Manually prevent certain kinds of taxon changes based on rank types

Feature Requests
1

Improper taxon swaps cost identifiers and curators more time to unravel than almost anything else on the site. There are some combinations that have appeared lately that caused entirely avoidable problems. These often create excess identifier time commitments to fix, or duplicating the same IDs just to get everything back to where it was before. It’s a waste of time, and a lot of clutter on observations.

The main one being that Species1 → Complex should not be a valid option for a taxon swap. This improper version of a split prevents atlasing, and all batch functions (like “update your content”, which can only use taxa contained within that taxon swap). The correct version, of course, is Species1 → Species1, Species2, Species3 and so on, ideally with all output taxa having the parent taxon appropriate for the swap (complex in this example).

Species1 → Genus should also not be an option. I have only heard of this used as a curator breach of power when one is disgruntled with observations or IDs being of species in a genus they deem shouldn’t be identified, to force them all back to genus. Otherwise, it is only being used as a misinformed or incorrect taxon swap methodology.

There may be others as well, but these are the two most problematic ones that iNaturalist seems to allow to happen. Granted I feel like iNat needs a curator “upvote” system for taxon swaps anyway, but that’s a whole different story.

2

In the Species1>Complex case, I see no problem with it as long as the curator who downgrades stuff also has a duty to take proper care of the results, in a timely fashion (in a matter of hours, not months). In essence, it would make theoretically (rather than technically) impossible to downgrade 1600+ obs from Species to Complex, unless the curator properly organizes a “re-ID task force” beforehand.

Anyway, same old, same old. If only there weren’t outstanding feature requests aimed at alleviating curation issues… improve guidelines (RFC-like?), technical safeguards (e.g. hands-on experience of curators, automatic delay, mandatory notifications before any commit, etc.)

3

I would say the easier and more correct option for this scenario is still to do the swap “properly”, so that the curator doesn’t need the duty to take care of the results to begin with. Species1>Complex prevents the atlas from applying, which ensures every observation ends up thrown to the Complex pile. This might not matter for species that are indeed cryptic and cannot be distinguished by geography, but for many, this does matter.

1 Like
4

Perhaps not directly related but an even worse taxon ‘swap’ occurred not too long ago where a common species (50,000+ observations) was split up into a few different species, and instead of performing any sort of taxon swap at all, someone simply changed the rank of the former species to ‘complex’ and added the new species under it. It caused huge indexing problems in addition to the problems listed for Species → Complex swaps above, and I’m still fixing the problems it caused to this day.

4 Likes
5

I wholeheartedly support this request.

That said, though, I think this request addresses a more fundamental misunderstanding among some curators about what taxon changes are for. Taxon changes are only for changes in taxonomy, and not for changing or “fixing” identifications en masse.

Some changes in taxonomy will necessarily result in identification changes (particularly lumps and splits at genus or species level and below), but they should only happen when the actual taxonomy of the organisms changes, and not merely as a convenient way to “correct” or “downgrade” a batch of existing identifications. The latter would be a misuse of curatorial privileges. Everyone should be correcting misidentifications, or overly precise identifications, through the established mechanism of adding a disagreeing ID. Being a Curator does not confer any special privilege to override the identifications of others. We all still have one vote at a time per observation, one observation at a time.

This issue comes to mind because of the example you gave

I agree, and the only reason I can think of why someone would make such a swap is if they decided that observations of the species are impossible to identify at species level on iNaturalist, and that the community should only be identifying them to Complex level instead. If so, that is not a taxon change, it is a suggested identification, and should be expressed as such.

For that matter, a Complex should never be the direct output taxon of any taxon change, since Complex is not a valid taxonomic rank (outside of iNaturalist), and therefore can never be part of a change in taxonomy. It should only be an indirect output of an atlased taxon split, where the output taxa can’t be resolved geographically and their immediate shared ancestor happens to be at the iNat “rank” of Complex. A Complex in iNaturalist is only an informal “container” to which species can be added or removed, and has no meaning if it doesn’t contain at least two species-rank taxa.

More generally, I can only think of one uncommon situation where it would be appropriate to swap a species-rank taxon into any taxon of higher rank - when the species-rank taxon is “made-up” or otherwise not validly published, and has no clear correspondence to one or more other valid species or subspecies. Only then would it be valid to swap it into whichever ancestor it could confidently be assigned. But I think this situation is rare enough that it can be handled through manual invalidation of the taxon and (re-)identification of the affected observations, instead of requiring an exception to this feature request.

3 Likes
6

From a pure ID standpoint “Species1 → Species1, Species2, Species3” isn’t accurate. What happens in this case is actually Species1 s.l. → Species1 s.s., Species2, Species3. Species1 s.l. = Species1 s.s. + Species2 + Species3 = Species1 complex. So, by going from Species1 s.l. to the Species1 complex, you are not changing the ID. When you do “Species1 s.l. = Species1 s.s. + Species2 + Species3”, you are using the swap to change the ID in all cases, often based on geography. Species1 s.l. does not equal Species1 s.s, so you are changing the ID there even though you are not changing the name. I think every situation is different. Ultimately, whatever is used should result in the least amount of effort needed for identifiers to get the most accurate end IDs.

3 Likes
7

Yes, fundamentally Species1 → Complex and Species1sl → Species1ss,2,3 can be identical. In many cases, it would have no impact which version you use, because in both cases it can be best that all observations are in fact put to complex level following the change.

The impact occurs solely when an atlas is possible (species having geographic separations), wherein sending observations to complex or genus level is unnecessary for certain regions. However, instead of thinking about this case-by-case I think it should still be encouraged to do the swaps “properly”. The normal Species1sl → 1ss,2,3 swap consistently accomplishes both scenarios, as long as the output 1ss,2,3 species have the correct parent (complex in this case) prior to the swap, and as far as I know is how it is intended to be done.

1 Like
8

Because anything not atlased would go to the new complex or the next level up, right? If so, this feature request makes perfect sense to me. The one I think you are talking about seemed like there wasn’t enough known regarding the distribution of the taxa that were split to properly atlas them as they were widespread weeds. So, the end result would possibly be the same. That said, that example is one where an identifier should identify a bunch before the swap so as to possibly figure out the probable distribution on each and then they could possibly atlas them.

9

Correct, atlas conflicts (multiple species overlapping) or no atlas will cause all observations to end up at the shared parent (usually genus, or complex in this case).

In the situations that inspired this feature requests, there were decent sources for geographic distributions already, and at least in one iNat data did not provide any novel distribution info for the atlasing process.

10

It’s probably worth defining ‘decent’ for each and every case (even POWO the reference has its flaws, shortcomings, and erroneous maps; and regional identification habits die hard). If considering iNat obs and iNatters’ posts as a primary scientific source for taxonomy and geographic distribution, there’s a risk of circular reasoning. More generally, I think “ID by location” is largely frowned upon (I got chastised for it more than once), hence my surprise that “mass ID by atlas” is a thing here. Even if tedious, better re-ID based on usual criteria like morphology.

1 Like
11

If I understand you correctly, this is not actually what’s done in practice. In a couple big recent splits the IDs of Species1 were not changed despite the definition of Species1 being changed significantly:

1 Like
12

Just because using slightly more concrete (yet fictional) examples helps me understand this better, and because I don’t do much taxon editing, would this be something like swapping species Crotalus exampleus → into complex Crotalus exampleus? So there would be no more active Crotalus exampleus species taxon?

And this would be taking species Crotalus exampleus and just swapping it into genus Crotalus?

13

While definitely problematic for some taxa, location is pretty reliable for distinguishing others. Every situation is different. Some taxa definitely do not have good enough range maps. However, a good look at iNat observations, if there are a lot that can be IDed morphologically, could inform a better one to be used in a taxon swap or inform that one should not be used.

14

would this be something like swapping species Crotalus exampleus → into complex Crotalus exampleus ?

Yes, this is a good fictional substitute.

And this would be taking species Crotalus exampleus and just swapping it into genus Crotalus ?

Also yes.

However:

So there would be no more active Crotalus exampleus species taxon?

Yes. In a taxon swap the input taxon becomes inactive. This means that these two examples will both mark the species taxon inactive without a species level replacement, and so that species is just “missing” on iNat unless manually recreated.

However people will instead choose to fix this by just reviving the original species taxon. For instance you can see this in the history of [link removed by Moderator], it was marked “active” again to revive the species.
That action was taken after this taxon swap [link removed by Moderator], which marked it inactive.

That in itself is not how it is supposed to be done, and is somewhat an exploit to fix a problem that should never have happened to begin with. But that is a lesser symptom of the bigger problem here, in the topic.

15

[Per Forum policy of not calling out other users in a negative light, I removed the links in the above post. They can be shared privately with Staff if needed.]

1 Like