There is always going to be difficulty in trying to put categorical labels on continuous distributions.
I've long wondered whether abandoning the concept of "species" in favor of some continuous measure of genetic relatedness would work.
From a practical, everyday standpoint, it would obviously be unworkable. We created the idea of species because in day-to-day life it's incredibly useful and only breaks down in some pretty niche edge cases (again: from the perspective of a lay person, those edge cases are both more common and more important to the scientists working with them). But as far as scientists, and particularly phylogenists and evolutionary biologists/ecologists go, it seems like we might save, if nothing else, a ton of effort that currently goes into trying to fit the round peg of nature into the square hole of our categorical system.
That being said, phylogeny is at best only tangentially related to my own expertise so I'm sure there would be difficulties that I haven't though of.
I work on computational phylogenetics. A practical problem is that while we have decent models of many things, gene transfer, hybridization, etc., practical methods tend to focus on the simpler models. I recall that a popular reason why phylogenetic trees are AFAIK more popular than phylogenetic networks (despite networks being more powerful) is because simpler models lead to more effective algorithms that do well under the oversimplified-but-probably-good-enough models. Complicated models are much more intractable (well, simple models usually lead to NP-hard problems. Complicated models are... even more hairy).
Also theory is much easier understood on simpler models. Just like big O notation abstracts over many things, simple models are much easier to reason about. There is a time and place for considering the memory hierarchy, how floating point division is really not the same level of speed as integer addition, but inferring the canine phylogeny under some easy-to-reason-about model of evolution is much simpler (both theoretically and practically), especially if you have already gathered data for five or six years and people are getting tired of no results.
Most people use genetic relatedness when constructing phylogenies today. People in the field understand species labels are just discrete terms. Some fields like metagenomics don’t even consider the species much and group things by genus or some higher level instead because of the difficulty of defining this stuff. Not to mentioned most of what you sample using metagenomics is going to be a population pool (think someone extracting dna out of soil or stool, there’s billions of organisms you are sampling potentially in that one sample). You start to think less of “do i have this organism and what does this mean” and more “what gene sequences am I detecting out of this pool and what does this mean for the resulting proteins being released in the environment and what that does”
Yep - metagenomics/phylogenomics practitioners have learned the hard way that trying to focus on species definitions gets you nowhere (if you think defining a species in multicellular sexually reproducing eukaryotes is hard, wait till you see what bacteria do - it's basically a free-for-all; and on top of that, short read metagenetic DNA sequencing gives you a soup of DNA that is very hard to precisely partition by source).
I consider the focus on species and speciation to be a science social phenomenon more than anything else, a lot of it attributable to the prestige of being the one to "discover" the species (a.k.a. describe it in a way that sticks).
The flipside is that the science of phylogenomics is incredibly fruitful, in that as we collect more complete information about genomes and their diversity and patterns of conservation, we are able to better understand life in all of its forms. It's also very data intensive and computationally complex - some of the most sophisticated data compression, suffix tree construction, hashing, and clustering algorithms are used in this field.
An unsung group of heroes in the face of this insane complexity is the scientists at NCBI whose job it is to organize all of the taxonomic and reference genetic/genomic sequence information (they are the ones who build the GenBank, RefSeq, and NCBI Taxonomy databases). Their work is as close as it comes to the source of truth on this. While the GenBank FTP server is easily overwhelmed by modern workloads trying to ingest this data, AWS and GCP provide high throughput mirrors (https://aws.amazon.com/opendata/, s3://ncbi-blast-databases, gs://blast-db).
I wonder if the Hamming distance between the DNA sequences would be useful or even good enough. Maybe there's a threshold where over a certain distance, two organisms wouldn't be able to reproduce.
You just slap a topology on things and you get an effective type theory that gets mostly good results which allow you to elide most of the complexity. This in turn allows you to reason about more complex things at a loss of fidelity — what we call abstraction.
If you think being able to recognize a forest rather than be overwhelmed by the number of leaves on trees is useful, then you understand why humans do that.
Says the person who has uttered a sentence chock full of categories!
The "what" of something is part and parcel of intelligibility. If reality is intelligible, and we are truly rational creatures that are able to know this reality, then this would entail that reality is divided into kinds that we can recognize. What those kinds are, and which things belong to which kinds, is a separate question, indeed the question in this article as it relates to living things.
A couple years ago I worked on an internal tool for the regulatory division of an agriculture company that involved mapping endangered species boundaries. The Environmental Protection Agency requires the company review potential impact of a proposed herbicide on endangered species inhabiting areas where the herbicide might be used (or drift).
The question of "what is a species" made me go a bit insane. I really wanted a unique identifier for each "species", but kept running into edge cases like species that changed scientific names (so you'd need name + date to resolve it). Occasionally people would refer to species by common name, but a common name can resolve to dozens of different scientific names. I don't think people realize how difficult and awkward it is to build software that deals with ambiguous entities.
This is a tangent, but the US Fish and Wildlife service has a cool tool for exploring endangered species [0]. I had a few issues with the data, but was overall impressed with how easy it was to access and how deep it goes (there are gigabytes of shapefiles representing endangered species boundaries offered in zip files). Preble's meadow jumping mouse [1] had a range so complex that its shapefile was 250 MB, the complete set of shapefiles covering all species' ranges was 2.5 GB--this single species of mouse took up 10% of this! For comparison, the Topeka Shiner's range shapefile was only 147 KB.
This reminds me exactly of how difficult it is to create decent security groups.
I've had a few opportunities to build active directory from the ground up. It always starts nice and clean with Accounting, Sales, Production, etc. Then Directors, Managers, Supervisors etc.
Everything maps out nicely in the beginning but then you run into things like "well Susie is only a supervisor, but she's in account so needs access to X. And Bob is a director but shouldn't have access to Y. And Managers should only be able to access Z but only if they are in Marketing. Etc."
You end up with a bunch of custom groups and the whole idea of a big venn diagram disintegrates.
In theory everyone has a defined role on paper, but it completely falls apart when the rubber meets the road.
I've assumed for a long time that the number of species is way over-inflated, just because it sounds cooler to discover a "species" rather than some variation within a species.
I cannot recall whether this was in Oderberg's "Real Essentialism"[0] or Gilson's "From Aristotle to Darwin and Back Again"[1], but in one of the two, I believe the author cites Maritain's view that there is essentially only one species of living things (in addition to human beings, whom Maritain, of course, recognized as possessing rational faculties, faculties that by their nature transcend non-intellectual life). This would seem to avoid the various questions about speciation and how it is possible, and certainly where there exists such ostensible radical potentiality for producing such variation along a lineage over time.
And this agrees with the former book, in which Oderberg distinguishes between a phylogenetic understanding of species, and an ontological understanding. The traditional ontological definition for "human" is "rational animal". It is composed of the genus "animal" and the specific difference "rational" that differentiates the human species from among the other species in the genus "animal"[2]. Under this traditional way of defining species, rational life on other planets would, therefore, qualify as human. "The aliens" would be, in some sense, "one of us", even if they were to differ phylogenetically. This becomes quite interesting in the context of an analysis along the lines of Aristotle's "De Anima", as, for example, rationality has definite entailments when it comes to anatomy and physiology.
[2] I am using the terms "genus" and "species" in the relative, ontological, intensional sense[3], not the narrow biological sense, as more or less meaning "broader category" and "narrower subcategory", respectively. Thus, we can say that the species of triangle is a plane figure that has three straight bounding sides, where "plane figure" is the genus, while having three straight bounding sides is the specific difference.
I just pulled out my copy of Oderberg (disclaimer: I've only read about half of it, and it's the first half that doesn't discuss biology much), and can't find any mention of Maritain in the index, so it must be Gilson, which I don't have a copy of.
In any case, what Maritain says is astonishing, and I fear you've just given me another reason to be suspicious of him :-). Is he claiming that the difference between (say) a shark and a dog is merely a question of accidents?
Many of these categorization issues can be resolved with a simple "I know it when I see it" test, made famous by the Supreme Court case Jacobellis v. Ohio
The problem with that test is people tend to see things differently, and historically this test has been used to justify inhumane treatment to certain races of humans.
It goes back much further than that. GP linked to the Wikipedia article on universals. During the late middle ages there was a furious debate about whether species (taking the term in its very broad philosophical sense) could be said to exist, or whether all categorization was done by human beings applying labels. To me, it seems one of the most fundamental questions in all philosophy.
*Any group of organisms that can have sex to make viable offspring. (So bacteria don't have species?)
Doesn't change the fact that there are 49 million kangaroos in Australia and 3.5 million people in Uruguay which means if the kangaroos were to invade Uruguay each person will have to fight 14 kangaroos.
"Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that.""
Y'all should read the article. It goes into more depth than these two comments! For example, two species can have sex, AND produce viable offspring, but NOT produce fertile offspring. Like a donkey and horse producing a typically infertile mule. Horses and donkeys are not the same species.
There is the long standing joke around the Geeks and Jocks. The don't usually have sex with each other. Genetically they have even diverged with genetic attributes that you typically find in one and not the other. They are even more successful at having grandchildren if they mate their "own type" due to to certain genetic combinations that both parents will give to the kids. Are they two species?
Thats certainly a valid way to do it. It really doesn’t matter how you define species as long as you say how you’ve done the defining. Its just a human word like wall or chair. That being said depending on your research question, certain definitions of species might be better suited than others. For example if you just consider reproductive compatibility, then certain human couples with fertility issues would be considered different species because certain genetic incompatibilities mean they are unable to produce viable offspring.
Well, then you have problems like how speciation occurs. If the ability to produce offspring is your definition, then interbreeding does not exist by definition. And then you're left with the question of how species come about. Let us suppose some member of species A produces an offspring of species B where B is not A. Then for the species B to survive, there must arise another member of that species B (i.e., one that can interbreed). So you have at least three low chance events: (1) two members of the same new species being born (2) at roughly the same time (3) meeting each other and (4?) breeding. You also could not say that A and B could interbreed, and it is only some species C arising form B that cannot interbreed with A, because you've already make interbreeding impossible; if you can breed, you are the same species. So if A and B can interbreed, and B and C can interbreed, then by transitive property, so can A and C. So interbreeding, full stop, cannot be criterion, certainly not the sole criterion.
> Its just a human word like wall or chair.
But this undermines the notion of species as objectively true. Artifacts of that sort differ from natural kinds in the sense that such artifacts have no inherent tendency to be what they are, whereas natural kinds do. A chair is not inherently to be something we sit on. That is merely a telos in the mind of the craftman, and this telos is what that determined the form he would give the artifact. The affordance and suitability of such an artifact for sitting is something that a user of that artifact may recognize in the chair and take advantage of. What qualifies as a chair is at least in part subject to convention for practical reasons. But a bird is not like that. It is not a matter of perspective that a bird is "birdy" and does birdy things and realizes objective ends in birdy ways that are not merely observer relative. Sure, we can categorize animals for practical purposes, and we may do so according to various ends for practical rather than theoretical reasons (like classifying beaver as a fish, and fish as not meat, for Lent), but that sort of practical classification is merely practical and secondary to the question of what something actually is.
And we cannot avoid saying what a thing is as that is inherent to the notion that reality is intelligible. If you were to reduce "birdiness" to mere convention, then you would indeed be characterizing birds in the same manner as artifacts (this is called mechanism, and it has severe problems). But even a mechanist would still need to classify the things that go into making this object of merely conventional identity, at some level, or else be left to conclude that reality is unintelligible, though strangely conventions manage to magically escape this problem, as mental entities would remain intelligible. One who denies the intelligibility of the world would be left with the task of explaining how he can even make the distinction between his intelligible "mental world" and "the world out there" without some intelligibility in the first place.
Ideas have consequences, consequences greater than many realize.
Speciation is a human concept. Mutation leading to population differences is not. You can still understand mutation and how phenotypes emerge without binning these into discrete species. When a lot of biologists do a phylogeny today they might group their tree by sequence divergence.
Thank you. Saved me posting it as a counter example. Basically nature will nearly always be able to defeat human efforts at putting things in neat boxes.
You are, wrongly, scaling up from the potential of individuals to the potential of the group of individuals/species.
Saying that "humans" are capable of locomotion or movement using their legs implicitly refers to an expectation, and nobody with that intends to say that individuals who are confined to a wheelchair are not humans.
Saying that humans have 5 fingers on each hand, which is the expectation, does not imply that individuals born with a different number of fingers are not humans, but simply that is not the expectation.
But when biologists finds an insect type with different number of something they will call it a new species. That we don't do that for humans has to do with politics more than anything else, some believe we would discriminate against such humans if we said they weren't really the same species.
> But when biologists finds an insect type with different number of something they will call it a new species.
This comment cannot be answered properly without knowing who the "biologists" are, which insect types we are talking about, and what is going along with the different number of something.
The fact that individuals with four or six fingers are considered to be of the same species has nothing to do with politics, but with the fact that people with "4 fingers" can reproduce (this is the expectation) with individuals with "5 fingers" and "6 fingers" and produce fertile offspring.
Politics comes into play when it comes to, for example, species conservation.
Populations of the same species that have remained isolated from each other for centuries, millennia or more may accumulate due to selection processes, genetic drift or founder effect, genetic and phenotypic characteristics that do not allow individuals from the two populations to reproduce successfully. Fertilization may not occur, the progeny may show morphological abnormalities, or the offspring may not be fertile. But the above may occur in, say, 20 percent, 50 percent or 90 percent of cases, and for any of these percentages the existence of a new species, variant or subspecies can be argued.
Should the two populations be protected since they are (or are not) different species?
Then defect is no longer intelligible. And yet, we can determine if a heart is defective, because we know the function of a heart, we know how that function composes into the telos of the organism (which is what transmits the function to organ, so to speak). Thus, any modification of the heart, certainly one that we see has ill effects for the organism, is a defect.
It is clear that a human being who is sterile, or one who cannot move about, is in some sense a defective specimen, as these deficiencies impede the realization of human telos. (If you deny telos, then truly nothing is intelligible. Not only does the notion of defect no longer make sense, not only are biology and biological function impossible and anatomy made unintelligible, but even efficient causality is made unintelligible and science impossible, as the very fact that an effect follows from a cause, that is is not arbitrary, presupposes telos.)
I've long wondered whether abandoning the concept of "species" in favor of some continuous measure of genetic relatedness would work.
From a practical, everyday standpoint, it would obviously be unworkable. We created the idea of species because in day-to-day life it's incredibly useful and only breaks down in some pretty niche edge cases (again: from the perspective of a lay person, those edge cases are both more common and more important to the scientists working with them). But as far as scientists, and particularly phylogenists and evolutionary biologists/ecologists go, it seems like we might save, if nothing else, a ton of effort that currently goes into trying to fit the round peg of nature into the square hole of our categorical system.
That being said, phylogeny is at best only tangentially related to my own expertise so I'm sure there would be difficulties that I haven't though of.