search

tdwg / tcs2

The TCS 2 Task Group will turn TCS into a form in which it can be maintained. The new version of TCS will be a vocabulary standard like Darwin Core and Audiovisual Core and will complement these other existing TDWG standards.
6 stars 0 forks source link

class:TaxonRelationship #43

Open nielsklazenga opened 3 years ago

nielsklazenga commented 3 years ago

TaxonConceptRelationship (property)
Label Taxon Concept Relationship
Definition Assertion of a relationship between two taxa
Usage notes
Comments
Required
Repeatable
Constraints

Mapping
TCS DataSet/TaxonConcepts/TaxonConcept/TaxonRelationships/TaxonRelationship | DataSet/TaxonRelationshipAssertions/TaxonRelationshipAssertion
TDWG Ontology http://rs.tdwg.org/ontology/voc/TaxonConcept#Relationship
Darwin Core —
nielsklazenga commented 2 years ago

We had a long discussion about the label for this class in https://github.com/tdwg/tnc/issues/48, but I do not think the last thing has been said yet.

At the moment, TaxonRelationship mainly seems too broad to me, as we want it to contain only the RCC-5-like relationships, so I think it needs to be qualified somehow.

camwebb commented 2 years ago

If we changed it to TaxonConceptRelationship, then, for me at least, it more naturally excludes many of the other kinds of relationship originally in TCS (child, anamorph, parent...). I'd rather not use RCC5 in the name, since RCC5 is a specific (spatial) definition of general set relationships. Using the characters Set somewhere in the name would work for me too.

deepreef commented 2 years ago

I know it's more letters, but I would prefer TaxonCircumscriptionRelationship. The word "Concept" is so overloaded in taxonomy, with different meanings and interpretations and assumptions, etc.

nielsklazenga commented 2 years ago

I am wondering if we can come up with a term that qualifies the type of relationship, rather than the things the relationship is between.

deepreef commented 2 years ago

I am wondering if we can come up with a term that qualifies the type of relationship, rather than the things the relationship is between.

I'm not sure I follow. Do you mean something like SetRelationship or something? Are you just trying to come up with a replacement to TaxonRelationshipAssertion of TCS 1?

nielsklazenga commented 2 years ago

This is the TaxonRelationshipAssertion/TaxonRelationship element of TCS 1. We cut back the relationship types that will be covered in this object to only the set relationships, so it would be good to reflect that in the label. I am thinking especially about separating them from the hierarchical relationships. So yes, something like 'SetRelationship', only I do not like that particular term. But it is good to get the options out there, so we can choose the best one when it comes to it.

I am not thinking about removing the Taxon bit, but the objects the relationship is between has not changed, so I do not think we have to change that bit. It is the Relationship bit that has changed, as that will be much narrower than it is in TCS 1.

jliljeblad commented 2 years ago

Maybe TaxonOverlap is more explicit on what kind of relationship this is about. In the same way, maybe the intersectscould be replaced by simply overlapsAtLeastSome. Not pretty though, I agree.

nielsklazenga commented 2 years ago

I am sort of with @deepreef. This is a comparison of circumscriptions. Only thing is that this introduces a new term and I would not want to define circumscription, especially since I think that in specimenCircumscription and characterCircumscription it is used in a (slightly) different meaning than it would here. You would get discussions about whether a characterCircumscription can be congruent with a specimenCircumscription, or whether a circumscription based on morphological characters is disjoint with one based on molecular characters (or that a circumscription based on one character is disjoint with one based on another character in general). There might be an alternative for 'circumscription' that means approximately the same thing and does not have the same problems.

nielsklazenga commented 2 years ago

I think @jliljeblad is onto something as well. Region Connection Calculus is about connectedness. This does not make much sense for taxa (and also RCC-5 is not about connectedness anymore), so we could make it about intersection. So, maybe not the exact word 'overlap', but it is in the right direction. Maybe 'TaxonIntersection'? Then we can leave the controlled terms as they are.

I still would like something in between 'Taxon' and 'Intersection', like 'Circumscription' (but again, maybe not that exact word, unless we define it) and I think I would like the 'Assertion' back at the end.

Just throwing this out there in the hope the perfect term will reveal itself at some point.

nielsklazenga commented 2 years ago

From a taxonomic perspective, it might also make sense to make it about 'congruence'. I think I might like 'TaxonConceptCongruence'. If taxon concepts are not congruent, there is some sort of conflict and the other terms: has proper part, is proper part of and partially overlaps describe the nature of the conflict. intersects indicates uncertainty (it could be congruent, or it could not be) and is disjoint with is like 'not applicable'.

ghwhitbread commented 2 years ago

I will suggest TaxonAlignment: Given that these relationships are about comparing circumscriptions, which, in all but instances of congruence, are about the relative delimitation of taxa (plural) in different taxonomic treatments. An unambiguous label, previously used by @nfranz.

nielsklazenga commented 2 years ago

I like 'Alignment' too, as it implies an opinion, so we do not have to consider adding 'Assertion' to the term. I think I'd like TaxonConceptAlignment better than TaxonAlignment.

deepreef commented 2 years ago

I also like 'Alignment', but I will again caution against using "Concept" in any of these terms. That word just carries too much baggage (i.e., it means too many different things to different people). Although it's more cumbersome, I think "Circumscription" is the word we want to stick with. If TaxonCircumscriptionAlignment is too many characters, then I would go with TaxonAlignment over TaxonConceptAlignment.

WUlate commented 2 years ago

What's the problem with many characters if we are able to get the right idea perfectly through? (Look at parentNameUsageID or acceptedNameUsageID in DwC for example). If TaxonCircumscriptionAlignment is too long, I prefer to take TaxonConceptAlignment and let's explain what needs to be explained.

deepreef commented 2 years ago

Just to be clear, I don't have any problem with the number of characters, but I really, strongly think we should avoid the word "Concept" as much as possible in any of these terms.

nielsklazenga commented 2 years ago

It is not just 'Concept', but 'TaxonConcept', which we define. I do not think the meaning is ambiguous, just that the word 'concept' might be used incorrectly, but taxonomists have been talking about 'taxon concepts' long before TCS, so I think we should stick with it. The subject and object of these alignments are TaxonConcepts, so I think it might be more confusing if we do not use it in the label.

'Circumscription' on the other hand is not defined and is used in TCS 1 in a different sense than is proposed here.

I am fine with TaxonAlignment though.

deepreef commented 2 years ago

Taxon Concept = circumscribed set of organisms Taxon Concept = particular treatment, including both synonymy and classification Taxon Concept = "Species Concept" (BSC, PSC, etc.) Taxon Concept = concept of what a "taxon" is

The word "circumscription" has been defined in biodiversity informatics literature and is often used in TDWG meetings and such, and has only one meaning* (the first one in the list above, which is what we want). I've lost countless minutes/hours at informatics discussions (Including the formulation of TCS 1) because people who thought they disagreed with each other turned out to be operating under different assumed meanings of "Taxon Concept". The only term that leads to even more confusion, misunderstanding, etc. in this space is "Taxon Name" -- but that's a rant for another issue.

Often it's best to leverage a term that people are familiar with, to facilitate a shared understanding of the meaning. But sometimes a term has multiple meanings within a community, leading to incorrect presumptions of its meaning.

TCS 1 includes the qualified* terms "SpecimenCircumscription" and "CharacterCircumscription". The unqualified term "circumscription" has been consistently used more broadly (and in the documentation of TCS 1) in the sense of "complete set of organisms comprising this taxon".

Maybe the easiest path forward is just to go with TaxonAlignment -- which several of us seem happy with. Does anyone object to that term?

jgerbracht commented 2 years ago

Even though I really want to use TaxonConcept, and still do when conversing with folks. I also have spent hours in various conversations only to realize part way through that I'm speaking of a stable, circumscribed set of organisms and the other person is thinking of the usages of a taxonomic name, which is variable depending on the authority. The only issue I have with TaxonCircumscription is that it could also mean the specific circumscription text and not the taxon or set of organisms being described. I can go either way as long as we very clearly define what we mean. Jeff

Jeff Gerbracht Lead Application Developer Birds of the World Cornell Lab of Ornithology 607-254-2117

From: Richard L. Pyle @.> Sent: Tuesday, November 23, 2021 9:14 PM To: tdwg/tcs2 @.> Cc: Subscribed @.***> Subject: Re: [tdwg/tcs2] class:TaxonRelationship (#43)

Taxon Concept = circumscribed set of organisms Taxon Concept = particular treatment, including both synonymy and classification Taxon Concept = "Species Concept" (BSC, PSC, etc.) Taxon Concept = concept of what a "taxon" is

The word "circumscription" has been defined in biodiversity informatics literature and is often used in TDWG meetings and such, and has only one meaning* (the first one in the list above, which is what we want). I've lost countless minutes/hours at informatics discussions (Including the formulation of TCS 1) because people who thought they disagreed with each other turned out to be operating under different assumed meanings of "Taxon Concept". The only term that leads to even more confusion, misunderstanding, etc. in this space is "Taxon Name" -- but that's a rant for another issue.

Often it's best to leverage a term that people are familiar with, to facilitate a shared understanding of the meaning. But sometimes a term has multiple meanings within a community, leading to incorrect presumptions of its meaning.

*TCS 1 includes the qualified terms "SpecimenCircumscription" and "CharacterCircumscription". The unqualified term "circumscription" has been consistently used more broadly (and in the documentation of TCS 1) in the sense of "complete set of organisms comprising this taxon".

Maybe the easiest path forward is just to go with TaxonAlignment -- which several of us seem happy with. Does anyone object to that term?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/tdwg/tcs2/issues/43#issuecomment-977401309, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAO4SSJKDYHN46MB6H7SPKTUNRDBNANCNFSM446PI2SA. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

nielsklazenga commented 2 years ago

Taxon Concept = circumscribed set of organisms Taxon Concept = particular treatment, including both synonymy and classification Taxon Concept = "Species Concept" (BSC, PSC, etc.) Taxon Concept = concept of what a "taxon" is

I would say a 'circumscribed set of organisms' is a Taxon. The treatment is not the same thing as the Taxon Concept in it, so the second and fourth are the same. Just because a species is a taxon does not mean that a taxon concept is the same as a species concept. And again, we are defining TaxonConcept, so possible different meanings are irrelevant.

Defining 'Circumscription' as the set of organisms it is a circumscription for is circular to me.

From @jgerbracht 's post:

The only issue I have with TaxonCircumscription is that it could also mean the specific circumscription text and not the taxon or set of organisms being described.

That is what a circumscription is, certainly in TCS, although not necessarily text, but a set of characters or representative specimens. Giving it the extra/different meaning of a shared idea of what the delimitation of a group of organisms is makes that TaxonCircumscription has all the problems TaxonConcept is purported to have.

nielsklazenga commented 2 years ago

The Taxon is the set of organisms, which is very much in accordance with the definition in Darwin Core, and since it is the sets of organisms we are trying to align – and also because it seems to be something we can agree on – it makes sense to go with TaxonAlignment (so @ghwhitbread had it right from the start).

ghwhitbread commented 2 years ago

I was also considering 'TaxonomicAlignment'.

nielsklazenga commented 2 years ago

Maybe you should quit while you are ahead :grin:. I have a problem with taxonomic, because it is the adjective for taxonomy – not taxon – which is much broader. Even within our broader discipline (outside systematics) it can be used for vegetation communities, soil types, geological layers and periods etc. and it is broadly used outside our discipline as well.

ghwhitbread commented 2 years ago

@nielsklazenga along with many others (and why the original choice), but this class exists for the alignment of taxonomies, no?

nielsklazenga commented 2 years ago

@ghwhitbread yes, for the particular kind of taxonomies in which the elements are taxa.

deepreef commented 2 years ago

OK, well it's now more clear than ever that we all have different interpretations of these words. And if our little group has such a diversity of interpretations, then the spectrum of interpretations among the broader community is likely to be even wider. (A common joke is that "a room with 10 taxonomists/informaticians will include at least 12 definitions of [taxon/concept/name/etc.]".)

Maybe it's best to avoid both "concept" and "circumscription".

Again... are there any objections to simply sticking with TaxonAlignment? I agree that @ghwhitbread had it right from the start.

nielsklazenga commented 2 years ago

Let's do a poll.

TaxonAlignment thumbs-up or thumbs-down (you'll have to explain in this case).

jgerbracht commented 2 years ago

TaxonAlignment is fine with me, again, as long as we unambiguously define what this means. I could see this being applied to specific TNUs as well as to the underlying concepts.

deepreef commented 2 years ago

I could see this being applied to specific TNUs as well as to the underlying concepts

As an aside, Dave Remsen, Nicolas Bailey and I have been meeting most Thursdays for an hour or two to think through the relationships of Protonyms, TNUs, and Taxon "Circumscriptions"/"Concepts"/Whatever-we-call-them. Dave and I in particular have been discussing this for the past couple of decades. We're very, very close to working out a model that we think solves it (in the sense of minting identifiers that represent meaningful circumscriptions, which can be algorithmically compared to other circumscriptions in terms of computing values of TaxonAlignment -- among other cool things). Dave mentioned it at the Catalogue of Life Global Team meeting yesterday, as we're almost to the stage where we're ready to explain it to others and open it up to broader discussion (we're only just approaching the point where we're pretty sure we understand it ourselves).

As soon as we have some documents & diagrams & sample datasets to illustrate our ideas, I'll share with this group -- as I think it has a lot of potential for informing what we're trying to achieve here. Watch this space...

nielsklazenga commented 2 years ago

Any further discussion on this is better had in #1 or the TNC repo, but, just to be clear about the terminology, the TaxonConcept we are talking about here is the TCS (1) Taxon Concept, which includes Taxonomic Name Usages (TNU) (or overlaps, only including accepted TNUs) as well as the "shared underlying" concepts like those in Avibase (which are more like "Usages" than the individual TNUs if you ask me). It is not in the purview of the TCS 2 Task Group to change that.

jliljeblad commented 2 years ago

Thumbs up on TaxonAlignment from me. (Yeah, I'm lagging behind a bit)

nielsklazenga commented 2 years ago

I would like to make the case again to name this class TaxonConceptAlignment. I think leaving out the 'Concept' because of perceived issues with the word 'concept' is getting too hung up on words. These are alignments of TCS TaxonConcept's, not Darwin Core Taxon's.

nielsklazenga commented 11 months ago

There is a proposal, in issue #225, to remove this class and have the terms in the Taxon Relationship Type Vocabulary as properties of a Taxon Concept. I think the proposed change is significant enough to not just rely on comments in GitHub issues and take silence as agreement, so we are going to have a Task Group meeting about this in a few weeks' time.

jliljeblad commented 11 months ago

Yeah, I'm not comfortable enough with classes and such to understand the impact of this proposal.