technically, there is. if you know that something is evidence of a living thing, but you’re not sure about anything below that, you can just identify it as “state of matter life”. if you’re convinced there is no evidence of a living thing, than you can state that in the data quality assessment.
i disagree. the second option is the closest to the true intention of the action.
i’ll use horses and equines as an example since i’m probably most experienced with those. i often see tracks that most likely belong to equines, but really there’s no telling what kind of equine. as far as i know, there’s no noticeable different between the hoofprints of a mule and the hoofprints of a horse, so realistically, a lot of these observations probably shouldn’t be marked as horses.
so, if i wanted to be that guy, i could disagree by saying i believe it’s in the equine genus. by doing this, i’m not saying it’s not a domestic horse, i’m saying i don’t think there’s enough evidence to prove the tracks were made by a horse. they could also realistically belong to a mule, or in some areas, a zebra.
i take it less to mean “this is definitely not this species” and more to mean “there is definitely not enough material here for me to make a specific distinction”.
however, i do agree the wording is imperfect. i’m not sure how i’d change it, but it would be inaccurate to say that i don’t think tracks were made by horses… rather, i think they could’ve been made by horses, but there’s no way for me to tell definitively whether they were or were not.
1 Like
On this I agree with you. But I review all observations for my region/taxon, both Needs ID and Research grade (baring those explicitly obscured by the observer), so it doesn’t affect me if somebody uses that button. I’ll see the observation either way.
I don’t see the harm in leaving IDs that are uncertain at the Needs ID level, so I tend not to use this button. Also, it would require several extra clicks, and who has time for that?
Sorry for repeating myself a lot. The important distinction is that your intention is to say, “i don’t think there’s enough evidence to prove the tracks were made by a horse”. But what ends up under your ID is, “imareallygoodphotographer disagrees this is a horse”. So you’re actually most definitely end up saying, “it’s not a domestic horse” even though you didn’t mean to.
Logic is expressed in words, so yes, this is all about wording. But I realize that the fix for this issue may be more involved than I originally thought.
2 Likes
I have to look at it again, and again. Let me offer another take. There is a backdrop. Someone has just bumped an observation up to a higher level classification. So the software wants to know the reason of your decision. It asked " Is the evidence provided enough to confirm this is XYZ species." If you picked option 1, well it is a non-confrontational choice, considering that the original observer may want to ask why ? I shouldn’t use such logics. The reason to pick option 1 is that I simply don’t know, as due to several missing clues.
If you pick Option 2: You are sure it can’t be the species, but agree that it belongs in that Taxon/Genus/Group. Now I’m seeing some variations in interpretations by other users. You said No to that question. It is a simply a Yes or No question. You said No, so you have disagreed.
The [higher level Taxon] and the [species] in that observation may represent separate Identity. You disagree it is the [species].
I mentioned yesterday that it doesn’t matter that much. The reason is that identifier is one person. The person is a contributor, whether he or she is a novice or an expert. They are the community. Without them, the 2/3 community rule cannot be realised. There are tons of observations without verification. I’ve no problem picking a hard disagree. I do it all the time. Whether I’m right or wrong, the choice is belongs to an independent identifier. The rule of this game is that 2/3 community agreement rule. We are conditioned to be immune from alternative suggestion. ofcourse, we would second guess our intelligence when someone said something else. We don’t dwell into it too much. Let’s make science enjoyable. Mistakes are bound to happen. The mechanism of the system is it is designed to correct itself over time. Only Time will tell.
I like the way your brain works! “X disagrees that this is DEFINITELY species A” sounds closer to the intention of saying “It is species A or species B”, doesn’t it? Of course, given that a lot of people miss the illogic of the original version, the additional subtlety of “definitely” is likely to confuse even more. But is the new version equivalent to “A OR B”? Let’s see. The statement, “it is definitely species A” translates to, “it is species A AND NOT species B”. Disagreement with this statement means negation of the whole thing, in short, you’d be saying, “NOT(A AND NOT B)”. Now, I find it difficult to explain this colloquially, but if you evaluate the truth table, you’ll find that it is equivalent to saying, “it is NOT A OR it is B”. Obviously, this is still very different from the intended, “it is A OR B”. So no, it would still violate basic logic. Isn’t formal logic fun? 
Of course, none of this would be an issue if iNat would simply add something like, “X disagrees that an ID at the proposed taxon level is possible”.
I generally agree. It can be useful for newbies to get feedback and confirmation, and it tends to be important for folks who get drawn into the “social media” aspect of iNat.
But of my roughly 5,000 moth observations contributed in the past 2 1/2 years, only ~2,000 are Research Grade. And that’s despite the fact that almost all the photos are among the highest quality (if I say so myself).
One reason why I would want to see more IDs confirmed to RG is that only those observations get included in training of iNat’s computer-vision algorithm, I believe. Among my greatest iNat moments were the occasions where species for which I am top identifier went from “pending” to “included”.
Bravo! More power to you. Seriously.
I have personally looked at ~150,000 moth observations from California (not just the thumbnails) and contributed almost 30,000 moth IDs in the past 2 1/2 years. No time to stand in the yard and shake fists at the clouds. What clouds? It’s California!
Same here. Those are the other kind of great moments: when I find, and manage to identify, other folks’ observations that are firsts on iNat. Plus my own firsts, of course.
I contribute photos and help to correct mistakes. The contributions get looked at by very smart people, of course. But with very few glowing exceptions, the country’s top moth experts are just too busy to help with reviews.
Maybe it’s some kind of OCD? But other than that, issues like the present one are what iNat developers can do something about. I may be completely wrong, but at the moment I believe they wouldn’t be too hard to fix. Of course, I worry about the other stuff as well, and even more. But those larger worries mostly concern the habits of some users. I doubt that iNat will ever be able to keep them in check.
1 Like
I agree with your approach. I think you should pick option 2 ONLY if you are SURE it can’t be the species.
The issue I addressed here is that option 2 is also widely used in cases where the identifier is NOT sure it can’t be the species. To the contrary, the identifier often knows that it may well be the species, but their point is that we can’t know for sure from the photo(s). I think we need a 3rd option for this case.
1 Like
Absolutely! I have no idea how easy a sliding scale would be to implement, but (for me) it seems the perfect, and perhaps the only, solution. I can’t see any other way to resolve the thorny issue of, let’s say, an observation which is 90% certain to be a species common in a particular zone, but with a 10% possibility that it may be another extremely rare and virtually indistinguishable species. An ID with 90% certainty plus a brief explanation could remove the observation from the needs ID pile, make it available (with the due reserve) for research purposes and satisfy the observer’s curiosity.
So it’s either a problem of not enough identifiers, or it’s that the moths themselves are difficult or impossible to ID to species based on even the best photos. Based on what I hear is going on with moths here in Ontario, I suspect it’s a little of both. Neither of those problems are easy to address directly, so I guess messing around with how disagreements are implemented is one way around it.
Hmmm…in one breath you say you want people to be very careful about using the “this is as good as it gets” button, yet you still say you want to see more IDs confirmed to RG. Reading between the lines, I get the impression that your real complaint is that you don’t like your observations being bumped up to higher taxonomic levels, and the complaint about the wording/logic is a red herring. I’m sorry if I’m putting words into your mouth, but this is the impression I get.
I don’t believe it’s what you actually want, but what you seem to be asking for is a small change to the disagreement options. You seem to be asking for a 3rd option that has the same result as option 2 (changes the taxon), but logs a different reason text (as indicated above). I wouldn’t have a big problem with that personally, but I’m not sure it would be a significant improvement. Novice users will find the disagreement options even more confusing (3 options vs 2), and experienced users will have to pause for a few milliseconds to decide if they should click option 2 or 3 on any given observation. But is that what you’re really after?
1 Like
I wasn’t going to respond to this thread because I don’t kick things to higher taxa often anyway, but I really take offense to this:
I’m also amazed and quite scared to see how many contributors are perfectly fine with clear violations of logic as long as the “end result” fits their needs. There are even declarations in this thread that read something like, “yes, for me the statement ‘it’s species A or species B’ is the same as ‘it’s not species A’”. They simply refuse to accept one of the most basic and otherwise universally accepted rules of logic.
The notion that the desired end result justifies the means, irrespective of the abandonment of logic and of the falsification of an identifier’s choice along the way, is unscientific to the core. It’s the stuff that gets folks like Marc Tessier-Lavigne (former president of Stanford University) or Francesca Gino (Harvard professor who, ironically, studies dishonesty) come under scrutiny, lose their jobs, get their papers retracted.
It’s honestly so disrespectful, unfair, and really melodramatic, to accuse the other commenters in the thread, who all have written what seem to me thoughtful comments in good faith, of being willfully incapable of logic, and of being like people who literally fabricated research data for years for their own personal benefit and career advancement. Like, come on. No matter how you look at it, this and that are not even in the same universe.
The selection when you pick a higher level taxon is to either count it as a vote against the child taxa or not. There isn’t another option in terms of how the community taxon works. The fact that it displays “username disagrees this is x” that displays when you vote to, in fact, disagree with the species-level ID, whether that’s because you think it’s a different species or just can’t be determined, seems like a non-issue to me, and not an “abandonment of logic.” It’s website UI text, not a formal logical statement.
As for OP possibly having other factors going into their ID, that’s why I don’t kick to a higher taxon unless I’m very sure of myself, but I also think it’s somewhat on them to add a description or comment to that effect, because how else are identifiers to know? The vast majority of users are amateurs, and I think it’s reasonable for identifiers to assume there’s not some critical additional information that’s known but not included in the post.
Not the case in point, but I do think it’s a bit of an issue that disagreeing with child taxa in the case where CID is wrong but you don’t know specifically, still counts as a disagreement even when someone eventually adds the correct ID. You have to go back and agree - if there are enough wrong IDs, retracting will put it back at the wrong taxon - often meaning adding an ID you are not sure about, and also requiring you to keep on top of your notifications.
6 Likes
Admittedly, there is an implied non-zero uncertainty in all of our IDs. Even with very obvious IDs, there’s always a finite, non-zero possibility that a look-alike species from the opposite side of the world was accidentally (or deliberately) transported to the location under consideration. So if we had numeric confidence levels on IDs, we should never (or rarely) see 100% confidence. Perhaps many would be 95-99%. You give a simple example where it seems reasonably clear how to assign a confidence level. But where does your 90% vs 10% come from? Who can compute those probabilities for any given date/location? I would be at a loss to assign a numeric value to the probability of finding species A vs species B in a given area on a given date, even though I have a huge database of observations at my fingertips. I might be able to give you more qualitative probabilities such as likely vs unlikely vs not-in-a-million-years, but that’s about it.
What do we do with blurry photo of an individual that could be any one of 3 fairly common species? Is the IDer supposed to evaluate the habitat and compute the probability of finding each of the 3 species at that specific date/time so that they can pick the one with the highest probability? That would waste a great deal of time. Better to simply assign an ID at the genus (or higher) level and move on.
I think there is still a place for punting IDs up to higher levels. For a non-expert, it’s hard to know what to do with a (hypothetical) species level ID that has a percentage confidence level assigned to it. It provides no hint as to what the other plausible possibilities for an ID might be. Having the ID at a higher level like genus, sub-genus, or complex does narrow down the ID to a limited number of possibilities - and it gives you a qualitative feel for how much uncertainty there is that you wouldn’t get from a percentage (which would be meaningless outside of a simple artificial scenario). What would a 10% confidence level really mean in concrete terms? Saying that the ID belongs at the family level vs the genus level means something real in taxonomic terms, even if it doesn’t tell you relative probabilities between the alternative IDs.
Higher level taxa are also useful in a broader context. If a time comes in the future when some new discovery is made that can help resolve uncertainties in a problem taxon, it makes it easy to find the observations that need to be re-evaluated.
1 Like
No thanks. That is ‘unknown’. IDing as Life serves no purpose.
Actually MORE than two thirds. 2 against 1 doesn’t do it. Nor 4 against 2 …
Here’s my suggestion
Would this wording be a reasonable approach for covering, with few words, both cases (disagreement for being a different taxon OR disagreement that enough evidence is provided for making such a specific ID)?
It also lays focus on the ID process rather than on the specimen’s identity
1 Like
Hi folks, I’m going to close this topic as it’s essentially the duplicate of an existing feature request discussion, where you can see others have already proposed many of the suggestions also raised here.
Feel free to vote for and continue the discussion there: https://forum.inaturalist.org/t/change-wording-used-by-the-system-when-downgrading-an-observation-to-an-higher-level-taxa/3862/
It’s a long topic so one thing you can try is clicking the “Show top replies” button at the bottom of the first post to help summarize it.
(Also generally if you have a suggestion for a change to how iNat works in any way, you can use the search feature to see if that change has already been proposed and submit a new topic to Feature Requests if it hasn’t been.)
4 Likes