgenus as separate entity and link species parent taxon to subgenus entity, or should we keep subgenus name as some kind of property of species entity and leave parent taxon linked to genus? -- Alexander Vasenin (talk) 20:17, 29 May 2014 (UTC)

We take care of them. So again: what's your problem? --Succu (talk) 20:27, 29 May 2014 (UTC)

@Succu: I just wanted to know if it's OK to link parent taxon like this: species->subgenus->genus -- Alexander Vasenin (talk) 20:51, 29 May 2014 (UTC)

Sure it is. But why do you think it is'nt? --Succu (talk) 21:05, 29 May 2014 (UTC)

I've never seen it and wanted to be sure -- Alexander Vasenin (talk) 21:18, 29 May 2014 (UTC)

Yes, this is possible, and is sometimes done. It is useful only when a taxonomic reference is given (sourced statement).- Brya (talk) 05:02, 30 May 2014 (UTC)