Skip to content

Can we teach logic reasoners to correctly apply nomenclatural rules?

No, this not concept taxonomy. Instead this post is intended to set up a nomenclaturally and taxonomically correct use case for an Answer Set Programming (ASP) reasoning project. The post will limit itself to: (1) getting the required nomenclatural emendations “right” (Section 4), given (2) a specific set of starting conditions (Sections 1 & 2), and (3) a new taxonomic insight that necessitates change (Section 3).

1. Starting conditions

This is a zoological use case (ICZN). The year is 2000. The family Agenidae contains one subfamily, two tribes, four genera, and 12 species, totalling 20 published names, as detailed below (Figure 1). The respective years of publication for the ranked names are provided, ranging from 1775 to 1985. * indicates Priority at the relevant rank. This leads to “Priority chains” such as [9] – [5] – [3] – [2] – [1] based on the designation of A. primus, 1775*. The four species-level type names were published from 1775-1790; all non-type species-level names are younger (20th century).

All 12 species represent taxonomically distinct, valid taxa. Each of the type species (and name) has a type specimen whose “apparent” morphology – square (primus), rectangle (secundus), hemiellipse (tertius), and pentagon (quartus) – serves to anchor the circumscription of the entailing genus (up to tribe, subfamily, and family – where applicable). “Apparent” in preceding phrase means that at the original time of publication (1775-1790, respectively), the type specimens were asserted to have these properties. The younger species types were classified accordingly; e.g. [10] and [11] have type specimens that exhibit a square, hence they were assigned to Agenus, 1775*.


Figure 1. Use case starting conditions prior to affecting the taxonomic change.

2. Complication – some type specimens are not what they seem(ed)

Notice that in Figure 1, each of the four oldest type specimens is somehow “compromised”, as represented by a grey circle that (symbolically) interferes with an unambiguous recognition of their morphology (or molecular identity, for that matter). The details are not relevant – the general point is that specific interpretations of the types’ morphological identities were made at the original times of publication (1775-1790); yet now (in the year 2000) it is evident that these assertions of morphology were in error.

3. Change scenario 1 – the type specimen of Agenus primus is rectangular

One particular change scenario following re-examination of type specimens in 2000 is shown in Figure 2. Accordingly, the type specimen of A. primus, 1775* is no longer considered to have a square morphology (!); instead exhibiting a rectangular morphology (green rectangle) that also characterizes the type specimen of E. secundus, 1780* – with additional nomenclatural and taxonomic contingencies (see Section 4). All other type specimens are “resolved” to exhibit the properties as originally asserted (hence no grey circles are shown for secundus, tertius, and quartus).


Figure 2. The type specimen for Agenus primus is re-examined and asserted to be rectangular.

4. Nomenclatural consequences

In this particular change scenario, the type species name primus has “global” Priority. It cannot be moved into synonymy of any other (and necessarily younger) name. Hence:

  • E. secundus => A. secundus, new placement, new combination
  • E. septimus => A. septimus, new placement, new combination
  • E. octavus => A. octavus, new placement, new combination
  • Egenus => new synonym of Agenus, genus level
  • Novagenus => new genus name, child of Agenini *
  • Novagenus quintus => new combination
  • Novagenus quintus => new type designation for Novagenus
  • Novagenus sextus =>new combination

* The prefix Nov- is chosen arbitrarily; the main point is that a new genus name is needed to accommodate [10] and [11].

5. Other change scenarios

Thirty-six such scenarios are possible if (1) all species-level taxa remain valid and (2) only one taxon is “moved”.

7 Comments Post a comment
  1. Neal Evenhuis #

    1. In order to be absolutely correct in the scenario that a re-examination of primus was actually that of the type, you need to put into the parameters that before examination, it was verified that the types of each specimen examined were the actual type specimens. Many older 1700s types and even into the 1800s had type specimens replaced by “better looking” specimens or for other reasons — the chances of very old types not being types are better than younger types. But in any case, that possibility always exists so for the sake of a case dealing with a logical process, it needs to be included and dispensed with.
    2. You do not consider Article 70.3 that fixes misidentified types (primus is essentially a misidentified type [= secundus]. That needs to be included as well since it is a part of the process of resolving generic circumscriptions and common usage.

    April 22, 2014
  2. Stephen Thorpe #

    I disagree with Neal over his point (2). Agenus primus cannot be a misidentified type species of Agenus, as they were both established by the same author in the same publication! Misattribution of characters is an entirely different issue to misidentification!

    April 22, 2014
  3. Stephen Thorpe #

    Oops, I am assuming that Agenus primus was fixed as type species of Agenus in the original publication (e.g. by original monotypy). If it was subsequently fixed, then it could be a misidentified type species, and Neal would be correct!

    April 22, 2014
  4. Thank you, Stephen (and Neal as well). This is helping me make the use case more specific; i.e., as specific as intended. To clarify – yes: Agenus and primus are presumably established by the same author in the same publication in 1775. So this is supposed to be a case of fixation of the Agenus/primus type relationship by monotypy and in the same (original) publication.

    April 22, 2014
  5. Stephen Thorpe #

    On another issue, you have to assume that none of the names are junior homonyms, and that strict priority does apply, which might not necessarily be the case if an older name has not been used for a very long time. Reversal of precedence can happen under some circumstances.

    April 22, 2014
  6. Stephen Thorpe #

    Also, conveniently in your example, all of the publication dates (at each level) are different. If they weren’t, then first reviser choices may be needed to determine the valid name in cases of synonymy.

    April 22, 2014
  7. Thanks again, Stephen. Yes – no homonyms in this set; priority does apply because (presumably) the names have been in comparable and regular use; and no human-centric first reviser choices are indicated either. The use case would need to specify all of this.

    April 22, 2014

Leave a comment to Stephen Thorpe Cancel reply

You may use basic HTML in your comments. Your email address will not be published.

Subscribe to this comment feed via RSS