In the Cellinese Lab we investigate the evolution, systematics and biogeography of angiosperms (flowering plants). In addition, we are interested in semantic data integration and interoperability and are developing tools that serve the biodiversity and phyloinformatics communities (e.g., TOLKIN, BiSciCol and RegNum).

Monday, January 16, 2012

Species and Phylogenetic Nomenclature

DISCLAIMER: This post may inflame you if you are not a supporter of phylogenetic nomenclature, so my advise to you before you go on is don't ruin your day and skip reading it.  If you choose to read it, please keep in mind this is not a debate on the validity of the PhyloCode (that's a given to us). However, if you fundamentally share the philosophy at the core of phylogenetic nomenclature, then you may want to know what's going on among us proponents.

Last October, Brent Mishler (University of California, Berkeley), David Baum (University of Wisconsin, Madison) and I submitted a proposal to the International Society for Phylogenetic Nomenclature (ISPN) in order to remove all mention of "species" from the PhyloCode.  We provide justifications for this goal in a supporting paper (Cellinese et al. 2012). The proposal is currently being discussed by the ISPN Committee of Phylogenetic Nomenclature, and given our paper is also in review, this seems to be the right time to stir the pot a little more.

All proponents of phylogenetic nomenclature would agree that the PhyloCode is about naming clades, and removing ranks from nomenclature.  Yet, unfortunately there remain aspects in the current wording of the Code that are not about naming clades and that retain rank considerations.  These aspects all surround the traditional Linnaean rank of species. The PhyloCode currently states that "In this code, the terms "species" and "clade" refer to different kinds of biological entities, not ranks" (note 3.1.1.).  While the concept of species implicitly endorsed by the PhyloCode, that of de Queiroz (2007), does indeed assume that species and clade are distinct entities, this is not universally accepted by any means.  Many different species concepts are currently applied across the different biological domains (Mayden1997).  Some, like de Queiroz (2007) view species as lineages.  Others (e.g. Mishler 1999; Pleijel 1999; Mishler and Theriot 2000 a,b,c; Baum 2009) view species as ranked or unranked clades.  But this disagreement is not relevant to a Code of nomenclature that is (supposedly) rank-agnostic and concerned only with naming clades.  The debate over species concepts does not need to be solved for the purpose of naming clades under the PhyloCode.  Phylogenetic nomenclature can and should remain logically independent from the philosophical debate about species.  Therefore, the PhyloCode need not and should not anoint any particular species concept as the correct one.  By leaving the word “species” out of the entire document it will be clear that the PhyloCode is available to all systematists regardless of their views on the nature of species.

The PhyloCode as currently constructed works under the assumption that species are not clades, yet paradoxically brings species into the naming of clades in a couple of damaging ways.  One is by its current ruling (Article 10.9) that traditional species names should be disallowed for clades.  The current version of the PhyloCode, influenced by Dayrat et al.(2008), effectively applies special rules at the “species” level.  When a clade happens to approximate to a traditional species in content, the PhyloCode now mandates that that clade be given a new name, distinct from the traditional species name.  This is unlike the case when clades approximate in content to genera or families (etc.), in which case the Code strongly encourages the conversion of the traditional name into a clade name.  In this way, the PhyloCode establishes parallel nomenclatural systems for species and species-approximating clades.  We argue that this is illogical; as there are no ranks under the PhyloCode, there should not be an explicit or implicit rank of species.  Phylogenetic nomenclature should accommodate clades whose content is roughly comparable to current species (under whatever species concept).  We propose that the PhyloCode be modified to be neutral about species and thus to accommodate all users, including those who wish to be able to attach appropriate names to clades that approximate taxa at the traditional species level.

Another way in which the current version of the PhyloCode generates problems for itself  is that it allows species to be used as specifiers for clade names (Art. 11), and relegates the governing of species names to the traditional Codes (Art. 21). This means that the PhyloCode is not a freestanding system of nomenclature.  This is unwise; the PhyloCode should be independent and self-contained.  No rules in the PhyloCode should depend on rules in the traditional Codes, including rules for species names.  This is easily solved by requiring that specifiers under the PhyloCode be museum or herbarium specimens (physical reference objects, possibly Linnaean types at the discretion of the taxonomist) at all levels, never Linnaean binomials per se.  Linnaean species binomials are incompatible with phylogenetic taxonomy because they naturally imply the existence of a genus rank, and they are incommensurable with phylogenetic taxa because they are named using only one type (Cellinese et al. 2012).

Since its purpose is to name clades, any mention of ‘species’ in the PhyloCode should be removed.  We propose to streamline the PhyloCode to focus solely on rational procedures for naming clades with uninomials at any level including the traditional species level.  The rules governing names of clades at and around the traditional species level should follow exactly the same rules and recommendations as at higher levels.  To view the original proposals we submitted to ISPN click here. Also see the posting at Phylonames.

References:

Cellinese, N., Baum D.A., Mishler B.D. 2012. Species and Phylogenetic Nomenclature. Syst. Biol. 61: 885-891.

Baum D. A. 2009. Species as ranked taxa. Syst. Biol. 58:74-86.

Dayrat B., Schander C. , Angielczyk, K. D. 2004. Suggestions for a new species nomenclature. Taxon 53:485–591.

Dayrat B., Cantino P.D., Clarke J.A., de Queiroz K. 2008. Species names in the PhyloCode: the approach adopted by the International Society for Phylogenetic Nomenclature. Syst. Biol. 57:507-514.

de Queiroz K. 2007. Species Concepts and Species Delimitation. Syst. Biol. 56:879-886.

Mayden R.L. 1997. A hierarchy of species concepts: The denouement in the saga of the species problem. In: Claridge M.F., Dawah H.A., Wilson M.R., editors. Species: the units of biodiversity. London: Chapman and Hall, p. 381-424.

Mishler B.D. 1999. Getting rid of species? In: Wilson R., editor.  Species: new interdisciplinary essays. Cambridge: MIT Press, p. 307-315.

Mishler B.D., Theriot E.C. 2000a. The phylogenetic species concept (sensu Mishler and Theriot): monophyly, apomorphy, and phylogenetic species concepts. In: Wheeler Q.D., Meier R., editors. Species concepts and phylogenetic theory.  New York: Columbia University Press, p. 44-54.

Mishler B.D., Theriot E.C. 2000b. A critique from the Mishler and Theriot phylogenetic species concept perspective: monophyly, apomorphy, and phylogenetic species concepts. In: Wheeler Q.D., Meier R., editors. Species concepts and phylogenetic theory.  New York: Columbia University Press, p.  119-132.

Mishler B.D., Theriot E.C. 2000c. A defense of the phylogenetic species concept (sensu Mishler and Theriot): monophyly, apomorphy, and phylogenetic species concepts. In:   In: Wheeler Q.D., Meier R., editors. Species concepts and phylogenetic theory.  New York: Columbia University Press, p.  179-184.

Pleijel F. 1999. Phylogenetic taxonomy, a farewell to species, and a revision of Heteropodarke (Hesionidae, Polychaeta, Annelida). Syst. Biol. 48:755-789.

13 comments:

  1. Please no. Already ICZN taxonomists think that the PhyloCode is dead, dying or mired in terminal indecision. If we now make a major change like this, which will entail completely redoing the Companion Volume, we won't get the code actually implemented for another decade, by which point every last iota of momentum will be gone.

    Sorry. I understand why you want to do this, and if it had come up ten years ago I would probably have supported it. But to do it now would be madness.

    ReplyDelete
  2. This is not about saving time. This is about getting it right! We will get the Code implemented. I have all interest in doing so. But making a difference between ranks is simply wrong, especially in a rank-free Code. The species rank is not special and clades should be treated equally across the border. If you understand why we want to do this, then that should be the focus of the discussion.

    ReplyDelete
  3. "We will get the Code implemented. I have all interest in doing so."

    When? Sorry, but "this is not about saving time" is no longer the reasonable response that it was back in 2000. I don't want to be crude, but it's time for the Code to poop or get off the pot: the rest of the world already sees this effort as moribund. Re-tooling now will reinforce that perception. We long ago lost the opportunity to strike while the iron is hot; let's at least catch it while it's lukewarm.

    As for "making a difference between ranks is simply wrong, especially in a rank-free Code" -- I do have sympathy with this. In my day job over the years I have spent many hundreds of hours working on specification for data formats and interchange protocols, and I too know the longing to Do The Job right. Unfortunately, I also know the pain of releasing a perfect standard to a world too apathetic even to say "meh", and I am very much afraid that is where we're headed.

    The brutal reality of biological nomenclature is that everything is anchored on species, and specifically on Linnaean binomials. I'm sorry, I wish it weren't so -- I'm on record as hating binomials, I think they were a terrible mistake -- but that is the world we live in. If the PhyloCode is to make any dent at all in the way nomenclature is done, it simply has to live in that world. Make our own parallel universe, where everything is anchored on specimen numbers, and no-one will come.

    Again, I am truly sorry to be so discouraging, especially as I've been through much the same experience as you (in other fields) and know how dispiriting it is to run into nay-saywers. But that is the reality.

    ReplyDelete
  4. There are no opportunities lost in science. This is not a business venture, and most of what we do now will stick with future generations, so it is very importan to do it right. This is not a matter of now or never. The Code will be published indeed but not tomorrow and hopefully by the end of the year, only hopefully. So, this is a good opportunity for refining it and make sure that it is widely accepted by a lage subset of the community. Proceeding otherwise may imply that many supporters won't use it. The brutal reality of anything at all is that changes are hard to embrace, new standards always hard to accept, but this should not imply that the implementation of new concepts should be impeded by the natural human reluctance for changes. If the PhyloCode is to make any dent at all, it needs to be flexible and inclusive and not bound by any particular philosophy. Living with something that is fundamentally wrong and a clear philosophical contradiction is not the way I like to do science. This is not a day job where compromises lead to a positive end of the day. Let's move on can only means let's do it right.

    ReplyDelete
  5. Well, that is a very appealing picture of the world.

    I wish I could be a so sanguine. But I am not. Political realities are real, and the more the PhyloCode annoys taxonomists (e.g. by requiring definitions to be anchored on specimens rather than species), the fewer of them will come on board.

    I am truly sorry; but it's a non-starter. It just isn't going to happen that people will define Titanosauriformes as (FMNH P25107 + PVL 4017-92). Replacing names with numbers in the specifiers would be just as unhelpful as doing so in the clade-names, so that instead of defining Titanosauriformes, we defined Clade #5873-1235. No matter what you do, however theoretically pure your position, however compelling your argument, you will never persuade taxonomists to give up names for numbers.

    ReplyDelete
  6. Hi Mike,

    I'm not sure what your background is, so I don't mean to be insulting, but you do know that names under the current codes are anchored on specimens. right? The only difference is that the PhyloCode insists on multiple specifiers (for good reason). But believe me, names under both the ICBN and ICZN are defined by type specimens.
    There is a compromise position here, at least as far as I am concerned -- I wouldn't mind so much if the Phylocode used existing species names as specifiers *as a shortcut* to referring to their type specimen, as long as it was clear that the specifier legally is the type specimen rather than the species name.
    You seem to be going off the deep end in arguing against the rational proposals we made (have you read the text of our proposal?) -- no one is arguing to replace names with numbers.

    ReplyDelete
    Replies
    1. "I wouldn't mind so much if the Phylocode used existing species names as specifiers *as a shortcut* to referring to their type specimen, as long as it was clear that the specifier legally is the type specimen rather than the species name."

      That's exactly what the current draft does:

      "Note 11.1.1. When a species is cited as a specifier, the implicit specifier is the type of that species name (if a type has been designated) under the appropriate rank-based code...."

      Delete
  7. Hi, Brent. My background is multilingual: I have a couple of published papers on aspects of PN, and a couple of others on rank-based zoological nomenclature. I also have a couple of chapters in press in the looong-awaited Phylonyms volume. So while I am very far from being expert in either form of nomenclature, I do have a reasonably balanced perspective.

    Yes, the current (i.e. rank-based) codes are anchored on specimens; but precisely so that we can then stop talking about individual specimens and start talking about names instead. You could say that the whole point of the ICZN is that I can say Brachiosaurus altithorax instead of FMNH P25107. To expect taxonomists to allow those specimen numbers to arise from their tombs for use in phylogenetic definitions is really not realistic.

    "There is a compromise position here, at least as far as I am concerned -- I wouldn't mind so much if the Phylocode used existing species names as specifiers *as a shortcut* to referring to their type specimen, as long as it was clear that the specifier legally is the type specimen rather than the species name." I guess that could fly. So long as it made no difference to what taxonomists had to actually do it shouldn't be too hard to sell. In fact, this is pretty much the status quo -- see Note 11.1.1 of the current draft PhyloCode: "When a species is cited as a specifier, the implicit specifier is the type of that species name (if a type has been designated) under the appropriate rank-based code."

    "You seem to be going off the deep end in arguing against the rational proposals we made." This is true, I admit it. Because I know exactly how this proposal will be met by mainstream taxonomists. And we need to win them over.

    Right now, we have a little bit of momentum with people like Mike Benton (my Head of Department at Bristol), who in the past has argued publicly against PN in general and the PhyloCode in general (e.g. Stems, nodes, crown clades, and rank-free lists: is Linnaeus dead? Biological Reviews 75:633-648). Despite his arguments in that paper and others, he is finding it expedient to include clade definitions in his work, which should be a cause for rejoicing. If the poor old PhyloCode ever makes it out of the door, I suspect Benton and other traditionalists will use it because they recognise that there is a need for PN, and it is better that it be regulated by a code -- any code -- than none. But if we throw that up in the air, we will lose what momentum we have with people like him.

    Finally, at the risk of repeating myself, let me say again I am truly sympathetic to what you're trying to do here. As science, I admire it. But as politics, it's a disaster. And I have been involved in too many scientifically admirable but politically disastrous standards efforts to stand idly by and watch another one die the same horrible death. Cruel to be kind.

    ReplyDelete
  8. Mike, what you are saying doesn't make much sense. We are trying to extend the Code to ALL clades including those that may correspond to traditional species. That does not mean that you can't use the Other Codes to name species, if you want to. All our proposal does is to allow us to name all clades including the most inclusives and make a reference to traditional names, just as you are doing it already at higher level. We are not even implying that those least inclusive clades are 'species' and that is why we want to term 'species' removed from the Code. In fact, we are trying to simplify the Code and make all clades equal. It seems to be that there is a fundamental misunderstanding of the issues. We are not even advocating to switch numbers in lieu of names, I am not sure where you see that.

    ReplyDelete
  9. "It seems to be that there is a fundamental misunderstanding of the issues. We are not even advocating to switch numbers in lieu of names, I am not sure where you see that."

    From your proposal, in the section on Articles 11.1-11.10: "All reference to species as specifiers should be removed from these articles and notes. Replace "species, specimens, or apomorphies" with "specimens or apomorphies" throughout. Remove Note 11.1.1, Article 11.3, and recommendation 11.B."

    Doesn't that mean I would have to define Titanosauriformes as (FMNH P25107 + PVL 4017-92) instead of (Brachiosaurus altithorax + Saltasaurus loricatus)?

    ReplyDelete
  10. Hi Mike,

    Yes, that is what it means. But in reference to your response to me (above), once we define Titanosauriformes under the PhyloCode using specimens, from then on we just refer to Titanosauriformes. The details about specifiers are available in RegNum, but you just use Titanosauriformes in everyday parlance.

    So this is just like the current codes, like you say, once someone names Brachiosaurus altithorax using a type, then the user doesn't need to mention the type every day. But the type still is behind the name in databases. So there is no difference here between the PhyloCode and the traditional codes, other than the Pylocode using multiple "types" rather than one. In many ways what we are suggesting is no major change for current practice.

    Traditionalists can see why one needs multiple "types" to specify a clade precisely, and many are starting to names clades, as you say.

    ReplyDelete
  11. The problem isn't in using the name Titanosauriformes, it's in defining it (and reading that definition). A definition based on numbers is less comprehensible than one based on names. We don't want to be erecting additional barriers to using the PhyloCode. Using a specifier such as Brachiosaurus altithorax is already totally unambiguous: however many species concepts might be floating around out there, the only only that matters for nomenclatural purposes is "whatever this species has been defined as under the ICZN". Which is why the existing Note 11.1.1 is perfectly adequate: "When a species is cited as a specifier, the implicit specifier is the type of that species name (if a type has been designated) under the appropriate rank-based code".

    ReplyDelete
    Replies
    1. I still think you are over-reacting on this issue. We are not advocating anything here that is not in the current codes. When you name a species under the current codes you cite a type specimen including its number (read a species description). Just like we suggest for the PhyloCode.

      And as you point out, the PhyloCode already has note. 11.1.1. We just want to extend it a bit and remove the "if" clause to make it clear that the type specimen is the official, legal specifier, even if a taxonomic name is given for reference.

      Delete