1. Introduction
Something potentially revolutionary is brewing in biological taxonomy. For hundreds of years biologists have used the Linnaean system for representing the organic world's diversity. Organisms are sorted into species, species into genera, genera into families, and so on. The Linnaean system also tells us how to name biological taxa. Recently, a handful of biologists and philosophers have called for the replacement of the Linnaean system (de Quieroz and Gauthier Reference de Queiroz and Gauthier1992, Reference de Queiroz and Gauthier1994; Ereshefsky Reference Ereshefsky1994, Reference Ereshefsky2001). They argue that the Linnaean system is theoretically outdated: it was designed with creationism and essentialism in mind, whereas now we live in a Darwinian age. Moreover, the continued use of the Linnaean system causes a number of practical problems.
The question of whether we should continue using the Linnaean system is a pressing one for biology and those fields that rely on biological classification. Nothing less than how we represent the organic world is at stake. The Linnaean system tells us what types of biological taxa exist and how we should name those taxa. Consequently, the terms and concepts of the Linnaean system frame all theoretical questions concerning groups of organisms above the level of the local population. If the Linnaean system is replaced, much of how we talk about the organic world will be different. Consider the effect on environmental policy. Environmental legislation and debates over the conservation of biodiversity are framed by the terms of the Linnaean system. Species and other types of Linnaean taxa, we are told, should be preserved. If the Linnaean ranks no longer have a theoretical basis, then the concepts used in environmental policy need to be re-evaluated.
Critics of the Linnaean system cite a number of problems with that system, everything from whether species should be given binomial names to the assumption that the organic world can be adequately captured by hierarchical classification. This paper will highlight just one problematic aspect of the Linnaean system, the Linnaean ranks. In what follows I will suggest that the Linnaean ranks should be junked and that an alternative system for representing the relations among taxa should be adopted. Two general reasons will be offered for getting rid of the Linnaean ranks. One is theoretical: the Linnaean ranks have little meaning in contemporary biology. The other is pragmatic: the continued use of the Linnaean ranks causes a number of practical problems. In the penultimate section of this paper I will speculate on the prospects of replacing the Linnaean system with an alternative system of classification.
2. Theoretical Problems
The Linnaean ranks are often divided into two types: the higher Linnaean ranks, such as genus, family, and class; and the rank of species. The theoretical problems of the Linnaean ranks divide along similar lines. Let us start with the higher Linnaean ranks.
2.1. Higher Taxa
From Linnaeus to Hennig, biologists have attempted to provide criteria for determining the ranks of higher taxa. In doing so, they have tried to describe a feature that makes all taxa of a particular rank taxa of that rank. For example, they have looked for a feature that makes all genera genera rather than species or tribes. As we shall see, the search for an adequate ranking criterion for higher taxa has come up empty-handed. Two prominent proposals were offered in the twentieth century, one by the founders of evolutionary taxonomy, Ernst Mayr and G. G. Simpson, and another by the founder of cladism, Willi Hennig.
According to Mayr (Reference Mayr1969, 223ff.) and Simpson (Reference Simpson1961, 222ff.), the rank of a higher taxon is determined by a number of factors. One is the phenotypic diversity found within a taxon. Another is the uniqueness of a taxon's adaptive zone. A third is the breadth of a taxon's adaptive zone. Generally, the higher a taxon scores on these factors, the more inclusive the taxon. The members of an order, for example, are more phenotypically varied than the members of a genus. Similarly, the difference between the adaptive zones of two families in a tribe will be greater than the difference between the adaptive zones of two genera in a family.
As any student of biological taxonomy knows, Mayr and Simpson's criteria for ranking higher taxa have been subject to much criticism. The arguments launched against them are standard ones used to criticise the taxonomic school evolutionary taxonomy. Hennig (Reference Hennig1966, 156) argues that there is no nonarbitrary method for measuring phenotypic diversity. He playfully asks “whether the morphological divergence between an earthworm and a lion is more or less than between a snail and a chimpanzee”? Measures of phenotypic diversity, he concludes, are based on “sheer prejudice.” Similar objections have been launched against the concept of ‘adaptive zone.’ Evolutionary taxonomists raise the taxon Aves to the rank of class, because birds, they argue, live in a significantly different adaptive zone than reptiles. Cladists respond that the concept of ‘adaptive zone’ is ambiguous and is applied inconsistently across phyla (Wiley Reference Wiley1981, 254). In brief, the criticism launched against Mayr and Simpson's criteria is that the concepts of phenotypic diversity and adaptive zone are too malleable to serve as measures of a taxon's rank.
Hennig ([1965] Reference Hennig1994, 272) offers an alternative way of defining the higher Linnaean ranks. He suggests that the taxa of a higher Linnaean rank should originate during a single time period. Here Hennig draws an analogy with geology. Just as geological strata fall into different geological categories depending on their time of origin, biological taxa should be assigned Linnaean ranks according to their time of origin. Classes, for example, could be defined as all and only those taxa that originate during the Late Cretaceous. Orders would be taxa that originated during a more recent time period.
Unfortunately, Hennig's suggestion for defining the higher ranks is problematic as well. Ideally, cladists would like the taxa of the same higher rank to have a similar degree of branching. However, taxa originating in the same period often have different genealogical structures (Eldredge and Cracraft Reference Eldredge and Cracraft1980, 222). For example, some taxa originating during the Late Cretaceous are quite successful and contain a number of orders and genera. But other taxa originating during the same period are monotypic and contain only a single basal taxon. According to Hennig's criterion, some classes are monotypic while other classes contain an array of subtaxa. From a phylogenetic perspective, Hennig's criterion places different types of taxa under a single rank. This result has caused cladists, including later Hennig (Reference Hennig and Pont1969), to abandon the idea of correlating the rank of a taxon with time of origin.
Stepping back, we have looked at two methods for ranking higher taxa: one offered by evolutionary taxonomists, another by the founder of cladism. Neither method is successful. The current status of the higher Linnaean ranks is the following. Higher taxa of the same rank are not comparable in any absolute sense. Families, for instance, can vary in their age, their degree of inclusiveness, their internal diversity, and the breadth of their adaptive zone. At best, calling a taxon a ‘family’ indicates that within a particular classification, that taxon is more inclusive than a genus and less inclusive than a class. Beyond that, calling a taxon a ‘family’ or an ‘order’ is an empty designation.
2.2. Species
Thus far, I have argued that the higher Linnaean ranks lack theoretical meaning. What about the other type of Linnaean rank, the rank of species? Is there a property distinctive of all species taxa? Once again, I will argue no. The taxa we call ‘species’ are non-comparable as well. Consequently, and this is a very controversial claim, the rank of species lacks a theoretical foundation.
Confusion over the distinctive property of species goes back to Darwin and earlier. In a letter to Joseph Hooker, Darwin writes:
It is really laughable to see what different ideas are prominent in various naturalists’ minds, when they speak of “species”; in some, resemblance is everything and descent of little weight—in some, resemblance seems to go for nothing, and Creation the reigning idea—in some, sterility an unfailing test, with others it is not worth a farthing. It all comes, I believe, from trying to define the indefinable. (December 24, 1856; in Darwin Reference Darwin1887, Vol. 2, 88)
So according to Darwin, there is no common and distinctive property among those taxa we call ‘species.’
Today, we still have a wealth of species concepts. Many of those concepts fall into two groups: interbreeding concepts (Mayr Reference Mayr1970; Patterson Reference Paterson and Vrba1985) and phylogenetic concepts (Mishler and Brandon Reference Mishler and Brandon1987; Ridley Reference Ridley1989). Roughly, proponents of interbreeding concepts believe that species are groups of organisms that successfully interbreed and produce fertile offspring. Proponents of phylogenetic concepts believe that species are basal monophyletic taxa. The problem that Darwin saw with the species category remains today. The interbreeding and phylogenetic approaches to species highlight different types of taxa.
Consider the processes that maintain the existence of such taxa. Interbreeding species are bound by gene flow. Some phylogenetic species are bound by interbreeding, but many are not. Those phylogenetic species not bound by interbreeding are held together by other processes, such as selection and genetic homeostasis. Examples of such phylogenetic species are monophyletic basal taxa consisting of asexual organisms and monophyletic basal taxa consisting of isolated groups of sexual organisms (Ereshefsky Reference Ereshefsky and Wilson1999, 291–292).
Alternatively, consider the patterns of interbreeding and phylogenetic species. Phylogenetic species must be monophyletic. Some interbreeding species are monophyletic, but, again, many are not. Ancestral species are a good example (Ereshefsky Reference Ereshefsky2001, 136–137). Suppose a handful of organisms forms a peripheral isolate that gives rise to a new species. The original species, the ancestral species, continues to survive as a successful interbreeding species. However, the ancestral species does not contain all and only the offspring of a common ancestor. The ancestral species is not monophyletic and fails to be a phylogenetic species, even though it is a good interbreeding species.
The conclusion to draw from this discussion is that the species category is a heterogeneous class. Some taxa we call ‘species’ are paraphyletic and form cohesive gene pools, others are monophyletic and consist of populations that do not exchange genetic material. The heterogeneity found within the species category is not limited to a few isolated cases, but is widespread (Ereshefsky Reference Ereshefsky2001, 130ff.). Many phylogenetic species fail to be interbreeding species, and many interbreeding species fail to phylogenetic species. The species category lacks a significant unifying feature and is in the same boat as the higher Linnaean ranks. The Linnaean ranks are problematic all the way down.
3. Practical Problems
I would like to switch gears now and discuss the practical problems that come from using the Linnaean ranks. Many of these problems stem from the theoretical difficulties just outlined. Consider the suggestion that the higher Linnaean ranks lack significant meaning. As we have seen, the higher Linnaean ranks have no interclassificatory meaning; they are ontologically empty designations. Nevertheless, many biologists believe that taxa assigned the same higher Linnaean rank are similar regardless of their classification. A common but incorrect assumption is that the Linnaean ranks represent real levels of genealogical inclusiveness in nature. Unfortunately such assumptions encourage debate over the ranks of taxa, even though such disagreements lack an objective basis for resolution. Accordingly, Hennig ([1969] Reference Hennig and Pont1981, xviii) suggests that the continued use of the Linnaean ranks is a source of many “unfruitful debates” in biological taxonomy. So one practical problem with the Linnaean ranks is that valuable research time is wasted assigning ranks to higher taxa when no such ranks exist in nature.
A similar practical problem occurs when we assign taxa the rank of species. All species taxa, it is often assumed, occupy a unique and common role in nature. As we have seen, however, the species category may be nothing more than a set of non-comparable taxa. If that is the case, then we have laboured under a false assumption. This mistake has practical implications. If the species category lacks a significant unifying feature, then time and energy is wasted when we argue over the correct definition of the species category. Similarly, time and energy are wasted when we argue over whether a taxon should be called a ‘subspecies,’ a ‘species,’ or a ‘genus.’ It is perhaps worth mentioning that the reality of particular taxa, such as Homo or Homo sapiens, is not being called into question. Nor is the classificatory hypothesis that Homo sapiens is part of Homo. What is being called into question is whether the designations ‘species’ and ‘genus’ add anything to those facts. If they do not, then the continued use of the Linnaean ranks wastes valuable research time.
Theoretical problems with the Linnaean ranks lead to further practical problems. These have to do with nomenclature. The current Linnaean system contains various rules of nomenclature. A centrepiece of those rules is the idea that the rank of a taxon be incorporated in its name. Species, for example, are given binomial names, such as Fundulus heteroclita. Higher taxa are assigned uninomial names, such as Homo. In addition, the names of many higher taxa are given rank-specific endings. The rank of the family Hominidae, for instance, is indicated by the suffix - idea. The rank of the tribe Hominini is indicated by the suffix - ini.
Suppose that the Linnaean ranks do indeed lack meaning. If that is the case, then we should drop the requirement that the rank of a taxon be indicated in its name. That requirement enshrines misleading and theoretically useless information in the name of a taxon. It requires that a name tells us the rank of a taxon even though no such rank exists in nature. If the higher Linnaean ranks have no basis in nature, then biologists should not be required to assign rank-specific endings to the names of taxa. Similarly, if there is no significant species/higher taxa divide in nature, then biologists should not be required to highlight that divide in the names of taxa.
One might reply that surely there is no harm in incorporating information about Linnaean ranks in taxon names even though no such ranks exist in nature. After all, we are just talking about names, and what harm can names do? One harm is that the continued use of Linnaean names furthers the misleading impression that the ranks of the Linnaean hierarchy correspond to levels in nature.
Perhaps a greater harm is that the use of Linnaean names places biologists in the awkward position of having to determine the rank of a taxon even though no Linnaean ranks exist. Suppose a biologist discovers a new taxon. She believes that it has a high degree of inclusiveness, so she must attach an appropriate suffix to its name. To determine which suffix to use, she must first determine that taxon's Linnaean rank. If the Linnaean ranks lack objective existence, then she is being asked to find a non-existent distinction before she can even name the taxon. Theoretical problems with the Linnaean ranks ramify into practical problems: non-existence distinctions must be found, and then those distinctions must be codified in the names of taxa.
Stepping back from these problems, there is a general point to be made. The Linnaean rules place an undue burden upon biologists. Time spent deciding the correct ranks of taxa would be saved if we stopped assigning Linnaean ranks. Arguments over the correct definition of the species category would be avoided if we opted out of the Linnaean system. And, taxon names would be more easily assigned if we dropped the rule that a taxon's name reflects its rank.
We have only scratched the surface of the practical problems that face the Linnaean system. There are many more, and many of these occur regardless of whether one is sceptical of the Linnaean ranks. One such problem occurs when classifications are revised, when, for example, a taxon thought to be a family is reclassified as a genus. Taxonomic revision is the norm and not the exception in biological taxonomy. Unfortunately, the Linnaean rules of nomenclature make such revisions doubly hard. Biologists must not only reclassify taxa during taxonomic revision, they must also alter the names of those taxa. The Linnaean rules of nomenclature are themselves a source of instability in classification. Drop the Linnaean requirement that a taxon's name indicates a taxon's rank and one source of instability in biological taxonomy will disappear.
The Linnaean rules lead to further problems, for example, when biologists disagree on the rank of a taxon. In such situations the Linnaean rules require that biologists assign two different names to what they agree is the very same taxon. The list of practical problems facing the Linnaean system goes on (see Ereshefsky Reference Ereshefsky2001, 221ff. for an overview). Instead of introducing more problems, I would like to ask some programmatic questions about the future of the Linnaean system.
4. The Future of the Linnaean Hierarchy
It is one thing to highlight the problems of the Linnaean hierarchy, it is quite another thing to successfully build a case for its replacement. In the remainder of this paper I will explore the possibility of replacing the Linnaean hierarchy with an alternative system of classification. A successful case for replacing the Linnaean hierarchy needs to address three questions. Are there alternatives to the Linnaean hierarchy? Are those alternatives any better than the old system? Even if there are better systems, is it practically feasible to abandon the Linnaean hierarchy and adopt an alternative? The quick answer to these questions is “yes, yes, and yes.”
Consider the first question: Are there alternatives to the Linnaean hierarchy? A number of them have been suggested in the last 40 years. In the 1960's, Hull (Reference Hull1966) and Hennig (Reference Hennig and Pont1969) offered alternative systems. More recently, de Queiroz and Gauthier (Reference de Queiroz and Gauthier1992, Reference de Queiroz and Gauthier1994) have written a series of papers articulating a phylogenetic system. And just in the past year, a handful of biologists have introduced a system of phylogenetic nomenclature (Cantino Reference Cantino2000). I've been working on a post-Linnaean system as well (Ereshefsky Reference Ereshefsky2001). So, there are lots of alternatives out there.
Turning to the next question: Are any of these systems better than the Linnaean hierarchy? A full-scale comparison of the Linnaean hierarchy to its alternatives cannot be conducted here. Let us just consider the problems of the Linnaean hierarchy highlighted in the previous section. If an alternative system is to overcome those problems, it must do two things. First, it must indicate the positions of taxa in classifications without using ranks. Second, it must divorce the function of naming taxa from the function of classifying taxa—the name of a taxon should merely be a name, it should not indicate the rank of a taxon.
All of the alternative systems cited above satisfy these two requirements. Each indicates the positions of taxa without Linnaean ranks. Instead, they indicate the position of a taxon within a classification by using either numerials or indentations. Consider the numerical system proposed by Hull (Reference Hull1966) and Hennig (Reference Hennig and Pont1969). A standard Linnaean classification looks like this:
Subclass Reptilomorpha
Infraclass Aves
Infraclass Mammalia
Division Monotremata
Division Theria
Cohort Metaheria
Cohort Eutheria.
A classification of the same taxa using positional numbers would be the following:
2.4. Reptilomorpha
2.4.1. Aves
2.4.2. Mammalia
2.4.2.1 Monotremata
2.4.2.2. Theria
2.4.2.2.1. Metaheria
2.4.2.2.2. Eutheria.
Positional numbers capture various types of information. A taxon's degree of inclusiveness is indicated by the number of digits in its positional number: the fewer the digits, the more inclusive the taxon. Positional numbers indicate which taxa are parts of more inclusive taxa. Eutheria's number shows that Eutheria is a part of the more inclusive taxon Theria, which is part of Mammalia.
Positional numbers have advantages over Linnaean ranks. Positional numbers are not part of a taxon's name, so the functions of classifying taxa and naming them are divorced. If a taxon needs to be reclassified there is no need to change that taxon's name, only its positional number needs to be altered. Positional numbers also have the advantage of not corresponding to any absolute ranks. A positional number merely indicates the placement of a taxon within a particular classification. Recall that the Linnaean ranks carry with them the misleading assumption that taxa of the same rank are comparable; they also carry the false assumption that the Linnaean ranks correspond to real divisions in nature. Positional numbers are free of such assumptions. They indicate the positions of taxa within classifications without carrying any suspect theoretical baggage.
Post-Linnaean systems avoid the problems associated with the Linnaean ranks. Do they also avoid the problems of Linnaean nomenclature? They do, and in a very simple way: when a taxon is assigned a name in a post-Linnaean system that name does not indicate the rank of the taxon, it is merely a name. Consequently, the practical problems of the Linnaean hierarchy highlighted earlier are avoided. A taxon can be named prior to knowing its classification. Biologists can assign a single name to a taxon even though they disagree on its classification. Furthermore, taxon names remain stable during taxonomic revision.
There is one last question I would like to consider: Is it practically feasible to abandon the Linnaean hierarchy and adopt an alternative? Suppose that an alternative system is more theoretically consistent with evolutionary theory than the Linnaean hierarchy. Suppose, also, that the adoption of that alternative would make the job of biologists easier. Are these reasons sufficient for switching? A common response is “no.” Some argue that the Linnaean hierarchy is so deeply entrenched in biology that there is little hope of replacing it (Jeffrey Reference Jeffrey1989).
I find this pessimism too self-fulfilling. If at the start we decide that scientific change is too hard, then there will be no change. If inconvenience were an overriding concern in science, then scientific progress would grind to a halt. Writing new textbooks is not easy, nor is telling people that the world is different than they've previously thought. Nonetheless, those are the results of scientific progress. Progress in biological taxonomy is no different: it might be inconvenient at first, but well worth the effort in the end.
Having said that, however, I would like to suggest that the adoption of a post-Linnaean system would not be grossly inconvenient. The primary change that a post-Linnaean system would bring is the elimination of the Linnaean ranks from biological classification. That modification leaves much of current classification untouched. Biological classifications are hypotheses of relations among taxa. That is the primary purpose of classification. In a post-Linnaean taxonomy, hypotheses concerning the relations among taxa would not be affected. What would change is the how we represent those relations. Instead of Linnaean ranks, positional numbers or indentations would indicate the relations among taxa.
Another change would concern the meaning of taxon names. In a post-Linnaean system, taxon names would no longer convey information concerning the Linnaean ranks of taxa. Taxon names would be considered merely names, not indicators of rank. This change would not affect the names of taxa. We would continue calling a taxon by its current name. We would just need to recognize that the various endings of taxon names no longer indicate the ranks of taxa.
One might nevertheless protest that the Linnaean hierarchy will not be replaced because its terms are too entrenched in biology and general society. The term ‘species,’ for example, is not only found in biological taxonomy, but in many textbooks, field guides, television nature programs, and environmental laws. One might argue that even though some biologists and philosophers might want to get rid of the term ‘species’ for worthy technical reasons, its occurrence is too prevalent to eliminate. The response to this sort of concern is that the switch to a post-Linnaean system does not have to be done in one fell swoop. A post-Linnaean system could be introduced in a piece-meal fashion. For example, at first we could abandon the higher Linnaean ranks but keep the species rank in place. Eventually, the species rank itself would be eliminated, but again that too can be done in steps. Perhaps the first step would be to eliminate the term ‘species’ from biological taxonomy. Later the term would be dropped from general biology texts. How this would be done does not need to be spelled out here. The point is that the switch to a post-Linnaean system of classification does not have to be all or nothing. It can come in stages.
From a practical perspective, I do not think that the switch to a post-Linnaean system is as radical as it sounds. Taxon names would remain constant, so would classifications. Just the theoretical overlay of the Linnaean hierarchy would be altered. As we have seen, the Linnaean ranks are a remnant of pre-evolutionary biology. Moreover, their continued use causes a number of practical problems: time is wasted assigning taxa meaningless and misleading ranks, and it is further wasted when taxon names must be altered during taxonomic revision. The practical thing, it seems, is to drop the Linnaean ranks altogether.
5. Conclusion
A potential revolution is brewing in biological taxonomy. The very way we represent the organic world is being challenged. This challenge has broad implications. From theoretical questions concerning the ontology of nature, to practical ones concerning how to frame environmental legislation. This is an area relatively untouched by philosophers. Yet, it is an area that would benefit from the attention of philosophers. Many central questions in the philosophy of science are present here. The question of scientific realism arises when we ask whether Linnaean ranks should be treated realistically or instrumentally. The question of how theoretical entities should be named arises when we ask whether taxon names should be defined descriptively or as rigid designators. The debate over the Linnaean hierarchy is as much philosophical as it is biological.
