Basic Methods

This section describes briefly the principal methods used in taxonomic research. They are considered from a philosophical standpoint in quite a general manner (as particular methodologies), so all technical details are omitted [Pavlinov 2018].

The sampling method is used at the stage of preparing taxonomic research. The importance of this method is determined by the epistemic sampling principle, according to which taxonomic research is always carried out not on the taxonomic reality itself (diversity as such), but on a research sample representing it. The latter acts as an operational model of an Umwelt and outlines a kind of empirical reality: its formation marks the final step in the cognitive reduction of the Umgebung. The main elements of the sample are organisms and their characters; or, more precisely, their descriptions (see below). In traditional “museum” systematics, the basic research sample is equal to the world collection pool, i.e., a totality of the fixed and preserved derivatives of biological organisms in museums, herbaria, etc. [Pavlinov 2016].

The sampling method is shaped by two basic principles. One is the principle of sample selectivity, which explicates the principles of method correspondence: it presumes that the sample, instead of being composed on a random basis (one of the conditions of classification phenetics [Sneath and Sokal 1973]), should be formed under the purposeful guidance of a conceptually defined research task. Indeed, the requirements for samples designed for phylogenetic or phenetic, comparative morphological or experimental physiological, macro- or microsystematic research will be different. Another is the principle of sample correspondence: this means that the sample must be representative, i.e., it must reflect adequately the structure of an Umwelt under study. Sample correspondence determines the reliability of extrapolations of the results of its study to the respective Umwelt.

The main qualitative characteristic of the sample is its composition, which depends on the content of the sample elements. The main quantitative characteristic of the sample is its size, determined by the number of constituent elements. The cumulative quantitative characteristic of the sample is its dimension defined by a combination of the amount of specimens and their characters.

The descriptive method serves as a means of recording information about sample elements in a form that provides the technical possibility of conducting taxonomic research. The primary description constitutes a basis of empirical analysis of the raw data themselves: it fixes in some way the information about the sample elements. This can be a verbal description of organisms or their derivatives, their analog or digital images (drawings, photos, etc.), voice recorded on a phonogram, figures of individual measurements in a table, molecular sequences, etc. The result of the primary description of a specimen/character is reasonable to consider as its individual descriptive model. It is the latter that constitutes an individual sample element, so at an operational level, the sample-as-model is represented as an aggregate of descriptions of organisms by their features. The secondary’ description is based on a certain set of primary descriptions, and provides their generalized characteristics. This may include a summation of several comparative morphological series, a table with calculated statistical characteristics, etc., as well as a description of the results of experiments or other manipulations with objects.

The comparative method serves as an analytical tool for revealing a certain structure of diversity in the research sample by comparing its elements. The main task of this method is some ordering of sample elements according to their similarities. In the primary’ comparison, the compared objects are the sample elements themselves: specimens are compared over their characters (taxonomic analysis) or characters are compared over the specimens (partonomic analysis). In the secondary comparison, the compared objects are the results of previously conducted analyses: these objects may represent the results of various experiments (e.g., hybridization under different conditions) or comparisons based on different methods (e.g., different clustering algorithms), or comparisons by different sets of characters (e.g., molecular and morphological).

The comparison is contextual and therefore relative: its context is set by a certain comparison basis (standard), with reference to which the similarity relations of the compared objects are assessed. This means that the comparison of any two isolated objects does not have a soundly interpretable meaning; the latter is provided using a certain comparison basis and depends on it. This basis can be represented by some fixed scale (for example, “more or less”) or a certain third object of the same “kind” as the compared ones, including an abstract archetype [Shreyder 1983; Rautian 2001]. Thus, in general, the comparative method is carried out in the form of the ternary relation between three objects, with two of them being compared while the third provides a basis for comparison.

Depending on the particular procedures, the comparative method can take the following forms. The phenetic method is based on comparing sample elements directly on a pairwise basis within a given sample, with the latter constituting the generalized comparison basis. In the typological method, a certain generalized (secondary) element of the sample, usually (arche)type or Bauplan, serves as a comparison basis. The comparative historical method includes the historical time scale as a supplementary comparison basis. If the similarity is assessed quantitatively, the comparative method becomes the mathematical one.

The experimental method involves direct or indirect manipulations with objects, which may be specimens or aggregates thereof. This method is only applicable to fairly simple objects, so it is of quite limited use in systematics; it is most developed in biosystematics and in some numerical approaches. Experiments can be active or passive: the former are carried out under controlled and (preferably) manipulated conditions, whereas the latter are not. If real experiments are impossible, they are replaced by imaginary (mental, virtual) ones, in which the elements of a subjective or virtual reality are manipulated.

The mathematical method is based on numerical estimates and computational operations. Depending on the nature of the latter, this method can be analytical (based on precise calculations by formulas), numerical (based on calculations by sample parameters), or graphical (based on graphical constructions). Currently, these operations are significantly automated by the use of computers.

Standard algorithms of the mathematical method in systematics serve as auxiliary tools for solving analytical tasks. In simple cases, this method, as part of the comparative one, implements the above operations with the objects (sample elements or assemblages thereof) on a numerical basis, including: (a) quantitative assessment of similarities between them; (b) quantitative analysis of the entire structure of similarity relations; and (c) transformation of the latter structure into classification. In more complex cases, this method, as part of the experimental one, serves for conducting virtual experiments: for example, to reveal an effect of changing certain sample parameters on the results of comparative or experimental analyses. Another area of application of the mathematical method is the analytical comparison and/or generalization of different classifications, as well as the analysis of their structure (for example, their compliance with Zipf-Mandelbrot distribution).

The classification method is a kind of quintessence of research activity in systematics as a classifying discipline, implementing the above-mentioned classification methodology at an operational level. It summarizes the results obtained by other research methods outlined above and serves as a kind of “common denominator.” This method is based on the logical procedure of classifying: it is an operation on the descriptive models of the sample elements representing the objects in some way, but not on the objects themselves.4 Its main result is a presentation of the similarity structure of the research sample in the form of classification.

An operational basis of the classification method is the classification algorithm, which recognizes particular classification units (taxa or partons) of different levels of generality and unites them in a single classification system. This algorithm is applied to estimates of similarities between objects that resulted from their previous comparative or experimental analyses. It can be deductive or inductive depending on the basic argumentation scheme. In contemporary taxonomic research, it usually incorporates certain elements of the mathematical method.

In this, classifying differs from sorting in the sense of dividing a set of physical objects into subsets; the sorting may be extensional (based on counting the objects) or intensional (based on comparing the objects by their characters). A variant of the latter is the identification of objects, i.e., their allocation to already-recognized classification units (for example, to different species).

 
Source
< Prev   CONTENTS   Source   Next >