Lumper or Splitter, Tag Yourself

August 18, 2022 9:22 AM

Lumping or splitting is an evergreen taxonomic tug-of-war. There is no correct method, but the choice can often have surprising practical consequences. Our brains are built to categorize; how does yours prefer to do it?
posted by thoroughburro (34 comments total)

Splitter. Then you have more birds on your list. Sometimes, without even having to go out into the field.
posted by Bee'sWing at 9:32 AM on August 18, 2022 [2 favorites]

While lumpers consistently tried to create coherent patterns, splitters preferred incoherent complexity.

There's the rub. If "incoherent complexity" is an accurate description of how one perceives reality (and it certainly describes how I perceive it) then one is a splitter.

Of course, one could also argue that "lumpers" and "splitters" are really just members of the same group (i.e., living creatures with finite lifespans) and that the difference is so minor as to be inconsequential.
posted by Faint of Butt at 9:37 AM on August 18, 2022 [2 favorites]

Lumper or splitter? Sounds like the kind of question a splitter would ask.
posted by kcalder at 9:41 AM on August 18, 2022 [29 favorites]

I am a persistent lumper of the "and what species definition do you roll with?" variety, but I acknowledge that this is at least as much about the level of joy I take from fucking with classification systems as anything else.

I'd think that splitting would be most dangerous for conservation efforts that result in very small effective population sizes rather than allowing outcrossing / infusions of new genetics from related individuals, as with the near loss of the Florida Panther.
posted by sciatrix at 9:48 AM on August 18, 2022 [1 favorite]

I like to think that I make a judgment about what degree (and kind) of classification is appropriate to the goal at hand. Taxonomy is an art.
posted by escape from the potato planet at 9:57 AM on August 18, 2022 [3 favorites]

(Working as a developer, or any other kind of IT professional, means that you get plenty of practice at designing and working with taxonomies.)
posted by escape from the potato planet at 10:00 AM on August 18, 2022 [5 favorites]

I'm a lumper. My counterpart at work is a splitter.

I'm constantly having to talk him out of for example adding more and more granular ticket relationships in JIRA, because he thinks it would be useful to know whether this ticket "causes" that ticket, or "would be helped by", or "solves", or "partially solves", and I think people just want to link ticket A to ticket B and will choose whatever random category gets them there
posted by ook at 10:16 AM on August 18, 2022 [3 favorites]

... splitters make very small units – their critics say that if they can tell two animals apart, they place them in different genera ... and if they cannot tell them apart, they place them in different species. ... Lumpers make large units – their critics say that if a carnivore is neither a dog nor a bear, they call it a cat.

Thanks, thoroughburro! Had never heard of lumpers and splitters. Love this. (BTW, I identify as a lump.)
posted by Bella Donna at 10:24 AM on August 18, 2022 [1 favorite]

Lumper or splitter? Sounds like the kind of question a splitter would ask.

Putting people into just two categories seems like excessive lumping to me.
posted by Foosnark at 10:26 AM on August 18, 2022 [1 favorite]

SPLITTERS!
posted by lalochezia at 10:28 AM on August 18, 2022 [1 favorite]

Funny, I was wondering the same thing about hypergraphs.

Insisting on discrete categories of any size (large or small) seems like the fundamental problem, because at some level any such category is just an approximation.
posted by nat at 10:44 AM on August 18, 2022 [2 favorites]

if a carnivore is neither a dog nor a bear, they call it a cat

Except bears, having non-retractable claws, are canids. If everything that's not in the order Canidae is a cat (Felidae), then everything that's not a felid is a dog.
posted by fedward at 10:47 AM on August 18, 2022 [2 favorites]

Also this is why object typing in programming tends to rely on "can it do this" tests. If it supports a specific method, then treat it as the most generic class object that supports that method unless you have a programmatic need to know more. (And yes, if you're a non-programmer thinking that this means if the object quacks you'd treat it like a duck, I'm sad to report that you, too, could suffer from the condition of being a programmer).
posted by fedward at 10:53 AM on August 18, 2022 [2 favorites]

you, too, could suffer from the condition of being a programmer

Not that! Anything but that!! What's next, I get taken to Detroit??
posted by Greg_Ace at 11:34 AM on August 18, 2022 [1 favorite]

I don’t know what I am. Mainly because I place categories into the epistemological realm as higher order abstractions, mental constructs, that have no ontological status except as mental constructs. That feathered, web footed creature, that makes a quacking sound, is a unique individual. In my head, it appears to be a duck, similar to those other things I see floating on the lake. Lumpers, I guess, work with existing abstractions and put the unique individuals they encounter into those. Splitters, on the other hand, seem to take individuals and then begin abstracting them into new abstractions. Reality for me is infinitely complex, the perceived incoherence is once more just in your head. Lumpers seem to be trying to simplify the complexity but the simplification moves them farther from reality. Splitters appear to be trying to stay closer to reality and even mentally increase the complexity. The key for me is to know that all categories, though quite useful and maybe necessary, are only mental constructs and should not be identified as existing out there. They are useful tools. You need to be able to swing both ways as you try to comprehend your surroundings.
posted by njohnson23 at 11:37 AM on August 18, 2022 [6 favorites]

Lumpers and splitters are opposing factions in any discipline that has to place individual examples into rigorously defined categories.

well thank fuck I don't have to do that
posted by chavenet at 11:48 AM on August 18, 2022 [5 favorites]

In the 1970s or so there was a faction trying to split Canada Goose into 17 species. It's since been split into 2, with the new one being the smallest previously-subspecies Cackling Goose.

The 17 species faction was all about morphology and size. That's bunk science. Do it with genetics, and back it up with some evidence from the field that the "split" species isn't freely interbreeding and producing fertile offspring.

This approach works mostly okay with vertebrate animals. Outside of that, especially in the plant world, don't know no idea nature is super weird and doesn't want to be categorized.
posted by mcstayinskool at 11:50 AM on August 18, 2022 [1 favorite]

I just wish they'd settle down and solve everything, once and for all, so I don't need to keep going "Oh yeah that's a Lutra canadensis... no, wait, it's Lontra now, isn't it?" I mean I am willing to address them by the correct names but whereas I once felt smug for knowing some scientific names I now feel stupid when someone says "Actually it's Pantherophis guttatus now" which is not only harder to say but doesn't scan with the Lion King song and I wish they'd just
posted by The otter lady at 11:54 AM on August 18, 2022 [1 favorite]

Cackling Goose would be a badass username, FWIW
posted by Windopaene at 11:55 AM on August 18, 2022 [1 favorite]

reminds me of the old jim barksdale quote: “in business, there are two ways to make money. you can bundle, or you can unbundle.”
posted by bruceo at 12:17 PM on August 18, 2022

Insisting on discrete categories of any size (large or small) seems like the fundamental problem, because at some level any such category is just an approximation.

Let's not re-litigate "is a hot dog a taco".
posted by nickmark at 1:01 PM on August 18, 2022 [1 favorite]

If one considers Pluto a planet, does that mean one is a lumper or a splitter?
posted by nickmark at 1:02 PM on August 18, 2022

I have no position on Pluto.

Is that because it's too small to stand on?
posted by Greg_Ace at 1:28 PM on August 18, 2022 [3 favorites]

I often think that, given perfect information, the ideally explicative boundaries between taxa would also describe a Voronoi diagram made from points for each taxon.

But what about the got-dang intergeneric hybrids? What am I supposed to do with those little bastards, introducing hideous merge points into my cladogram?
posted by agentofselection at 2:29 PM on August 18, 2022 [1 favorite]

escape from the potato planet, you said it.
posted by rrrrrrrrrt at 2:29 PM on August 18, 2022 [1 favorite]

Neuroanatomy is a good topic for finding out who's a lumper and who's a splitter. If you ask two different neuroanatomists how to partition the brain, you'll get at least three answers. Even things that are sometimes taught in introductory psychology courses aren't universally agreed upon. Does the cerebral cortex have six layers, as is commonly taught to psych students? Or three? Or eleven? Different researchers may give different answers depending on how lumpy or splitty they are, or just depending on what's most convenient for a particular research question. Is the striatum really divided into the caudate, putamen, and nucleus accumbens, or is it just one continuous nucleus with a gradient of function that happens to be perforated by white matter in certain places in humans? Or is it actually best conceptualized as two intercalated nuclei, due to the striosomes? The most extreme answers to this question probably don't have too many strong defenders, but many neuroscientists would argue that the conventional lines along which we cleave the striatum are fuzzy at best.

I think biology in general tends to reward the ability to be kind of flexible in how much of a lumper or splitter you are. Few of the boundaries we impose on the natural world, whether they're taxonomic, anatomical, or otherwise, are completely unambiguous, and because these categorizations are just tools we use to understand the world, it's valuable to be able to modify those tools as necessary. That's one of the real advantages of modern phylogenetics, I think: because it's fundamentally hierarchical, you can flexibly choose your level of lumpiness/splittiness depending on your need.

I also think the Pluto / dwarf planet controversy is a good example of how astronomers could have benefited from talking to biologists. The question of how to categorize Pluto and other similar bodies is essentially a taxonomic one, and instead of getting caught up in the question of whether Pluto is "really" a planet, they could have learned something from biologists about how to build a more productive taxonomic system that allows celestial objects to be categorized more flexibly depending on how finely-grained one needs to be for a particular research question.
posted by biogeo at 5:41 PM on August 18, 2022 [3 favorites]

Oh yeah this is the whole microlabel vs queer thing. As a teenager I was a splitter for sure, as an adult I'm a lumper. For identity stuff, at least. My bookshelf is an entirely different matter.
posted by brook horse at 7:38 PM on August 18, 2022 [1 favorite]

Ah, for identity stuff I started as much more of a splitter but have come comfortably to live in lumper-land, for sure.

I think biology in general tends to reward the ability to be kind of flexible in how much of a lumper or splitter you are. Few of the boundaries we impose on the natural world, whether they're taxonomic, anatomical, or otherwise, are completely unambiguous, and because these categorizations are just tools we use to understand the world, it's valuable to be able to modify those tools as necessary. That's one of the real advantages of modern phylogenetics, I think: because it's fundamentally hierarchical, you can flexibly choose your level of lumpiness/splittiness depending on your need.

There are a lot of deeply enjoyable and thoughtful field specific pieces in this paragraph but I fully admit that I keep snagging on the concept of the OTU as the base classification unit being extended into anything I personally work on or have to keep track of. No! Anything but that! Please, I have canine and feline dependents! Have mercy! OTUs drive me a little nuts explicitly because I care a lot about drawing my boundaries in terms of natural categories rather than artificially imposed ones. OTUs as a framework sidestep the question of whether any given categories reflect natural variation or not: they cut the Gordian Knot by explicitly containing no theory of the relationships of any particular sample to any other. I find that having shorthand for conceptualizing the expected relationships between samples is useful for orienting myself, even if the categories aren't perfect.

Hell, you can quite easily split into groupings more granular than even individuals are: as witness taxonomies of the genomes of cells taken from an individual with cancerous tissue, a uterus that has carried a pregnancy, or somatic mutations in various types of tissue. The natural distributions of bounded variation are always fuzzier than I think we'd all like them to be.
posted by sciatrix at 9:27 PM on August 18, 2022 [1 favorite]

Progress vs. Categories[*]

As humans we like to put things into categories. It makes communicating and thinking easier. Scratch that. It makes communicating and thinking possible. Categories go hand in hand with words as providing us with crucial compression of reality. Just like a 1:1 map is completely useless (it is the terrain itself), so would be a need to describe every detail of every person or object before being able to make a point. We would never get anywhere...

People and institutions who derive their power from controlling one of the existing categories will fight progress that might undo the importance of the category. For example, in academia the chairs of existing departments resist the creation of “inter-disciplinary” degrees (a coinage that by itself is meant to appease the existing disciplines)...

That’s why change will be hard and take a long time. Categories everywhere persist far past their usefulness because of vested interests. Anyone interested in inventing the future instead of clinging to the past will need to overcome that. The work of establishing new categories is hard and success in doing so a huge accomplishment that moves all of us forward.

posted by kliuless at 12:48 AM on August 19, 2022

So what do we call people who reject the concept of categorization entirely?

well thank fuck I don't have to do that

Insisting on discrete categories of any size (large or small) seems like the fundamental problem

There are certainly situations where overapplication (or misapplication) of taxonomy can become dysfunctional.

But your computer, and the entire internet, would immediately melt into goo, if it weren't for the thousands of different either-or taxonomies that its constituent parts are based upon. The entire concept of a digital computer, or a computer network, becomes incoherent without those binary hard edges.

In the digital realm, the taxonomy isn't a messy approximation of reality. The taxonomy is the reality.
posted by escape from the potato planet at 3:32 AM on August 19, 2022 [2 favorites]

There are 10 kinds of programmers...
posted by neuron at 2:13 PM on August 19, 2022

As I get older the taxonomic grouping of "Whatchamacallit" grows and grows. "Thingamabob" comes in a close second.

I'm a splitter when it comes to plant collecting because it provides greater precision and ability to know what you are getting will look like but I really don't like being too fussy conversationally so it really depends on the context and domains.

Splitter. Then you have more birds on your list. Sometimes, without even having to go out into the field.

Go beyond taxonomy and get into morphs, molts, sexes and bird ages! Ordinary people see gulls but a really serious twitcher sees up to 4 different age groupings per flock of gulls!
posted by srboisvert at 5:09 AM on August 20, 2022 [1 favorite]

We have a tendency not just to categorize, but to reverse the relationships between the categories and the world, seeing categorizations as properties of the thing, when they are only useful mental tools. You get silly arguments like "is a hotdog a sandwich" because people feel like they are arguing about properties of a hotdog, but they are only arguing about properties of their personal categorizations. The granularity of a categorization which is useful will necessarily be somewhat individual - as it depends heavily on domain knowledge, interest, utility, and individual differences in how we think. I suppose that last one could be called "lumper or splitter" but I am not sure it is a useful abstraction.
posted by Nothing at 11:37 AM on August 22, 2022

It reminds of of a cookbook my grandfather once read me a chapter from. The Supper of the Lamb, which I think I have quoted here before:

But back to the onion itself. As nearly as possible now, try to look at it as if you had never seen an onion before. Try, in other words, to meet it on its own terms, not to dictate yours to it. You are convinced, of course, that you know what an onion is. You think perhaps that it is a brownish yellow vegetable, basically spherical in shape, composed of fundamentally similar layers. All such prejudices should be abandoned. It is what it is, and your work here is to find it out.

For a start, therefore, notice that your onion has two ends: a lower, now marked only by the blackish gray spot from which the root filaments descended into the earth; and an upper, which terminates (unless your onions are over the hill, or have begun to sprout because you store them under a leaky sink trap) in a withered peak of onion paper. Note once again what you have discovered: an onion is not a sphere in repose. It is a linear thing, a bloom of vectors thrusting upward from base to tip. Stand your onion, therefore, root end down upon the board and see it as the paradigm of life that it is—as one member of the vast living, gravity-defying troop that, across the face of the earth, moves light-and airward as long as the world

posted by Nothing at 11:42 AM on August 22, 2022 [2 favorites]