Make 'Needs ID' Independent of BOTH 'Research Grade' AND 'Casual' Statuses

I’m a taxonomy and botany neophyte, but my impression is that such is not necessarily the trend in botany. It doesn’t seem like infraspecies are slowly being chiseled away at. Some things even get lowered to subspecies after looking at the genetics. @jdmore could perhaps correct me if my perception of the general trend in botany is wrong here.

Perhaps this is where I’m feeling the disconnect. I’m virtually entirely ignorant of the world of zoological nomenclature and taxonomy. In botany, it doesn’t seem like subspecies are necessarily often-dubious or don’t really matter or that they’re more-often-then-not very difficult to tell apart (though sometimes they are!).

This conversation is starting to go in circles. The issue is not you and I agreeing that the same rigor should be applied, it’s that I am not convinced that the same rigor will be applied under the proposed system consistently by the larger community. As I’ve said, this is already a minor problem on iNat, and, in my opinion, the proposed system has the capacity to greatly exacerbate it. Others and myself have already explained why infraspecific taxa are at risk of greater misapplication than higher taxa.

I don’t know much about plants, but for animals, I feel like infraspecific taxa are added at a reasonable rate for the species that are actually determinable through iNat submissions. And so for those taxa, they should be reasonably searchable using the high rank filter. I’m not convinced the same isn’t true for plants without having more knowledge on the specifics of the random subset of plants. Just because infraspecific taxa exist doesn’t mean they can be IDed from typical iNat photos, and if they cannot, then 0 IDs to the infraspecific taxon is exactly what we should have.

3 Likes

Understood, and the same is true for subspecies in many groups of plants, although plants have a lot of pretty well-defined and recognizable subspecies too.

My point is that, to my mind, this is not a problem around which we can or should build a better identification system (the topic at hand). Any iNat identification system has to assume that the taxa available in iNaturalist, at any rank, are at least potentially identifiable using evidence in iNaturalist observations. To the extent that is not the case, then we have a problem with iNaturalist taxonomy, not with how the identification system is built.

iNaturalist taxonomy is not required to mirror outdated 1940s revisions. While it’s good that we follow global taxonomic authorities as much as possible for consistency and minimizing taxonomic disputes across the platform, we can and do deviate when sufficiently justified, and such authorities don’t exist at all for large swaths of the tree of life. In either case, the iNat community has the ultimate say-so.

In mollusks or any other group, if the situation in iNat taxonomy is that a particular set of subspecies “are a sort of placeholder” and “ideally … aren’t really used,” that to me is not sufficient reason to have the iNat identification system treat every subspecies on the site differently than species or any other taxonomic rank.

2 Likes

The link is broken. The working URL is https://doi.org/10.3389/fcosc.2022.971280

1 Like

Whoops! Thanks! I’ve fixed in my post as well.

I would expect the identifier of ssp to know the ssp first - then start from their preferred higher taxon. ‘Most’ identifiers won’t be able to ID the ssp so the current system works well.

Sometimes a taxon specialist does a sweep - with a copypasta this ssp because … that ssp because … With that info a willing identifier can help tidy up. For my iNat use, maybe a few a year, if even that many.

The example that springs to my mind is anyway at sp level. Anaxeton is laeve in the South and arborescens in the North Peninsula. Off the Cape Peninsula there are a few more sp. But sorting from sp for taxon specialists works just fine.

I did help to sort out these spiders, since it is (eye candy) the most observed one in Cape Town https://www.inaturalist.org/taxa/904618-Trichonephila-fenestrata-fenestrata

You have a tiny captive audience on the Forum. Imagine trying to explain this change to all iNatters.

There’s very little to explain though. I screwed up by beginning my explanation of this feature request at a high level, but if you look at https://forum.inaturalist.org/t/make-needs-id-independent-of-both-research-grade-and-casual-statuses/52795/29?u=regnierda, it’s far from mind bending.

When and how things go to GBIF and what observations you retrieve using the default search filters all remain the exact same. The new elements are new search filter options and an additional indicator at the top of the observation page.

You mention taxon specialists, and it feels like most comments thus far on this thread have evidenced a perception that across the entire taxonomic tree all or most subspecies are extremely difficult to ID if you’re not an expert in that specific taxon. That’s simply not true. Sure there are plenty that are dubious, or even if valid they’re still extremely difficult to tell apart, but (1) that’s not a problem specific to subspecies, and (2) there are plenty of subspecies that are not dubious and are not extremely difficult to tell apart.

I’m a generalist that focuses on the state where I live, and I actually can’t think of any subspecies offhand that were a a huge ordeal to learn to ID (except maybe Chenopodium berlandieri bushianum). While they often require smaller details to ID, details that people may be less likely to capture (because they don’t know it’s an important detail, that is until I explain it to them when I put a species-level suggestion and a comment on their observation), I’ve found the subspecies that I’ve encountered in my area to be as generally straightforward to ID as were their parent species.

I can however think of numerous species that were (or still are) extraordinarily difficult to learn, and/or ID even once learned – Stellaria spp., Polygonum spp. (not everything is P. aviculare, not everything is even in the P. aviculare complex!), and various members of Amaranthaceae, to name a few.

I continue to disagree wholeheartedly with the notion that subspecies (or variety, or form, or whatever other infraspecific ranks are out there) are vastly different than any other rank or that they should be treated any differently than species as far as iNaturalist’s “help me find stuff to ID” design goes.

The problem is that iNat doesn’t actually have a “help me find stuff to ID” mechanism. It has a “help me find things that won’t go to GBIF, or could go to GBIF, or did/will go to GBIF” mechanism, and it conflates that with whether something “Needs ID”. This proposal solves that conflation problem for both wild and non-wild observations in a way that keeps all the things in the same existing buckets. (Captive/cultivated stays over there. Wild is still right here in front of me, unpolluted by the riff-raff over there. GBIF-ability remains unchanged.)

Let me ask this of those of you that are for https://forum.inaturalist.org/t/make-captive-cultivated-not-automatically-no-id-needed/112 but aren’t a fan of my proposal here (and I’m perplexed at the number of you), what do you see as the undesirable difference(s)?

Is it primarily that this includes something that would change how subspecies might be interacted with? If that one detail got taken out, if “look at me, I need you to ID me” still stopped at species, how many of you would then be a fan?

The neat thing about this proposed design is that you don’t have to muck much with the existing engine parts. This is something that can just be built on top of the existing system. It seems like it would be less disruptive (and I’d have thought less contentious) than previous suggestions to create new categories or to redefine existing categories in the Casual → Needs ID → Research Grade flow?

It’s also flexible. We wouldn’t necessarily have to make it go all the way to subspecies (though I think it should, and I had thought it would actually help sell the idea since it’s then no longer just benefitting non-wild observations). You could also include or exclude other factors easily, whether they’re existing factors today or new factors in the future. Perhaps we think we want to ID fossils too (to be clear I’m not proposing that we do). Perhaps we think we don’t want to ID life observed from Mars. Those things could be included or excluded in this design as desired/needed.

I was actually going to request you rephrase the title to reflect it is about refining to taxa below species when they are present. If you want to modify your request substantially from what it says now, please submit a separate topic.

It’s not primarily about that though. That’s just a sub-aspect that has featured prominently in the conversation thus far. It’s really about this:

These are just the most apparent and impactful-to-many-people-or-observations ways (but not all of the ways) that this problem manifests:

  • Research Grade observations that could otherwise be ID’d further are obscured
  • Casual observations because they’re not wild that could otherwise be ID’d further are obscured

To be clear, I don’t want to modify my request, I’m just trying to draw out thoughts on what the huge perceived difference is between my proposal and the “just let us ID casual stuff already” proposal. I sense that it’s really just the details about subspecies handling, but few have really indicated clearly whether that’s their sole concern, thus my asking.

I have a row of URLs bookmarked to ID. I would need LIFETIMES to clear them all.
Perhaps you could try IDing out of your state, across the border where biodiversity is in a similar climate zone or geology? Or your preferred taxon somewhere that lacks identifiers? Or a taxon sweep thru RG obs to see if they are correct? Or annotate for phenology graphs?

Ha! I’ve no lack of things to identify either! And I wouldn’t dream of taking away all those good ID opportunities from you and the others who I know participate in coordinated ID sprees. :stuck_out_tongue:

I’m not even sure that identifying every plant observation from my own state is something I can actually achieve in my lifetime, but I certainly will try. ~11,000 down, ~260,000 left to go – and I have yet to learn much of Asteraceae, Fabaceae, Cyperaceae, or Poaceae beyond the particular species I’ve observed myself. I also am still totally ignorant when it comes to gymnosperms, ferns and friends, mosses and friends, and algae.

My Identify filter is just “all Nebraska observations of plant taxa that I know regardless of their quality grade”. (Well, that’s my main one. I also regularly check for unknowns and stuff of any kingdom at kingdom or above.) If it’s a plant in Nebraska, whether it’s casual, research grade, or needs ID, I’m going to try to review it eventually and add an ID and annotations of my own regardless of what label it has associated with it.

To be honest, with how I have my own Identify bookmarks set up, this feature request doesn’t actually change much for me either way, at least not day to day. It would be nice to be able to see, as taxonomy changes over time, what new opportunities to ID something to a more granular level arose because of the creation/resurrection of a subspecies, but it’s not something I lose sleep about.

I just believe that we should separate the question of “can someone ID this further”/“does this have scientific value”/“does this ultimately help people further engage with nature” from the question of " can this make it to GBIF / is this good enough to go to GBIF".

Interesting discussion, but we won’t be moving forward with this request, I’m sorry.

Such a change would be an enormous one (200+m observations would need to be reindexed), as taxonomy changes there would need to be more reindexing (which is already an issue with taxonomic changes), and I think it would encourage people to add finder IDs (particularly subspecies) that are already difficult and controversial (as reflected by iNat’s current treatment, where two ssp IDs are needed to move the observation taxon). Finally, a change this huge (basically updating all data quality grades and how those work throughout the site) would cause a lot of confusion without a major benefit. The data quality system is already confusing. More granularly identifying garden plants or getting observations identified beyond species are not major needs for iNaturalist.

Currently, anyone can use the filters to search for observations “stuck” at a certain taxonomic level and try to refine them (see this tutorial) I understand this doesn’t address the issue of knowing whether finer taxa are available for a certain taxon, but if someone is motivated to do so they can check it out.

6 Likes

Then we need to split
‘Casual’ into
Not Wild
and
actually Broken to be binned, until the observer resolves the missing or wrong info.

If an obs is both Broken and Not Wild, then Broken would take precedence.

That could be easily split using the DQA options that were applied. The two sets are here, but they need to be visible and usable as separate sets depending on the identifier’s preferences. If the obs is Broken, no one else needs to waste time on it.

3 Likes

I agree; while I find the original proposal intriguing, I think the biggest thing I liked about it is not the subspecies surfacing, but simply that it aims to move past the current system of equating captive/cultivated observations with “broken” observations and sweeping both of them under the algorithmic rug.

2 Likes

This topic was automatically closed after 16 hours. New replies are no longer allowed.