[Zoobank-list] "the real challenge"

Fri Oct 7 20:59:08 BST 2005

Hi Martin,

> > If availability were tied to (fast) registration only, 
> > uncoupled from (much slower) publication, what would 
> > be the safeguard prior to publication against misapplication 
> > (intentional or unintentional) of the available but 
> > insufficiently defined name?
> 
> Sorry, but I think you're aiming at a danger other than the one 
> I see. 

You are correct -- I did misinterpret your question, and I apologize for that.

> Secondly, lag times between registration=availability and subsequent 
> ("secondary", as you call it) publication, and in your scenario these 
> lags could be long or even unlimited, significantly increase the risk 
> of threats to stability from uses of an available name in taxonomic 
> senses other than that intended by the name's author. 

Yes, I understand the concern now, and I do not dispute that it is real.  However, keep in mind the alternative: that a name, possibly needed by the community for a well-recognized taxon, awaits formal published description for a long or even unlimited amount of time, and is unavailable for use by anyone until it is finally published.  So, the alternate risks are:

1) Name registered (=available) now, and described in detail later -- risk of mis-using the name by other authors for an inappropriate taxon between registration date and description publication date.

2) Name not made available until after publication of formal description, and therefore unavailable to broader taxonomic community during same period of lag-time before appearance of publication -- risk of being named a second time by an author unaware of the first in-prep or in-press description, and/or references to "Aus sp. of Jones" in other publications by authors who are waiting for Jones' description to be published.

I think both risks are real -- the first much reduced in the "published+registered = available" Scenario, and the second much-reduced in the "registered=available" scenario.

In fishes, the second problem is VERY real, and the first problem potentially small; but I realize that fishes comprise a very small part of the 20,000 new names/year, so I don't know how the problems compare in other taxonomic groups.

Here are some factors that would mitigate the first problem in the scenario I propose (registered=available):

1) As per the existing Code, and the proposed change to the Code for this Scenario (http://www.iczn.org/new%20ZooBank6.htm, Arts. 17.2.3 & 17.2.4), distinguishing characters of the new taxon must be provided [existing Art. 13.1.1, proposed Art. 17.2.3], and existing taxa from which it is distinguished [proposed Art. 17.2.4] would be a required component of Registration.  Competent taxonomists will provide adequate information of this sort; and incompetent taxonomists might not.  Just as is currently the case with paper-published descriptions.

2) I would STRONGLY advocate that at least one (and allowing for many) image(s) of the Holotype be a mandatory component of Registration.  I understand that this might impose undue hardship on practicing taxonomists who lack access to electronic imaging technology (digital camera or scanner), so this might need to be considered carefully.  But for many taxa, a picture paints a thousand characters, and would go a long way to mitigating mis-application of a registered name (especially in conjunction with type locality).

3) In order to register a new name, a taxonomist would have to be registered (with contact details) on ZooBank.  Thus, any taxonomist who wishes to apply the registered name during the lag time before its formal description is published, would have easy access to contacting the person who registered the name, requesting more information or access to specimens for nomenclatural comparison.

4) Registration includes mandatory citation of holotype specimen, and therefore the ability for other authors to compare it directly for their own use (admittedly not a major mitigating factor for the problem you describe, due to the cumbersome nature of accessing holotypes, and the low likelihood that authors would go to this much trouble).

5) Faith in the competence of those authors who wish to cite the newly registered name to ensure they understand the parameters of the taxon in which the holotype is a member, before they attempt to apply it in their own use.  This may be overly wishful thinking, of course.

Bear in mind also that there is nothing preventing taxonomists from registering their new names concurrently, or after the date on which their formal descriptions are published, thereby allowing the opportunity to eliminate the "risky" time-lag altogether.

I concede there is a risk here, but I maintain that, with mitigating factors described above (and ones we haven't thought of yet), the net cost-benefit ratio favors the registered=published scenario -- especially when considering the costs of having the name totally unavailable during the lag before publication in the alternative paradigm.

> Thus, the net result of your scenario would be more threats to 
> stability than under the current system, at least as regards 
> the specific issue we're discussing here.

Though I agree with the point you raised above, I disagree that the "net result" is more threats to stability -- because you have not considered the *existing* threats to stability (e.g., redundant names being proposed for the same taxon by different authors unaware of each other's work in prep/press; need to references non-names like "Aus sp. of Jones" while awaiting formal description, etc.) that would be eliminated in the proposed scenario of registered=available.

> In addition, I disagree with the uncoupling of availability and 
> publication for a principal reason: nomenclature is useless 
> without the taxonomic content it is supposed to help communicating 
> about; 

Here is where our core philosophical disagreement may lie.  According to the way the Code is now structured, the *only* objective taxonomic content of a name is the type specimen or syntype series.  Everything beyond this is subjective interpretation by taxonomists, and therefore outside the realm of nomenclatural rules.  One of the greatest challenges of informatics faced by the taxonomic community is the rampant historical muddling of taxon names with taxon concepts.  The organismal content of a *name*, per se, ends with the type specimen(s).  The *application* of that name to a broader scope of organisms representing a taxon concept is the science of taxonomy, and is currently outside the realm of nomenclatural rules.  For 250 years, taxonomists have confounded these two informational domains in traditional practice (e.g., stating a name without referencing a defined concept), and from my perspective, this is the single greatest impediment to digital information flow in the field of biology. The solution, from my perspective, is to clearly disambiguate what is meant by a taxon name, and what is meant by a taxon concept, so we can reduce or eliminate the muddling.

> therefore it should not be completely disassociated from 
> this content (in time and space). 

Due to existing 13.1.1, and proposed 17.2.3/17.2.4, they would not be completely disassociated.

> The increased risk of uncoupled, 
> parallel or even diverging worlds developing, and meaning being 
> lost in both worlds as a consequence, is too large to accept this 
> increase voluntarily and unnecessarily.

I don't know what to say, other than "I disagree".  When you consider that adherence to the Code *at all* is voluntary (yet essentially all taxonomists comply anyway), and that we are all at liberty to create Code-compliant names with equal or less information content than a plain registration entry (yet very few of us do), much *already* rides on faith in the competency of taxonomists not to abuse nomenclature.  Competent taxonomists will continue to produce competent taxonomic works, and incompetent taxonomists will continue to give the rest of us headaches.  But the safeguards against incompetent taxonomy afforded by mandatory electronic registration would dwarf, in my opinion, the shaky and non-enforced safeguards that we perceive exist (but don't really) in the current paradigm of mandatory paper publication of new names.

Incidentally, the reason I champion the registered=available paradigm is because I honestly believe that among the three proposed scenarios for mandatory registration, this is the one with the best combination of likelihood of adoption and increase to overall nomenclatural stability.  It's not that this scenario is perfect (it is definitely not); rather, it is that I see even more problems with the other two scenarios.  Right now the only fourth option is abandon mandatory registration.  Perhaps someone else can propose another scenario of mandatory registration that is even more likely to be adopted, and contributes even more to nomenclatural stability.

> > I think that this general approach you describe has been voiced 
> > several times on this list before, without much dissent.  I think 
> > everyone understands that mandatory registration of post-200X 
> > names will be a different system from retrospective registration of 
> > pre-200X names.
> 
> I don't see that as the essence of what I described, and I do see 
> differences to the two scenarios on the ICZN webpage and the three 
> discussed here.

Several people have endorsed the approach of mandatory registration only (soon), with gradual ramp-up as experience with online registration guides us.  The separate issue of retrospective registration has also been addressed, and seems to suffer less diversity of opinion.

> The steps you're concerned about are not part of my proposal. 
> I am not advocating stepwise changes to the registration/availability procedure, 
> but a quality control system - especially for previously established data  - 
> that is cumulative and open-end rather than inflexibly tied to a fixed date 
> of data closure. To clarify, the scenario I favor is:
> 
> 1. Mandatory registration as an additional requirement as 
> quickly as possible. I don't believe this needs much testing. This kind of system 
> has been working just fine with GenBank practically from the beginning of the latter, 
> with nobody I know having any problem with it. The GenBank system can even be 
> improved upon if open review of submitted data (nomenclatural only) is made possible 
> (but not mandatory) in ZooBank. That way, at least some homonymies, language errors 
> in name formation, etc., could be avoided prior to availability, even without the 
> distant-future comprehensive database being operational.

How is this different from Scenario 1 (published+registered = available)? One difference might be (if I read you correctly) is that during the period between registration and formal publication, the registered name will be publicly exposed (but not officially available), thereby allowing review and feedback prior to availability.  Is this correct?

> 2. Begin open-access entry into a registry of previously established data 
> as quickly as possible as well, but without making it mandatory for 
> availability, etc. 

This has been assumed by essentially all parties from the start.  Because it has not been contentious, it has not dominated discussions on this list.

> Install a stepwise data quality control system with ICZN Official list protection
> once a certain quality level is reached.

This is the discussion that took place on the ICZN list, and I think it's a fundamentally good approach, and one that flows naturally from large-scale efforts involving non-mandatory retrospective registration. 

> 3. Leave any additional changes to a thorough re-edition of the Code, that ought 
> to promote more than so far the most important principle of the scientific 
> method: that everything be based on reproducible evidence.

I guess the only difference of opinion we may have is that I believe that the time it takes to adequately test the registration system before making it mandatory, is commensurate with the amount of time it would take to carefully and thoroughly examine the changes needed for the next edition to the Code.  It seems to me that the bottleneck in terms of time required for a new Edition is more about making sure the broader taxonomic community is aware of the proposed changes, and has sufficient time to participate in the discussion and provide feedback about the changes.  This should apply equally to mandatory registration (of any sort), in my opinion.

Finally, you did not address one of my questions/points to you, which is the intermediate version of the Code.  If I understand you correctly, you propose that the 4th Edition be amended to include mandatory registration first, and later a fifth edition would go into effect.  For the sake of argument, let's say the implementation date of the amended 4th edition would be 2008, and the 5th Edition would be 2012.  We already have five time-spans to accommodate separately in the Code (pre-1930, 1931-1960, 1961-1984, 1985-1999, 2000-next).  If I understand you correctly, your proposal involves creating two more during the next decade (2007-2011, 2012-next).  If it will really take at least 2 more years before we can confidently implement mandatory registration, and 6 more years to establish a new Edition of the Code, I believe we will be doing future generations a great service by postponing the implementation of mandatory registration by a few years, and implement all the "major" changes (of which I consider simple mandatory registration one) at one time.

Apologies for the huge post.

Aloha,
Rich

Richard L. Pyle, PhD
Ichthyology, Bishop Museum
1525 Bernice St., Honolulu, HI 96817
Ph: (808)848-4115, Fax: (808)847-8252
email: deepreef at bishopmuseum.org
http://hbs.bishopmuseum.org/staff/pylerichard.html