Cladistic Systematics

Cladistic systematics (cladistics, cladogenetics) is a version of phylogenetic systematics, in which a one-to-one correspondence between phylogenetic pattern and phylogenetic classification is the guiding idea, and genealogy is the key parameter underlying the classification. To implement this idea, phylogenesis is reduced to cladogenesis as the only parameter that can be considered universal and capable of being reflected by a classification with minimal distortions. Such a reductionism means that claiming this approach most fully embodies the fundamental aims of evolutionary interpreted systematics [de Queiroz 1988; de Queiroz and Gauthier 1990, 1992] is not correct.

The cladistic concept originated in the depths of the German school of morphologists and phylogeneticists, in which Haeckelian systematic phylogeny emerged in the second half of the 19th century [Hamilton 2014; Rieppel 2016]. The botanist Walter Zimmermann was the first to formulate its main provisions and formalize some important concepts [Zimmermann 1931, 1934]. His approach was then further developed by the zoologist Willie Hennig: the first complete version of Hennig’s phylogenetics, published in 1950 in German [Hennig 1950], like Zimmermann’s papers, was left almost unnoticed; the beginning of the “cladistic revolutions” was laid by the publication of its revised edition in English [Hennig 1966]. Zimmermann and Hennig identified their simplified phylogenetic concept with the whole of phylogenetics, and with this, phylogenetics with phylogenetic systematics. Both identities are hardly correct because phylogeny is a multifaceted natural process, of which cladogenesis is but a part, and phylogenetics studying it is not identical to systematics construing classifications (see above). Therefore, the designation of Hennigean phylogenetics, with its focus on cladogenesis, as cladistics [Mayr 1965] seems more than justified. The latter’s main principles are presented in a number of monographs [Nelson and Platnick 1981;Wiley 1981;Ax 1987; Pavlinov 1990, 2005; Wagele 2005]; besides, many most recent textbooks on systematics deal basically with its cladistic version.

The conceptual space of cladistic systematics is built on a fairly rational basis using some elements of axiomatization. The respective formalisms were present first in the works of W. Zimmerman and used quite actively subsequently by some more recent authors [Lpvtrup 1975; Gaffney 1979; Wiley 1981; Pavlinov 1990, 1998, 2005,2018; Reif 2009]. For this reason, cladistics is distinguished by a fairly well-developed specific procedure of taxonomic research enriched with specific terminology.

The general position of cladistics with respect to the Umwelt it constructs and investigates is declared to be realistic: according to Hennig, all monophyletic groups (in their cladistic interpretation) “from species to higher category rank, have individuality and reality” [Hennig 1966: 81]. However, it was emphasized in Chapter 3 of this book that, the more reductionist an Umwelt is, the less of an objective reality of the entire Umgebung it embraces. This is obviously true in the case of cladistics: its evolutionary model is much more reductionist and therefore is less “realistic” in comparison with the classical phylogenetic one [Pavlinov 2005, 2018]. Be that as it may, cladists believe that their approach “provides all parts of the field studied by biological systematics with a common theoretical foundation” [Hennig 1965: 101] and therefore it is the phylogenetic (actually, cladistic) classification that has the status of “the universal reference system of biology” [Hennig 1966; Wiley 1981; Williams and Ebach 2008].

The cladistic evolutionary model implies a predominantly divergent evolution, in which the proportion of parallelisms and reversions is assumed to be minimal. According to the principle of dichotomy, cladogenesis is a dichotomous branching process; this is strengthened by the (rather artificial) supposition that, in each evolutionary event, the ancestral species “dies out,” giving rise to two descendant species [Hennig 1966]. The monophyletic group is specified as including the ancestral species and all its descendants, with its ancestor being by obligation a sole species and not any supra-species group [Hennig 1950, 1966; Wiley 1981; Ax 1987; de Queiroz 1992; de Queiroz and Gauthier 1992]. Such a group is denoted as holophyletic, and the kinship relation that unites its members is denoted as holophyly [Ashlock 1971; Mayr and Ashlock 1991; Mayr and Bock 2002; Envall 2008]. A group that includes not all descendants of the ancestral species is denoted as paraphyletic; if a group is derived from several ancestral species that are not closely related, it is polyphyletic. To emphasize these differences, the holophyletic group can be designated as inclusive, while the paraphyletic group is non-inclusive [Ebach et al. 2006]; according to another terminological variant, they are monocladistic and paracladistic groups, respectively [Podani 2010]. Connecting their recognition with the topology of the phylogenetic tree, it was proposed to designate holo- and paraphyletic groups as convex, and polyphyletic as concave [Meacham and Duncan 1987]. The term metaphyly was proposed for the situation where it is impossible to distinguish clearly between holo- and paraphyly [Kluge 1989].

The holophyly, or cladistic relationship, is outlined as follows:

taxa (or organs) B and C are closer to each other than to A [if] the common ancestor of B and C (X2) is later than the common ancestor (XI) of all three taxa or organs [...] the temporal relationship between the ancestors of XI and X2 is the only direct assessment of phylogenetic relationships.

[Zimmermann 1931: 989-990]

Thus, this relation is set relatively: two groups are related by closer kinship than each of them is with some third group, if the closest ancestor of the first two groups is not also the ancestor of this third group. The latter can be either a supposed ancestor of the first two groups or some (non-ancestral) group “external” to them. Thus, the concept of outgroup is introduced, and substantiated by the non-operationality of the concept of ancestor [Engelmann and Wiley 1977; Nelson and Platnick 1981; Watrous and Wheeler 1981; Wiley 1981; Pavlinov 1990, 2005; Nixon and Carpenter 1993; Reif 2005a], Accordingly, holophyly is defined strictly and operationally as follows: two groups constitute a holophyletic group if it is shown that they are close with respect to a certain outgroup; the former are usually called sister groups. This definition leads to exclusion of the ancestor-descendant relationship from the definition of monophyly and makes, paradoxically enough, the entire cladistics non-historical, since it implies neither ancestor nor origin, and therefore neither does it imply history [Reif 2003].

The concept of holophyletic group (= clade), central for cladistics, contains serious shortcomings, making it rather problematic [Pavlinov 2005 , 2007c, 2018]. Theoretical ones are caused by the so-called Platnick’s paradox: a monotypic genus cannot be considered holophyletic, since (by definition) it includes only one species, which is obviously a descendant, which leaves the latter without any ancestor membership in the respective genus and thus makes the latter paraphyletic by definition [Nelson and Platnick 1981; Wagele 2005]. This shows that the strict theoretical definition of monophyly in cladistics turns out to be logically contradictory [Ashlock 1984; Pavlinov 1990, 2005, 2007c; Reif 2005b; Envall 2008; Kwok and Bing 2011; Aubert 2015], and this is an example of the inverse relationship between the rigor and meaningfulness of the concept (see Section 3.5 on the latter). Among the “semi-theoretical” shortcomings, a clear dependence of the interpretation of monophyly on the scale at which a particular phylogeny is considered is evident. From this it follows that determining the holophyletic status of a high-rank group (say, a class) with reference to a single ancestral species is non-constructive and therefore not very meaningful [Pavlinov 1990, 2005, 2007c; Gordon 1999]. Finally, at an empirical level, any cladistically defined taxon in practice is always paraphyletic, since it is theoretically impossible to encompass all its representatives, both extant and even more so extinct ones, because of the limited nature of any empirical knowledge. However, this difficulty is easily removed by constructive interpretation of monophyly: when defining a holophyletic group, only those members should be presumed that are known at the time of its study [Pavlinov 2005, 2007c].

The procedure of reconstructing cladogenesis, including analysis of characters and similarities, is generally referred to as cladistic analysis. It is quite highly formalized, especially in numerical phyletics, in which many procedural elements are borrowed from phenetics. Substantially, it corresponds to the development of a cladistic hypothesis at two basic levels: lower-level hypotheses include cladistic characters as judgments about particular semogeneses, whereas the higher-level one is the final hypothesis about the relationship between holophyletic groups [Neff 1986], and operationally it is a multisemogenetic hypothesis as a generalization over the entire set of cladistic characters [Pavlinov 2005, 2007c], Besides, it is reasonable to identify third-level hypotheses in the form of posterior evolutionary scenarios [Eldredge and Cracraft 1980; Pavlinov 2005]: they can be involved in the iterative procedure of the sequential weighting of characters, in which their weights can change during the study.

An elementary operational unit of comparison is the character-bearing semaphoront, defined as “the individual during a certain, however brief, period of time” [Hennig 1966: 6]. This reduction radically distinguishes cladistic phylogenetics from the classical one, in which an organism is considered as a developing whole, which allows inclusion of embryological data in phylogenetic reconstructions. The basic units of cladistic analysis are terminal groups corresponding to the vertex nodes of the cladogram; groups corresponding to its internal nodes and (potentially) interpretable as ancestors are not considered.

The cladistic character is interpreted operationally as a transformation series presuming to reflect a historical sequence of transformations of a certain feature of organisms [Estabrook 1972; Mickevich 1982; Pavlinov 1990, 2005; Pogue and Mickevich 1990; Grant and Kluge 2004; Harris and Mishler 2009]. Its modalities are classified into two main categories: plesiomorphy denotes an initial state, while apomorphy denotes a derived state of the respective feature; the characters themselves are often thus denoted. This distinction sets the polarity of the cladistic character, with several empirical rules being used to determine it [Hennig 1966; Hecht and Edwards 1977; Wiley 1981; Pavlinov 1990, 2005]; some of them can be traced back to the works of early post-scholastic systematics. For example, the criterion of commonality corresponds to the principle of constancy by Jussieu and Cuvier, whereas the rule of reciprocal illumination is a variant of the principle of congruence by Adanson. Cladistic character is treated basically as an indirect indicator of cladistic events and the respective holophyletic groups; its classification significance is determined by its contribution to the final cladistic hypothesis. According to the principle of polarity, its significance is inversely proportional to the probability of its evolutionary reversions and parallelism. According to the principle of congruence, the significance of a set of characters is greater the more consistently they indicate the same sequence of cladstic events. According to Darwin’s principle, the cladistic significance of characters is inversely proportional to their adaptive significance. In all other respects, cladistic characters are assigned the (nearly) equal weights of indicators of holophyletic groups.

From the above definition of cladistic relationship it follows that, to identify a holophyletic group (clade), it is necessary to use only that similarity that indicates the position of its members relative to their closest ancestor or an outgroup. To fulfill this condition, the concept of special similarity is introduced based on the key principle of synapomorphy. For this, two components are distinguished in overall similarity: syn-apomorphy is similarity in apomorphic states and symplesiomorphy is similarity in plesiomorphic states. For the recognition of holophyletic groups, only synapo-morphies are significant; symplesiomorphies are not taken into account. With this, any differences between groups are also discarded when inferring the overall structure of cladistic relationships, which is formalized by the principle of irrelevance of differences (also goes back to Darwin); the latter makes cladistically treated similarity and difference asymmetrical, in contrast to phenetic. Synapomorphy is interpreted in two ways: true synapomorphy refers to the closest ancestor, while underlying synapomorphy means similarity due to parallel evolution within a monophyletic group [Saether 1979, 1982; Pavlinov 1990, 2005]. An advanced character state unique to a monotypic group is interpreted as the latter’s autapomorphy.

It is possible to consider the principle of synapomorphy as a special kind of similaritiy weighting [Pavlinov 1990,2005,2018]. It is based on one-state logic (see Section 3.5): only positive judgments (the presence of synapomorphies) are significant, while negative ones (the absence of synapomorphies) are insignificant for delineating holophyletic groups.

The principle of synapomorphy implies a quantitative assessment of the validity of holophyletic groups, which is formalized by the principle of summation of synapomorphies: the more apomorphies characterize a group, the more reasonably it can be treated as holophyletic [Farris 1983, 1986; Pavlinov 1990, 1998, 2005]. From this follows the relevance of the principle of total evidence, which presumes the inclusion of as many characters as possible in the analysis [Eernisse and Kluge 1993; Kluge 1998; Rieppel 2004, 2005a, 2009]; in phylogenomics, this condition is implemented by whole-genome analysis [Savva et al. 2003].

In cladistic analysis, special attention is paid to the principle of parsimony [Farris 1983, 1986. 2008; Kluge 1984; Pesenko 1989; Pavlinov 1990. 2005, 2018], which is sometimes emphasized by the definition of cladistics with direct reference to this principle [Kitching et al. 1998]. The latter is built into either the ontological background (evolutionary parsimony) in the form of the concept of minimal evolution [Camin and Sokal 1965] or the algorithms of cladistic analysis to minimize the amount of parallelisms and inversions in the resulting cladistic hypothesis (methodological parsimony) [Farris 1983].[1] In structural cladistics, the principle of parsimony serves as the basis for excluding any evolutionary model from the initial conditions of the elaboration of cladistic classifications [Nelson 1978, 1979; Platnick 1979; Nelson and Platnick 1981; Patterson 1983] (see below).

The structure of cladistic relations is represented by a stylized form of the phylogenetic tree called cladogram; in terms of similarity, it reflects the hierarchy of synapomorphies, i.e., it is synapomorphogram [Sneath 1963; Pavlinov 1990, 2005; Williams and Ebach 2008]. With reference to Hennig’s evolutionary model, it is assumed that each node of the cladogram must be dichotomous, in which case it is considered fully resolved; this serves as one of the criteria for its optimality [Nelson 1979; Nelson and Platnick 1981]. The base of the cladogram corresponds to an initial event in the phylogeny of the group studied and defines indirectly the entire hierarchy of its holophyletic subgroups. Each of the latter can be defined in one of the following ways: node-based by reference to a cladogram node, stembased by reference to its internode, and apomorphy-based by reference to the syn-apomorphies [de Queiroz and Gauthier 1990, 1992; Nixon and Carpenter 2000; Sereno 2005].

One of the important properties of the synapomorphogram is a decrease in the number of synapomorphies from top to bottom. Since holophyletic groups are determined through synapomorphies only, this generates a specific cladistic uncertainty that increases in the same direction. This means that judgments about cladistic relationships between taxa are least reliable at the base of the cladogram [Pavlinov 1990,2005,2018]. This regularity is responsible for frequent controversies in treating cladistic relations among basal groups.

The basis for the development of a cladistically natural classification is the interpretation of a cladogram in taxonomic terms, i.e., translation of its hierarchy into a taxonomic one by representing clades as taxa [Hennig 1950, 1966; Nelson and Platnick 1981; Wiley 1981; Williams and Ebach 2008]. To emphasize the specificity of such a taxonomic construction, in which only holophyletic (synapomorphic) groups are present, it was proposed that it should be called not a classification, but a cladification [Mayr and Bock 2002; Hbrandl 2010]. In the latter, the inclusive hierarchy of taxa is determined entirely by the sequence of the nodes in the respective cladogram: this means that the vertical component is dominant and the horizontal component is minimized. This condition is formalized by the principle of equality of ranks of sister groups: all clades converging to the same node of the cladogram are represented by the taxa of the same rank in the respective classification. The horizontal component is implicitly introduced by an analogy of the classical progression rule; in cladistics, a progression is defined by an increasing number of synapomorphies characterizing the respective groups [Wiley 1981; Pavlinov 2005]. The correspondence between the hierarchical structure of cladogram and cladistic classification may be strong or weak: in the first case, all sister groups identified in the cladogram must be reflected in the cladification; in the second, this is not necessary but the prohibition of paraphyletic taxa remains crucial.

The strong correspondence can make the cladification hierarchy very detailed and requiring the use of many additional taxonomic categories, therefore the principle of non-strict hierarchy is frequently followed in two versions. According to one of them, plesions as taxa with an unfixed rank are used, especially for fossil groups [Wiley 1981; Schoch 1986; Pavlinov 1990,2005]; in another variant, taxa of different ranks can be assigned to the same level of generality [McKenna and Bell 1997]. Interpretation of the hierarchy resulting from an asymmetric cladogram also poses a special problem: the ranks of taxa in different parts of the same cladification turn out to be unable to be compared if they are separated by a different number of nodes [Pavlinov 2005]. To avoid all problems with ranked hierarchies in cladifications, it was suggested that they should be abandoned and instead the rankless hierarchy should be used [L0vtrup 1977; de Queiroz and Gauthier 1992; Ereshefsky 1997, 2001a; de Queiroz and Cantino 2001; Reif 2003; Mishler 2009; Zachos 2011]; by this, cladistic hierarchy is formally similar to that of the scholastic genus-specific scheme (see Section 2.2.2).

The taxa in cladifications can be assigned special designations reflecting their phylogenetic status; such cladifications are called annotated [Wiley 1979, 1981]. The entire study group is denoted as pantaxone; crown groups and stem groups are distinguished within it [Ax 1987; Shatalkin 1988; Forey et al. 1992; Sereno 2005]; it is proposed that the taxa corresponding to the stem groups should be called adokimic, i.e., “untrue” [Boger 1989], or parataxa [Meier and Richter 1992], or plesions (mentioned abovej.The groups of uncertain phylogenetic status are termed metataxa, or they may be mixotaxa or ambitaxa depending on their position in the cladogram [Gauthier et al. 1988; Archibald 1994].

As was noted in Section 2.7, development of the cladistic research program, especially coupled with molecular phylogenetics, because of its reductionist character, led to a significant “de-biologization” of biological systematics. Consequently, ideas of possible perspectives of the latter’s post-cladistic development are beginning to be discussed [Wheeler 2008; Williams and Knapp 2010; Zander 2013; Pavlinov 2019, 2020]. However, this program still holds a leading position in systematics judging by both regularly published “primers” exposing its methods and the aggressive attitude of many journals and their peer reviewers against other approaches.

The post-Hennigian development of cladistics led to its differentiation into several subprograms [Charig 1982; Hull 1988; Pavlinov 1990, 2005. 2018; Ebach et al. 2008; Williams and Ebach 2008; Pavlinov and Lyubarsky 2011; Quinn 2017]. They are united by the use of a few emblematic terms: cladogram, holophyly, synapo-morphy, etc. With this, they are significantly different with respect to some important provisions of their onto-epistemology and methodology.

Evolutionary cladistics partly conserves some features of classical phylogenetics [Hill and Crane 1982; Hill and Camus 1986; Pavlinov 1990, 1998. 2005, 2018; Wägele 2005]. It presumes prior elaboration of rather complex evolutionary scenarios for all characters, from which synapomorphic schemes are derived as a basis for reconstruction of the phylogeneric hypothesis. Acknowledgment of the importance of internal parallelisms between closely related groups is a characteristic of this school: they allow underlying synapomorphies to be recognized and, according to one of Darwin’s principles, they serve as an additional support of monophyly [Tuomikoski 1967;Brundin 1972;Saether 1979, 1982. 1986]. In terms of cladistics, this actually means a “legalization” of at least some paraphyletic groups; this position is defended by some botanists actively using cladistic terminology [Cronquist 1987; Brummitt 1996; Hörandl 2006; Stuessy and Hörandl 2014; Willner et al. 2014; Lachance 2016]. It is proposed that classifications thus developed are called para- or patrocladistic [Stuessy and König 2008; Wiley 2009; Carter et al. 2015]. Stratocladistics, also in a traditional style, takes into account geochronological dating of fossil forms for clarification of their cladistic relations [Fisher 2008]. The zoologist Quentin Wheeler believes that the development of this approach means the formation of another “new taxonomy” [Wheeler 2008], but in fact this is a partial return to classical phylogenetic origins.

Parsimony cladistics develops the ideas of the “founding fathers” in a much more formalized way. In it, as in the previous version, the idea of evolution is initially accepted, with the main task being to reconstruct phylogenies as the basis for developing the respective classifications [Gaffney 1979; Farris 1983, 1986; Kluge 1984]. But this approach presumes minimizing the traditional speculative contents of the evolutionary scenarios: with reference to the methodological principle of parsimony, it is argued that such scenarios specified for particular cladistic characters are equivalent to adopting undesirable ad hoc prior judgments. Accordingly, all the latter are excluded from the prior analyses of characters accepted without predetermined polarities; such a background model corresponds to a unified hypothesis of undirected (random) evolution. The latter drastically distinguishes this version of cladistics from the Hennigian one, where irreversibility of character evolution is considered one of the basic assumptions [Hennig 1966]. So, the only prior assumption adopted is that of monophyly of the studied group relative to a certain outgroup; reference to the latter serves as a decisive means in determining the base and hence the entire hierarchy of the cladogram, from which all conclusions about synapomorphies are derived on a posterior basis [Watrous and Wheeler 1981; Farris 1982, 1983; Pavlinov 1990, 2005; Nixon and Carpenter 1993; Barriel and Tassy 1997; Harlin 1999; Sereno 2005]. This approach is based on analysis of as many characters as possible without their prior substantive weighting, indicating a noticeable element of the phenetic idea in its methodology. In order to emphasize its “phenetic bias,” this school is sometimes called neocladism [Saether 1986] or phenetic cladistics [Wagele 2005].

Pattern cladistics develops the original ideas of W. Hennig in a very paradoxical way, and therefore it was called transformed cladistics by its leaders [Nelson 1978; Platnick 1979; Nelson and Platnick 1981; Patterson 1983]. It excludes any prior considerations of phylogeny from its background ontology, thus making the latter non-evolutionary [Beatty 1982; Brady 1985; Scott-Ram 1990; Ebach et al. 2008; Pavlinov 2018]; this gave a reason to call this approach methodological cladistics [Hill and Crane 1982]. Its ontic basis consists of an idea of the hierarchical structure (pattern) of the diversity of organisms as a consequence of the orderliness of their ontogenies [Patterson 1983, 1988]. The general argumentation scheme is almost directly derived from K. von Baer’s epigenetic typology (see Section 2.4.3), according to which the more generalized characters appear earlier than the more specialized ones in animal ontogeny. Based on this, the principle of synapomorphy is reformulated into the principle of the levels of generality’ of defining characters: each of the latter (or their combination) defines a certain synapomorphic group, and their combined internested hierarchy determines the entire hierarchy of the cladogram and thus the cladistic classification inferred from it. As a result, this version of cladistics loses its phylogenetic roots and turns into one of the versions of typology of an epigenetic kind [Tatarinov 1977; Charig 1982; Riedley 1986; Scott-Ram 1990; Pavlinov 2018; Brower 2019].

  • [1] It must be emphasized that methodological parsimony presumes implicitly evolutionary parsimony, because of the instrumentalism effect: no other evolutionary scenario can be inferred from the “most parsimonious” cladistic hypothesis than the one according to which evolution itself is parsimonious [Pavlinov 2005,2007c],
 
Source
< Prev   CONTENTS   Source   Next >