Life on Earth is organized into a hierarchical naming system that has been in continuous use for roughly 250 years, applied to more than two million described species, yet the framework is anything but settled. The familiar tiers most people learned in school (kingdom, phylum, class, order, family, genus, species) trace back to the work of Carl Linnaeus in the eighteenth century, and that basic scaffolding persists today. But the number of top-level groups, the rules for drawing boundaries between them, and even the question of what counts as “alive” have shifted dramatically as new tools have come along. What follows is a look at how biologists sort living things, why the system keeps changing, and where real disagreements remain.
The Linnaean Backbone
Linnaeus introduced a standardized way to name organisms using two Latin words (the genus and the species) and arranged those names into nested ranks. That system has survived because it does something genuinely useful: it signals hierarchical relationships at a glance.1The Botanical Review. The Linnaean system and its 250-year persistence When you see a species placed in a particular family, you immediately know it shares ancestry with other members of that family. The ranks themselves have been debated and tweaked over the centuries, but the core idea of binomial names anchored to physical type specimens has proven remarkably durable.
For most of that history, organisms were divided into two kingdoms: plants and animals. Fungi got lumped with plants, bacteria were an afterthought, and single-celled organisms were shoe-horned into whichever kingdom seemed least wrong. By the mid-twentieth century it was clear that two kingdoms could not capture the real diversity of life, and biologist Robert Whittaker proposed a five-kingdom scheme (animals, plants, fungi, protists, and bacteria) that better reflected evolutionary relationships.2PubMed. New concepts of kingdoms of organisms That model dominated textbooks for decades and is still the version many people remember from school.
The Three-Domain Revolution
The biggest upheaval in life classification came in the late 1970s, when Carl Woese began comparing ribosomal RNA sequences across organisms. Instead of looking at what organisms looked like or how they behaved, Woese read differences in a molecule that every living cell carries. The result was the discovery that a group of microbes previously lumped with bacteria were, in evolutionary terms, as different from bacteria as you are. These were the Archaea, and their recognition led Woese to propose three domains of life: Bacteria, Archaea, and Eukarya (the domain that includes animals, plants, fungi, and all other organisms whose cells have a nucleus).3PubMed Central. Carl Woese: Still ahead of our time
The three-domain model reshaped biology. It showed that the vast majority of life’s genetic diversity is microbial, invisible to anyone cataloguing organisms by shape alone. It also raised an uncomfortable question: if molecular sequences tell one story and physical traits tell another, which one wins? For the most part, molecules have won, and that shift has only accelerated as genome sequencing has become cheap and fast.
Two Domains or Three
Even the three-domain model is now under pressure. Phylogenetic analyses of a group of Archaea called the Asgard archaea have produced strong support for a “two-domain” tree of life, in which eukaryotes evolved from within the Archaea rather than forming a separate, equally ancient branch. One large-scale analysis found that eukaryotes group most closely with the Heimdallarchaeota and Wukongarchaeota branch inside the Asgard archaea, making Eukarya essentially a highly derived lineage of Archaea.4PubMed Central. Expanded diversity of Asgard archaea and their relationships with eukaryotes If this topology holds, the correct top-level split is Bacteria on one side and an Archaea-plus-Eukarya supergroup on the other. Not everyone is convinced the data are definitive, and the debate continues, but the trend in the literature over the past few years has leaned toward two domains.
The practical effect for how we classify life is significant. Under a two-domain model, the familiar “domain” of Eukarya would still exist as a useful label, but it would sit inside the archaeal branch rather than beside it. It is a bit like discovering that birds are dinosaurs rather than a separate group alongside dinosaurs: the label “bird” remains handy, but the evolutionary story changes.
Hidden Diversity Below the Surface
Even within the bacterial branch, the past decade has uncovered diversity that traditional methods missed entirely. Candidate Phyla Radiation (CPR) bacteria are a huge and poorly understood group characterized by extremely small cell sizes, stripped-down genomes, and limited ability to sustain themselves independently, so they generally survive as parasites or symbionts of other microbes.5PubMed Central. Widespread but Poorly Understood Bacteria: Candidate Phyla Radiation Standard laboratory techniques for surveying microbial communities, which rely on amplifying ribosomal RNA genes, miss roughly a tenth of environmental microbial sequences, with CPR bacteria and uncharacterized Archaea making up the bulk of those blind spots.6PubMed. Metagenomics uncovers gaps in amplicon-based detection of microbial diversity In other words, until metagenomics came along and allowed researchers to sequence DNA directly from environmental samples, a substantial fraction of bacterial diversity was invisible.
New cultivation techniques are starting to close that gap. Researchers using tiny droplet-based isolation methods have managed to enrich several CPR orders that had never been grown in the lab, including members of Parcubacteria and the candidate group Jorgensenbacteria.7ISME Communications. Enrichment of different taxa of the enigmatic candidate phyla radiation bacteria using a novel picolitre droplet technique These organisms challenge the assumption that every cell is a self-sufficient metabolic unit. They raise real classification headaches because their genomes are so minimal that placing them on a tree often yields unstable results.
When the Tree Becomes a Web
The metaphor of a “tree of life” implies neat branching: lineages split and never rejoin. In practice, horizontal gene transfer (HGT) means that organisms routinely swap genetic material across species and even across domains. In bacteria and archaea, HGT is so common that some researchers have described the history of life as a “web” rather than a tree.8PubMed. Evolution of genes and organisms: the tree/web of life in light of horizontal gene transfer Hybridization in plants and animals adds another thread. Even symbiosis, such as the ancient event that produced mitochondria, is a form of lineage merger that the simple tree model struggles to represent.
That said, the tree is not dead. Genome-scale comparisons suggest that even when HGT is frequent, the underlying branching history of lineages remains detectable, “much like cobwebs hanging from tree branches.”9PLoS Biology. The Cobweb of Life Revealed by Genome-Scale Estimates of Horizontal Gene Transfer For classification purposes, the tree metaphor still works as a first approximation, but anyone studying microbial evolution ignores horizontal connections at their peril.
What Counts as a Species
Below the grand domain-level debates, there is a quieter but equally consequential argument about what a “species” actually is. For sexually reproducing organisms, the traditional test is whether two populations can interbreed and produce fertile offspring. That criterion falls apart for bacteria, archaea, and many fungi, which reproduce asexually and swap genes laterally. It also struggles with plants, which hybridize freely across what anyone would call distinct species.
For prokaryotes (bacteria and archaea), classification has increasingly relied on a genomic yardstick called average nucleotide identity (ANI). Two genomes that share roughly 95 percent or more ANI are generally considered the same species; below about 83 percent, they are clearly different species. An analysis of around 90,000 prokaryotic genomes found that this boundary is surprisingly clean, with the vast majority of genome pairs falling either above 95 percent or below 83 percent, and very few sitting in between.10Nature Communications. High throughput ANI analysis of 90K prokaryotic genomes reveals clear species boundaries ANI has become a standard tool for verifying taxonomic assignments in genome databases.11PubMed Central. Using average nucleotide identity to improve taxonomic assignments in prokaryotic genomes at the NCBI The result is a more consistent and reproducible way to draw species lines in the microbial world, though it says nothing about whether two organisms look, behave, or function similarly.
Where Viruses Fit (or Don’t)
Viruses sit in an awkward position in any classification scheme. They have genetic material and evolve, but they cannot reproduce on their own. They hijack the machinery of living cells to copy themselves, which is why most biologists have historically excluded them from the tree of life. The International Committee on Taxonomy of Viruses (ICTV) maintains its own classification system, separate from the codes that govern bacteria, plants, and animals. The ICTV carefully distinguishes between a virus as a physical entity and the taxon to which it is assigned, and it does not classify below the level of species (strains and genotypes are left to individual research communities).12Journal of General Virology. Virus taxonomy and the role of the International Committee on Taxonomy of Viruses (ICTV)
The discovery of giant viruses, particularly Mimivirus and its relatives, reignited the debate about whether viruses deserve a domain of their own. These viruses are physically enormous by viral standards, visible under a light microscope, with genomes larger than those of some bacteria. One analysis of informational genes in large DNA viruses argued that they form a clade as ancient and distinct as Bacteria, Archaea, and Eukarya, suggesting a “fourth domain of life.”13PLoS ONE. Phylogenetic and Phyletic Studies of Informational Genes in Genomes Highlight Existence of a 4th Domain of Life Including Giant Viruses Others have countered that the phylogenetic trees supporting a viral fourth domain are artifacts of analytical methods, and that there is no solid evidence viruses constitute a true domain.14PubMed Central. Evolution of viruses and cells: do we need a fourth domain of life to explain the origin of eukaryotes? What is widely accepted is that these giant viruses likely derive from an ancestral cellular genome that shrank over time through reductive evolution; the nature of that ancestor remains debated.15PubMed Central. Genomics of Megavirus and the elusive fourth domain of Life
Beyond viruses, there are other biological entities with distinct genetic identities that sit outside any traditional tree of life: viroids, plasmids, and mobile genetic elements. These have been called “acytota,” a kingdom of neglected life, and they highlight just how much biological reality resists neat categorical boundaries.16PubMed. Acytota – associated kingdom of neglected life
Tools That Are Changing the Map
The way we discover and classify life has changed as much as the classification system itself. Environmental DNA (eDNA) metabarcoding allows researchers to identify multiple species from a single bulk sample of soil, water, or other environmental material without ever seeing the organisms.17PubMed. Towards next-generation biodiversity assessment using DNA metabarcoding This has been transformative for surveying ecosystems that are hard to access or where organisms are too small, rare, or cryptic to catch. A scoop of pond water can yield a species list that would take a field biologist weeks to compile by hand.18PubMed. Environmental DNA metabarcoding: Transforming how we survey animal and plant communities That said, the method requires careful optimization and standardization of protocols to achieve reliable species-level identification.19PubMed Central. Species-level biodiversity assessment using marine environmental DNA metabarcoding requires protocol optimization and standardization
Machine learning has also entered the picture. Deep learning models trained on images from camera traps and herbarium collections can now identify species from photographs with high accuracy. In one study testing automated identification of snakes, lizards, and toads from camera trap images, pretrained models reached roughly 86 to 87 percent accuracy for multiclass identification.20PubMed Central. Animal Species Recognition with Deep Convolutional Neural Networks from Ecological Camera Trap Images Consumer-facing apps like iNaturalist and Pl@ntNet rely on similar technology. These tools are not replacing taxonomists, but they are democratizing access to identification in the field and speeding up biodiversity surveys.21Methods in Ecology and Evolution. Machine learning for image based species identification
On the genomic side, building species trees from hundreds or thousands of genes at once has become standard. This approach forces biologists to reckon with the fact that different genes often tell different evolutionary stories. When gene trees disagree with each other, the culprit can be incomplete lineage sorting, hybridization, or gene duplication. Modern statistical methods for inferring species trees are designed to handle some of these complications and have proven robust even when the data include gene duplicates alongside true single-copy genes.22Systematic Biology. Species Tree Inference Methods Intended to Deal with Incomplete Lineage Sorting Are Robust to the Presence of Paralogs
Why Taxonomy Matters for Conservation
Classification is not just an intellectual exercise. The way we draw species boundaries directly affects which organisms receive legal protection and funding. When taxonomists “split” a species, dividing what was previously one species into two or more, it increases the number of lineages recognized as being at risk and can change conservation priorities.23Biological Conservation. Species splitting increases estimates of evolutionary history at risk A population that was considered part of a widespread species with a healthy global count might, after a split, become its own species with a much smaller range and much more precarious outlook.
The reverse can also be a problem. If two distinct species are lumped together, conservation attention may be spread too thin, or a genuinely threatened lineage may be masked behind healthy overall numbers. A review of real-world cases found no single consistent effect of taxonomic changes on conservation outcomes, but noted that splitting tends to increase protection. Name changes, meanwhile, have the least effect when the organism is already charismatic and well-known.24Biological Conservation. The impact of taxonomic change on conservation: Does it kill, can it save, or is it just irrelevant?
A persistent practical headache is that different databases use different names for the same organism. Harmonizing taxon names across biodiversity databases is an active area of work, with a growing ecosystem of tools and reference databases designed to reconcile names and track synonymies.25Methods in Ecology and Evolution. Harmonizing taxon names in biodiversity data: A review of tools, databases and best practices Without that harmonization, merging data from different countries, time periods, or research groups risks double-counting some species and missing others entirely.
Folk Taxonomy and Indigenous Classification
Scientific taxonomy is not the only way humans organize the living world. Indigenous communities around the globe have maintained their own classification systems for centuries, and these folk taxonomies are often remarkably congruent with the Linnaean system. A study of the Tének (Huastec Maya) in Mexico found that their naming and grouping of edible plants corresponded closely to scientific species boundaries, despite having been developed independently. The Tének system is practical and minimalistic, focused on distinctions that matter for agriculture and food, yet it captures both inter- and intraspecific diversity with impressive precision.26Journal of Ethnobiology. Folk Taxonomy of the Inter- and Intraspecific Edible Plant Diversity of Huastec Mayan Farmers in Mexico
Similarly, Chatino, Chontal, and Chinantec communities in Oaxaca, Mexico, have developed comprehensive folk taxonomies for wild mushrooms that enable effective use of fungal resources and support the transmission of traditional knowledge across generations.27PubMed Central. Folk taxonomy of wild mushrooms in communities of the indigenous groups Chatino, Chontal, and Chinantec in Oaxaca, Mexico These systems remind us that classification is, at its root, a human cognitive activity driven by the need to navigate biological diversity. The convergence between folk and scientific systems suggests that the biological world does have natural clusters that independent observers tend to recognize, even if the philosophical justifications differ.
Xenobiology and the Grammar of Life
All known life on Earth shares the same basic biochemistry: DNA or RNA as information carriers, the same set of twenty amino acids, and a nearly universal genetic code. Current classification systems assume this shared chemistry. But a new field, xenobiology, is attempting to engineer organisms that use different molecular building blocks entirely, such as non-standard amino acids or alternative genetic alphabets. One philosophical analysis frames this work as rewriting the “grammatical conditions of life” rather than merely editing biological content, and argues that it raises fundamental questions about what life is and how we should categorize it.28Frontiers in Synthetic Biology. The genetic code as an operational a priori: xenobiology and the rewriting of life’s grammar
If a synthetic organism runs on a genetic code that is fundamentally incompatible with all natural life, it would not fit on any existing branch of the tree. It would not share common ancestry with anything in Bacteria, Archaea, or Eukarya, because it was designed from scratch. Classification systems would need new top-level categories, or perhaps entirely new frameworks, to accommodate such entities. That prospect remains largely theoretical for now, but it underscores how tightly our classification of life is bound to the assumption of shared chemistry and common descent. Change the chemistry, and the whole edifice needs rethinking.
Can the Linnaean System and Phylogenetic Naming Coexist
Running alongside these biological discoveries is a long-standing procedural dispute about how names should be governed. The Linnaean system assigns organisms to ranked categories (genus, family, order, and so on), while phylogenetic nomenclature ties names directly to branches on an evolutionary tree without mandatory ranks. Some taxonomists have argued these two approaches are fundamentally incompatible; others have sought a middle ground. One set of recommendations proposed retaining the species binomial (the two-part Latin name) and the classical supergeneric ranks for practical reasons, while replacing the older concept of “type taxa” with explicit phylogenetic definitions that require groups to be monophyletic, meaning every member shares a single common ancestor not shared with outsiders.29Systematic Biology. Are the Linnean and Phylogenetic Nomenclatural Systems Combinable? Recommendations for Biological Nomenclature
In practice, most working biologists use a hybrid. They keep the Linnaean ranks because they are universally understood and embedded in databases, laws, and medical records, but they increasingly demand that named groups reflect evolutionary relationships. The tension between stability (names that do not change every time a new analysis comes out) and accuracy (names that reflect the best current understanding of evolution) is never fully resolved. It flares up whenever a major rearrangement of the tree makes a familiar name misleading, and it dies down whenever the community settles on a workable compromise. That push and pull is not a bug in the system; it is how classification stays connected to an ever-improving understanding of the living world.