Chapter 8 Chapter 4 Macro Taxonomy, Taxonomy-3

4.6 Buffon The 18th century was a prosperous period for natural history. Cook, Bougainville, and Commerson's voyages were carried out one after another, and the new upsurge in understanding nature was not only reflected in Rousseau's treatises but also in the works of the "French Revolutionary Philosophers" ("Philosonhes") of the Enlightenment.Museum showrooms and plant specimen exhibition rooms have appeared one after another in Britain and the European continent. They are all owned by princes, nobles or wealthy businessmen, such as George Clifford (1685-1760) in the Netherlands, Sir Hans Sloane (1660-1753) in the United Kingdom. ), Sir Joseph Banks (1743-1820), and aristocratic businessmen from various countries in continental Europe.One of the purposes of these patrons of natural history was to publish scientific catalogs of their collections.

Books on nature were growing in popularity at the time, but none could match Buffon's Natural History (Histoire naturelle).Although similar to Linnaeus's taxonomic work, it also discusses the diversity of nature, but Buffon's treatment is fundamentally different.He paid little attention to identification, and only wanted to describe the various animals vividly.He opposed the bookishness of the scholastics and humanists and their emphasis on logical categories, essences, and discontinuities.He is very inclined to and appreciates the idea advocated by Leibniz, which emphasizes the concept of continuity, completeness and Aristotle's degree of perfection.In Buffon's view, this view of nature is far superior to the dry classification of "nomenmologists" (this is Buffon's derogatory term for Linnaeus and his school).His studies of Newton also drove him in this same direction.Haven't the law of gravitation and other laws of physics already proved the unity of nature arising from universal laws?Why should it be cut into species, genus and class to dismember and destroy this unity?In the first volume of his Natural History, published in 1749, he claimed that nature does not know species, genus and other orders; outside of this exhaustive statement).

Buffon initially loved physics and mathematics. Although he had a certain understanding of natural history before, it was not until 1739 when he was appointed as the general manager of the Royal Garden (now the Botanical Garden) at the age of 32 that he became interested in the diversity of nature. Buffon and Linnaeus were both born in 1707, but the difference between them could not have been greater, nor among their followers.The Linnaean school emphasized all aspects of the classification process that help identification, while the Buffon and French schools concentrated on the study of natural (world) diversity.The Linnaean school emphasizes discontinuity, while Buffon emphasizes continuity.Linnaeus adhered to Plato philosophy and Thomas logic while Buffon was influenced by Newton, Leibniz and nominalism.

Linnaeus paid special attention to "basic" traits, often single discriminative traits.Because, as he said, attention to descriptive details would prevent the identification of basic traits, whereas Buffon insisted that "all parts of the object under study must be used," including internal anatomy, behavior, and distribution. Buffon's point of view is very suitable for dealing with the problem of mammals, and it is also a continuation of the tradition of earlier taxonomists such as Gessler.The number of mammalian species is limited and identification is not a problem.Only botanists like Ray and Linnaeus applied the principles of logical classification to animal classification.When Buffon divided mammals into domestic and wild animals, he thought it was "the most natural (reasonable)" classification.For him, "natural" is practical, not as "reflecting essence" as Linnaeus did.

By about 1749 Buffon's views began to change, eventually shifting considerably as knowledge of living organisms increased (Roger, 1963: 566). In 1749, he doubted the possibility of biological classification at all. In 1755, he admitted that there were related species. In 1758, he still laughed at the concept of genus. In 1770 he took the genus as the basis for his classification of birds, but he seems to have reserved his mind on its arbitrariness.Even if he admits that the "genera" of domestic animals share a common descent, they are of course no more than biological species. After 1761 he also adopted the concept of section.But it must be remembered that Buche never attempted to classify the entire kingdom of animals and plants.Most of his "Natural History" are monographs on individual species of mammals.These monographs are excellent, both literary and scientific, and have had a major impact on the education of young zoologists.But these monographs were not intended to be developed into general treatises on systematics, in which Buffon had no interest.

Although they started from completely different poles, as their research progressed, Linnaeus and Buffon became closer in thought.Liberating Linnaeus from the idea of ​​species being fixed, Buffon admitted (contrary to the nominalist view) that species could be defined involuntarily as reproductive communities (Natural History, 1753, IV: 384-386) .But Buffon always disagrees with Linnaeus' view on the nature of the genus, that is, he believes that the genus is the most objective and true view of all levels.Moreover, Buffon's basis for identifying advanced taxa is completely different from what Linnaeus publicly declared to adopt (the entire body type versus a single trait that shows the essence).

By their later years, the 1770s, the differences between Linnaeus' and Buffon's classifications had narrowed to such an extent that their respective traditions had fully merged among their pupils.Lamarck, Buffon's student, yelled that there were no classes, but only individuals, but after all the uproar he paid no attention to this in his taxonomic works. The same is true for Lacepede.Finally, by the time of Cuvier, Buffon's nominalist tradition was no longer recognizable. 4.7 The revival of zootaxonomy Little progress was made in animal taxonomy in the seventeenth and eighteenth centuries.Linnaeus' classification of invertebrates was indeed a step backwards from Aristotle's.But the situation changed dramatically after the publication of Cuvier's (1769-1832) Classification of Worms in 1795.The miscellaneous taxa that Linnaeus called "Vermes" were divided into six new classes of the same level by Cuvier: Molluscs, Crustaceans, Insecta, Worms, Echinoderms, and Phytophytes.Seventeen years later Cuvier deprived vertebrates of their superiority by placing certain invertebrates on the same level as vertebrates.He divided all animals into four phyla (classes, "embranchements"): Vertebrate, Mollusca, Arthrozoa, and Radial Animals (Cuvier, 1812).In these highest taxa a number of new, hitherto intermixed and completely neglected classes, orders and families were identified.He merged Linnaeus' molluscs and scales into Mollusca, and moved jellyfish and sea anemones from Molluscs to Phytophytes.

Cuvier's greatest contribution to animal classification was the discovery that the internal anatomy of invertebrates holds a wealth of information.He dissected many marine animals and discovered a large number of new traits and new structural types.This was the beginning of the famous tradition of comparative invertebrate zoology.These discoveries enabled Cuvier to identify for the first time some new taxa that are still in use today. One thing worth noting about Cuvier's remarkable contributions is that, while based on carefully crafted conceptual systems and laws, they did not go conceptually beyond the principles of Aristotelian logic.He still emphasizes descending classification and strives to discover the true nature of each phylum, and still weights traits according to their functional importance.Although Cuvier also innovated in these areas.

Cuvier believed that certain physiological systems were so important as to control the composition of all other traits.This is a new conceptual breakthrough.Taxonomists before Cuvier looked at the whole, as if each character was independent of the others, and it seemed that organisms with different characters had different essences.Buffon first disagrees with this view.An organism is not a random collection of characters as stated in the works of the Linnaean school. The composition of characters is determined by the "correlation" between them.Cuvier extended this rather broad notion of Buffon to a concrete principle: the correlation of parts (organs) (see Chapter 8).The various parts (organs) of an organism are interdependent, and if the teeth of an artiodactyla juvenile are replaced with those of other animals, the anatomist can immediately recall what changes might occur in the structure of other parts of it.All the functions of an organism are interdependent to such an extent that they cannot be changed independently: "It is precisely this interdependence of the various functions and the mutual coordination and complementarity of the functions that determine the mutual relations of the various organs. laws, which have the same inevitability as metaphysical principles or mathematical theorems, since the fine-grained harmony between the interacting organs is a necessary condition for the survival of the organisms to which these organs belong, and if one of these functions is altered and this If such a change is incompatible with changes in other functions, the organism will no longer be able to survive." ("Course in Comparative Anatomy", Lecons danatomie comparee, 1800I: 51).

Since Cuvier's public statement, experienced taxonomists have used correlated variation as a clue for evaluating characters.Associated changes can not only reveal specialization associated with the occupation of specific adaptive zones, but can also reveal deep-rooted genetic integration represented by trait stability in higher taxa.Lamarck in his Flora of France The importance of character stability was first noticed in (Flore Francoise, 1778) and subsequently by de Jussieu.However, Cuvier not only paid attention to structural correlation, but he also further developed a detailed system of weighted traits, which is embodied in the principle of trait affiliation (see next section). Principles rather than interested in identification schemes.Like Reaumur before him, he doesn't even mind genus or species descriptions.He states what his real purpose is in the following words: "In a word, I have not written this book for the purpose of classification; it would be easier and more suitable to establish the names of species by artificial systems. My purpose is to make The classes of invertebrates and their true relations to one another are better understood by general principles of their known structures and general properties."

Although Lamarck (Jean Baptlste Lamarck, 1744-1829) and Cuvier had different philosophical views, their contributions to taxonomy were very similar (Burkhardt, 1977).Lamarck has also made many valuable contributions to the classification of invertebrates, involving cirripedes, the status of tunicates, and confirming that spiders and annelids belong to different taxa.Indeed, Lamarck has made many contributions to the classification of protozoa to molluscs, but when it comes to taxonomy, his ideas are as old as Cuvier's.When Lamarck set out to study the classification of animals, he believed that animals were arranged in a single series from the simplest ciliates to humans.He thus attempted to classify each higher taxon according to its "degree of perfection".Later, under the influence of Cuvier changing the single series into four circles, and on the other hand, due to his own comparative research, Lamarck gradually gave up the concept of single series.He begins by admitting only that certain species and genera deviate from the straight line (series) due to "forces of circumstance", but eventually he also admits that higher taxa diverge.His final scheme of animal relationships (1815) was not in principle different from the phylogenetic trees contained in related literature of the late nineteenth century. Lamarck often emphasized the importance of the work of classification because "the study of affinity ... should now be regarded as the principal means of progress in the natural sciences." 4.8 Taxonomic traits Classification is the placement of organisms into taxa according to their similarity and mutual relationship; the similarity and relationship of organisms is determined by their taxonomic characters or inferred from their taxonomic characters.This definition indicates the decisive significance of taxonomic traits in establishing classifications.From the beginning of taxonomic history to the present, however, there have been many differences of opinion as to what traits are most useful and reasonable in taxonomic studies.Indeed, the history of taxonomy is largely the history of disputes over this question.The ancient Greeks were well aware that utilitarian traits, such as a plant's medicinal properties or whether it had thorns, had nothing to do with other, seemingly deeper, qualities of plants.The essentialist who classifies logically also feels that some traits are more important, more fundamental, than others.Although their use of the terms essential and accidental character bears the burden of scholastic dogma, they do have an awareness of what was not understood until centuries later.It has been recognized since Chechapino that non-morphological traits such as relationship to humans (raised or wild), seasonality (deciduous or evergreen), or habitat are less useful for classification than structural traits more useful.Structural characters have thus dominated taxonomy since the sixteenth century. Throughout the period from Cecha Pinot to the present, the issue of taxonomic traits has given rise to three main debates: (1) Should only a single key trait or multiple ("everything possible") traits be used? (2) Should only morphological traits be allowed or can ecological, physiological and behavioral traits also be used? (3) Should traits be weighted, and if so, by what criteria? Aristotle had long since suggested that some characters are more useful than others in distinguishing classes of animals; and in the whole history of taxonomy few have disagreed with this conclusion. (Equal weighting of all traits has been advocated in some early works by numerical taxonomists—Sokal and Sneath, 1963, which can be seen as exceptions).Therefore, the question is not whether to weight or not, but two new questions are derived: what principle should be used to determine the weight of traits?How to transfer weighted measures into categories?It should be noted that negating a particular basis for weighting does not mean negating the method of weighting.Buffon and Adanson, while in favor of utilizing "as many traits as possible", do not attempt to weight them equally at point h1. So long as classification is essentially an identification scheme, it necessarily relies on a single trait.As long as the purpose of identification can be achieved, it doesn't matter even if the categories obtained by this method are not uniform.As far as plants are concerned, the experienced botanist is well aware that no other plant structure provides more and better identifying features than fruiting (flowers, fruits, seeds).The special advantage of this structural system is that it contains many quantitative traits, such as the number of petals, and the number of stamens and stamens.Another advantage of flowers is that they are relatively invariant within a species yet contain a large number of variable parts showing species-specific differences.No one was more assiduous or successful than Linnaeus in studying and pointing out these differences, although some of his contemporaries complained that he used traits that could only be seen with a magnifying glass. No essentialist would admit that the use of fruitiness as a taxonomic trait is due to its practical value.Essentialists have created a myth believed only by themselves, that certain aspects of plants are somehow more important than others and thus better reflect the nature of plants.Cechapino ranks the nutrition of plants and their results (growth) in the most important position and puts the reproduction reflecting fruitfulness in a secondary position, because plant generation is the second most important clue to the nature of plants, and other aspects of fruiting ( flowers and seeds) is the third most important trait.The difference between Linnaeus and him is to rank fruit over growth, for the simple reason (Philosophy of Plants, section 88): "The essence of plants consists in fruit." Perhaps the best evidence for choosing flowers based on practical rather than philosophical reasons It is to this day that the flower is used most in the key, although the argument for its "functional importance" was abandoned two hundred years ago. Although all botanists from Geisler (1567) and Cechapino (1583) to Linnaeus recognized the importance of fruit set, there was still a problem of selection because of the variety of characters involved in fruit set.Different botanists choose different traits as the basis for the first step of classification: Tizi Lev and Rivinus choose the corolla, Magnol chooses the calyx, Boerhave chooses the fruit, Siegesbeck chooses the seed, and Linnaeus chooses the pistil and stamen.It would be difficult to prioritize the various components of the fruiting process in terms of their physiological significance.Therefore, botanists before Linnaeus were mostly divided into several factions according to nationality. British botanists followed Rui, German botanists followed Rivinus (Bachman), and French botanists followed Tiu Lev.Because the main purpose is identification, the Tiulev system was more and more widely used by botanists because it was more concise and easier to remember than the other two, until it was later replaced by the more applicable Linnaeus sex system. In the 17th and 18th centuries, the number of known animal species increased rapidly, and the use of morphological traits gradually increased, but among zoologists, very few zoologists were as interested in methodology as contemporary botanists.Ecological traits are still preferentially adopted, especially in animals other than vertebrates.For example, Vallisnieri (1713) divided insects into four categories according to their living environment, plants, animals, water (including crustaceans), rocks and soil.Even when morphological traits were used, they were not carefully selected, for example Linnaeus classified fish-like whales as fish, and he also classified most invertebrates as worms. Cuvier's principle of trait affiliation is a weighting system that refers to the fact that parts (organs) of an organism differ in their taxonomic value.In his early (before about 1805) writings, vegetative organs, especially those related to circulation, were the most important differential characters of the higher taxa of atrial fibrils. By 1807, however, the nervous system had been elevated in importance, taking on the most important role in the division and ordering of his four phyla (Coleman, 1964).At the lower levels of the hierarchy, Cuvier tended to assign different weights to the same trait in different classes of animals.For example, the tooth character specifies the order in mammals, the genus in reptiles, and only the species in fish.Another example is the structure of the foot, which is of order-determining value to mammals because it is the main tool of mammalian action.But in birds the wings are prominent, and the feet are of very low value as a taxonomic character.However, Cuvier believed that certain characters were related to a certain level of the hierarchical structure.Apparently the subordination of characters is identical to the established weighting system of the botanist except that in animals, in the traditional Aristotelian fashion, "receptivity" is regarded as the highest order, and therefore mainly Traits come from the nervous system. "While Cuvier revolutionized the taxonomy of invertebrates, it was not by introducing new concepts but by internal dissection that he found a whole new set of traits to work with. Another revolution in the use of animal traits did not involve new concepts but came from technological progress: the invention of the microscope.The introduction of optical instruments into natural history by Leeuwenhoek (about 1673) was a great innovation, the full impact of which has not yet been fully revealed (new discoveries made in recent years with the scanning electron microscope illustrates this point).Even the key traits of the Linnaeus system, female and stamen, need a magnifying glass to clearly identify them. The study of the engravings on the elytra or antennae of beetles, the wing veins, and the genital processes of all insects requires at least a magnifying glass.A microscope is essential for studying aquatic invertebrates, not to mention algae, protozoa, and other protozoa. After the 1720s, the study of biology under the microscope was greatly accelerated.Detailed histological observations of organisms of all sizes uncovered taxonomically important sensory organs, corpora, reproductive and digestive appendages, and previously unknown details of the nervous system.Entirely new traits (such as chromosomal differences and biochemical differences) were also discovered in time, also as a result of technological progress.Although the number of traits available to taxonomists has grown rapidly, the amount of information remains insufficient to clarify important debates about phylogenetic relationships. The dogmatic notion that a particular type of trait is best suited for classification was severely criticized during Linnaeus' lifetime.But it is not the weighting principle itself that is under attack but the reasons for weighting.Initially, the only weighting basis that was accepted as a weighting basis of practical value in determining the classification was functional importance.But later a whole new weighting basis was proposed.Lamarck, Cuvier, and de Jussieu emphasized the importance of "constant" characters.De Candolle (De Candolle) emphasized growth symmetry, and indeed growth symmetry in plants can often indicate the characteristics of genera and entire families.This kind of symmetry can often appear in the setting of flowers and leaves and other characters of plants. Essentialists believe that the genus is the totality (summation) of all "species" (ie, subordinate taxonomies) with a common essence; it is the totality of species with certain common "characters" as later taxonomists said.It has been a distressing finding since the earliest days of taxonomy that some individuals or species lack this or that "typical" (that is, fundamental) trait of the taxon.Conventionalists would assign such species to other genera; experienced taxonomists such as Linnaeus would completely disregard such contradictions.Indeed, some higher taxa can only be reliably defined by a combination of traits, one of which may be present outside the taxon or may occasionally not be present in members of the taxon. In such cases a single trait is not necessarily required for membership in the taxon.But it doesn't mean that just having this single trait is enough. It seems that Adanson was the first to recognize this explicitly, although Ray had mentioned it implicitly in some treatises before him. Vicq-dAzyr (1786) once said that "a class may be perfectly natural, though not possessing a single character common to all the species composing it." Heinccke pointed out that herring and sprat have mutual Eight structural traits differed, but only one in ten individuals differed from each other in all eight traits. Beckner first officially confirmed this principle by naming the taxonomic unit based on the combination of traits as "polytypic".But because the term "polymorphic" has been used taxonomically in a different sense, Sneath proposed to use the term "polythetic" instead. Allowing multi-character associations to identify higher taxa spells the end of essentialist definitions.Long before this, however, the entire concept of relying on traits of particular importance (necessary for logical classification) had been attacked, and in due course gave rise to entirely new concepts of classification. 4.9 Upward Classification By Empirical Grouping While European zoologists and botanists were dazzled by the abundance of new families and genera of plants and animals in the tropics, the classification method that had been dominant from the time of Cecatchino to Linnaeus-the descending classification of logical classification-became increasingly inapplicable.The method of classification according to the principle of logical classification hopes to achieve two purposes: to reveal the natural order (God's plan for creating all things) and to provide a convenient identification scheme.In carrying out this method, however, it has been found that these two aims are contradictory, and that the consistent application of the principles of logical classification usually leads to absurd results.A retrospective analysis of this classification theory reveals that it has at least the following three basic weaknesses: (1) When only a few flora and fauna need to be classified, it is sufficient to have an identification scheme, which logical classification can provide.This approach, however, fails to aggregate into the "natural" categories of species and genera that are required for classification when large numbers of flora and fauna are involved. (2) Only one single trait can be used in each step of classification.This trait was selected for its supposed ability to reflect the nature of the "genus".However, the claim that certain traits, such as greater functional significance, are better suited to reflect the nature of a taxon than others is both theoretically and practically untenable.Thus, the entire system of weighting traits according to assumed functional importance lacks plausibility. (3) The entire philosophy of essentialism on which the logical classification method is based cannot be established, so it is not suitable as the philosophical basis for the theory of classification. The radical revolution in philosophical thought that took place in the seventeenth and eighteenth centuries could not fail to influence the thinking of the classified naturalists. The relative influence of the Scientific Revolution and the Enlightenment on some people's minds has been a topic that many historians have grappled with. For example, Locke's philosophical thought that emphasizes nominalism and empiricism; Kant, Newton and Leibniz emphasized continuous thought; Buffon and Linnaeus and their schools of thought.The reason why Buffon ridicules "nomenclature" (he refers to Linnaeus) is the manifestation of the influence of the above-mentioned philosophical thinking. Yet a careful study of the taxonomic writings of the eighteenth century reveals that purely practical considerations played a major, if not dominant, role in forming taxonomic concepts.The practical difficulties encountered with the downward classification became increasingly apparent. A method that compels the famous Linnaeus to "deceive" and classify his species "secretly under the table"; is it not precisely because logical classification cannot solve the problem?How should his inexperienced followers avoid making utterly absurd classifications?In order to understand the nature of the radical change in taxonomy between 1750 and 1850, attention must be paid both to the new demands made by taxonomic practice and to the fact that the philosophical foundations of descending classifications have gradually been eroded. In the end it became clear that it would be futile to try to salvage the descending taxonomy, and that the only way out was to replace it with an entirely different method, the upward method or what is called the compositional method. it.In the ascending approach, similar species are grouped into classes starting from the bottom, and these classes are combined into a hierarchical structure of higher taxa.This method is, at least in principle, strictly empirical. Despite various arguments (see below), this approach is generally adopted by every modern taxonomist, at least at the beginning of the taxonomic procedure. Classification by inspection and categorization, rather than by division, marked a total methodological revolution.Not only is the direction of the classification step reversed, but rather than relying on a single trait, many traits are used and considered simultaneously, or, as some proponents of ascending taxonomies insist, "all traits should be used and considered simultaneously". ". Although the concepts of these two methods are very different.But the replacement of downward by upward was such a gradual process that apparently no one from the end of the seventeenth to the nineteenth century was fully aware that this change was taking place. There are several reasons for the gradual nature of this change.First, the method of distinguishing objects by "inspection" is of course not at all a new invention.Aristotle had long used association characters to divide his higher taxa.If you are sorting a basket of mixed fruit, you can check them into apples, pears and oranges without much effort.Such preliminary sorting was evidently done by all early botanists, even those who professed to practice logical classification. Bock and Baohin did it openly, Cechapino, Tiu Lev, and Linnai did it covertly.It is clear that a certain degree of upward classification has been blended into the downward from the outset. (Conversely, after the negation of logical classification, some parts of it were preserved because of their usefulness for identification.) There must be several prerequisites for this transformation (but the history of this transformation has not yet been thoroughly analyzed. Pass).First of all, it is necessary to understand what is being classified (here refers to the species) in order to carry out upward classification.Therefore, a prerequisite for ascending classification is understanding species, even those defined by essentialism.Early herbalists and other scholars before Linnaeus sometimes generally regarded all species or varieties in a genus as complete species (pure species), so they would have considerable difficulties in the face of ascending methods.The natural history tradition developed in the seventeenth and eighteenth centuries made a key contribution here (see Chapter 6).The second prerequisite is the weakening of the essentialist influence already mentioned.Finally, an empirical attitude arose during this period, one that favored results over principles, partly as a result of the decline of essentialism. In the 1680s, three botanists advocated the method of classifying species based on joint characters (multiple characters).The British botanist Morison once used multiple characters. Ray said that in order to make a correct inference about the nature of the genus, "there is no more definite evidence than having several common properties." He repeatedly emphasized in 1703 In view of this, "the best scheme of arrangement of plants is that of all genera, from the highest to the lowest, having some common character, or agreeing in several parts, or incidental to it (Plant Research Act: 6-- 7). Around the same time the French botanist Magnol (1689) objected to the method of dividing higher taxa by logical classification. In order to deduce the relationship, he used not only the character of fruiting but also the characters of the various parts of the plant.More importantly, he places particular emphasis on the importance of a holistic approach.That is, "by inspection" to classify species: even in many plants, there is a certain similarity, which is not composed of individual parts, but as a whole; this is a very important and inexpressible Similarity (ProNomus, 1689). Masnol is particularly important in taxonomy because he had a major impact on the formation of Adamson's ideas.His objection to the division of characters into essential and accidental characters (as required by essentialists) was adopted by Adanson and the entire empiricist school, though he did not pay much attention to it. Buffon attaches great importance to reverse classification by inspection. He once said (Oeuvr.phil, 1749: 13) "As far as I can see, the only way to devise a beneficial and natural method is to classify things that are so similar. separate things that are different from each other." He also emphasized that all traits should be considered. Merrem, Blumenbach, Pallas, Illiger, Meckel, and other zoologists adopted his advice (Sttesemann, 1975: 107). The first person who had the rational courage to stand up and openly doubt the correctness of logical classification was Michel Adanson (1727-1806).In his Natural Classes of Plants (Les families naturelles des plantes, 1763) he proposed to replace logical classification by empirical induction, "because only one part or a few parts of plants are considered methods are arbitrary, hypothetical, and abstract. These methods cannot be natural... The only natural method in botany is one that takes into account all similarities grouped together and divided into classes and families." Adanson went a step further to develop a detailed method for examining taxonomic characters. After the logical classification of single traits was denied, new problems were raised.If classes are based on more than one trait, how many traits should be used?Should certain traits be preferred?Adanson was the first botanist to explore these questions fairly systematically.In order to find out what effect the selection of traits will have on the classification.Adanson experimented with dividing plants into 65 artificial categories based on their special traits (such as corolla shape, seed position, and thorns).According to the test results, he believed that it is impossible to obtain a satisfactory natural classification system based on a single trait or a combination of only two traits.Adanson is sometimes called a numerical taxonomist (first proposed by Adrien de Jussieu in 1848) because he calculated the proportion of natural combinations that would result from each such arrangement.This is very wrong.Because Adanson did not use this mathematical method in the actual division of genera and families.When he divided the genera and families, he observed these categories by eye according to the method of Magnol.Although he first distinguished between genus and species, "the general examination of these comparative descriptions led me to realize that plants naturally divide themselves into classes or families" ("Natural Classes of Plants", 1763). Adanson clearly sees that different traits have different taxonomic meanings. “对一切性状予以同样的加权将会和阿丹森的归纳方法发生逻辑上的矛盾。这样的武断方法将意味着对性状在事前就作过评价”（Stsfleu，1963：201；Burtt，1966）。阿丹森所提倡的是对植物一切部分的可能考虑，而不单单只考虑结实。他特别强调两点：（1）某些性状对改进分类无关，可以忽略不顾；（2）具有最大信息量的性状在各个科之间各不相同。每个科都有自己的“天才”（genie）。 阿丹森的一些反对者批评他的“特殊”理由是他的方法所要求的植物知识太多。如果鉴定是分类的唯一目的这将是一种合理的批评，但是正像系统学史一再证明了的那样，令人满意的分类，以对一切证据作批判性评价为基础所作出的分类，只能由对有关类别透彻了解的人才能作出。可以把阿丹森对性状的态度总结归纳如下：他的确赞成性状加权，但这种加权并不是基于任何先入之见或既定原则（例如生理重要性）之上而是依据将事前通过检查已经确定了的门类加以比较的一种凭经验的方法。 由阿丹森所提倡的几乎每一项原则现在已经成为分类学方法论的一部分。然而在托马斯逻辑盛行和林奈的权威占统治地位的尉代，阿丹森几乎完全被忽视。现在很难说他的《植物类别》在当时究竟有多大影响。这书虽然得到林奈的赞赏，但是确实受到它的影响的人，例如A． L． de Jussieu在谈到他们的观念的源流时却没有雅量提起阿丹森。 后来的分类学家在实际工作中也得出了相同结论，但这并不是由于研究阿丹森大部分已被遗忘的着作的结果，而是独立地和凭经验得到的。几乎一直到一百年以后阿丹森的卓越贡献才重新被人们发现。 4．1O过渡时期（1758—1859） 林奈的《自然系统》第十版出版（1758）后的一百年是分类学活动空前活跃的时期。 这些活动大多数是由于林奈对多样性研究赋予的极高声誉促成的。由于发现的生物越来越多，有越来越多的年轻人成为了动物学家和植物学家。寻找新种及其分类的热潮威胁到生物学其他学科的发展。例如，Kolreuter和Sprengel对花的生物学的着名研究因为不能发现新种而被忽视。不是分类学家的内格里面对这股热潮曾为生物学的其他领域正在被“系统学潮流”所淹没而深为不安。 私人和国家收藏中动植物标本的大量积累在分类学这一行职业中引起了深刻变化。 分类学家愈加职业化和专业化。创办了许多新杂志以适应记叙大量新种的需要，一些业余爱好者发现专门研究一个单独的科能够使他们在能力和资格上达到高水平。每年的分类学研究论着稳定上升。 分类学的研究领域大大扩展。直到当时动物学的主要兴趣限于脊椎动物，植物学则限干研究显花植物。过渡阶段中动物学的兴趣已延伸到无脊椎动物，特别是海洋无脊椎动物，最后，甚至由Sars开始转向深海动物。植物学家也转向更加注意隐花植物。 正是在阿丹森的《植物类别》（1763）和达尔文的（1859）出版之间的这个阶段下行分类才逐渐被上行分类取而代之。法国在欧洲国家中是最少受到本质论影响的国家，率先地引用了分类学的新方法。这不仅明显地表现在Magnol，布丰和阿丹森早期的开拓性工作中，而且也反映在拉马克和居维叶的着作中。拉马克（1809，1815）虽然仍然依附于早已过时的哲学思想，但在分类上却按归类法而不是按逻辑分类；居维叶的部分相关原则大大加强了分类单位多性状概念化和寻找新性状的趋势。这样就在动物学中引进了一种新的实用主义传统。其中性状是由对形成似乎是“自然的”归类作出贡献的能力来评价，也就是说性状是由结果追溯到原因按经验来评价。而且认识到性状的相对重要性可以因高级分类单位的不同而发生变化，即性状的分类（学）价值不是绝对的。 这种情况也导致了分类阶元的重新概念化。阶元已不再被认为是逻辑分类中的步骤（从总类到最低级的种）而是等级结构中钩等级。属成为了集合性高级阶元，因而它在本体论上和认识论上与逻辑分类的本质论的属完全不同。属在涵义上和作用上的这种变化常常被分类学家和哲学家忽略掉，从而引起了误解和概念混乱。 另外，在不同阶元的相对重要性上也发生了微妙变化。林奈把属看作是宇宙的中心。 由于不断地发现新种，属就变得越来越大，大多数的属必须一再细分，着重点转移到比较高一级的科上。在很多（但不是全部）类生物中，科成为最稳定的分类单位。 由下行分类向上行分类转变（联同有关的方法论和概念上的变化）和几乎所有的科学“革命”一样是缓慢的、渐进的和参差一不一的。正如前面所指出的，Magnol（1689）开始着重科，Bauhin，Morison，瑞，Magnol，悌宇列弗等早已运用多个性状（往往来自不同的器官和系统）进行分类，虽然还多少有些犹疑。Bauhin（1623）首先“按照它们的自然相似性”将植物归类以进行分类。然而上述这些学者在做法上也并不是前后一贯的，特别是因为他们常常在不同程度上把分类当作鉴定体系来运用。 Stafleu（1963：126）正确地指出亚里斯多德逻辑分类的声誉之所以逐渐下落不仅是由于实际原因，而且也由于笛卡尔及其原则的影响。例如阿丹森就按照笛卡尔的有顺序的四条基本规定：怀疑、分析、综合，详叙来塑造自己的方法。笛卡尔的影响以及牛顿和莱布尼茨的影响（通过布丰）是为什么林奈对法国的影响不及其他国家的原因之一。 林奈的很多实用性改革（如双名法，命名法则等等）当然最终都被采纳了，但是他的亚里斯多德主义则只是作为一种方便的鉴定方法而不是作为分类的完善哲学基础被人们接受。林奈以后的分类学最引人注目的发展可能就是。分类越来越清楚地等级结构化了（见下文）。 植物分类学从切查皮诺（1583）到林奈的两百年间极其繁荣。到了林奈以后的阶段虽然继续稳步发展，却不是那样壮观动人。这段时期有三个特点，最重要的是努力于建立植物的“自然系统”（即使到了今天也尚未完全达到）。在这方面De Jussieu，deCandolle，Endlicher，Lindley，Bentham，Hooker等人都作出了自己的贡献。其次，对隐花植物日益重视，不仅是蕨类，藓类，而且包括真菌，藻类和单细胞水生植物（原生生物）。第三，植物学的专业化也有所发展（尽管植物学家专业化的程度远不及动物学家），研究植物特殊门类的专着先后出版，导致了对植物界某些门类的深入研究。 有一项很少被人注意的发展是在这个阶段中动物分类学成为了理论动物学的一个主要分支。像Siebold，Leuckart，EhrenbergSars，Dujardin等（甚至还可以将达尔文也列进去）博物学家开始时是分类学家，但是后来对作为整体的活的动物深感兴趣，对普通动物学作出了重要贡献。例如研究清楚了寄生虫的生活史、世代交替、海洋无脊椎动物动体期顺序、内部器官的结构与功能以及与活的动物有关的各个方面。上述这些研究往往可以确凿无疑地证明是直接源于分类学研究，然而分类学在生物学中的开拓作用却很少得到应有的重视。例如，只是近年来才充分认识到达．尔文关于蔓脚类动物的专着对其进化学说的发展与进一步完善是多么重要（Ghiselin，1969）。 下行分类的绝大多数支持者都充分地觉察到按照他们的方法作出的分类是“人为的”。林奈在他的一些着作中曾经提到作出真正的“自然的”分类（按他的解释）的时刻尚未到来并为之表示惋惜。他有几况发表过植物“自然”分类的片断（Stafleu，1971），而且无论他的某些主要分类是多么人为，其中他为大多数属所作的归类和现代进化分类学家所能做到的非常相似。但是仅仅用上行分类代替下行分类并不足以产生自然分类。必须有某些组织原则，某些基本概念作为分类学家的行动纲领。 自从古希腊时代以来有一种流行的信念，认为自然（界）的多样性反映了某种更深层的秩序或和谐。就我们从留传下来的极少资料所能判断的，没有其他哲学学派比毕达哥拉斯学派对这种和谐更加关注。自然神学复活了自然和谐（式）平衡的概念并且随处可见它的标志：各种适应的明显“设计”。但是乍一看来多样性十分混乱，似乎和这种哲学并不调合。到了林奈以后时期当已知的动植物的种和高级分类单位的数目几乎按指数增长时情况就变得更为严重。面对层出不穷复杂纷坛的新种人们不禁要问：“博物学家梦寐以求的自然和谐在什么地方？控制多样性的规律是什么？造物主在设计大大小小的各种生物时有没有什么计划？” 在自然神学全面统治的时代简直不可想像有机界的多样性竟然如此杂乱无章不可理解，竟然纯粹是“偶然”结果。因此分类学家的任务便是去寻找控制多样性的规律，或者说去发现造物主的计划。 “自然（的）系统”就是能最圆满地反映造物主计划的分类系统，每个博物学家的理想就是得到这种系统。但是在研究不同的学者，使用“自然的”这个词，思想中指的是什么时就会发现答案是多种多样的。对这个词的用法加以讨论将有助于了解这个时期的思想动态。为了理解透彻起见，下面在介绍这词的几种意义时也顺便提到它们的反义腻（1）“自然的”和“偶然的”相反，它是反映真正“本质”（nature亦即essence）的。本质论者的分类学家，从切查皮诺到林奈都是按这个意义试图提出自然分类（Cain，1958）原则上这就是林奈的理想，而且似乎也是他对其人为的性系统表示不满意时的内心思想。对林奈来说“自然的”所表示的意义和我们现在所理解的完全不是一回事。因为对他来说，种、属或高级分类单位的“本性”（nature，基本性质）就是其本质（essence）。林奈学派的所有学者都同意这一点（Stafleu，1971;Larson，1971）。 绝对不要忘记林奈认为属和高级分类单位作为造物主的创造代表了不变的本质。只有充分地认识了这些本质以后才能真正理解属和高级分类单位。Cain（1958：155）指出：“可能林奈认为“自然（的）”系统就是显示事物的本性的系统，本性实际上就是本质”。认清这一点有助于理解他对“自然方法”（意即“自然系统”）的论述。 林奈的关于纲和属起源的学说是彻底的神创论学说。由以上所有的这些可以明显看出当谈到“自然系统”时林奈头脑中真正想的是什么：在这个系统中，不是根据整体相似性为高级分类单位直觉地下定义，而是确定这些分类单位的真正本质。当然，在林奈的追随者之中，“自然系统”这个术语已逐渐取得了完全不同的意义。 （2）随着本质论哲学势力的削弱，“自然的”这个词开始指“合理的”（而不是指多变的）而言。这种解释反映了十八世纪流行的思想和态度。这就是认为自然秩序是合理的，可以通过理性推理来认识和理解。自然界的一切事物遵从神赐的规律，自然秩序和神的计划一致。“自然系统”（如果能找到的话）将反映造物主创造的蓝图（Azassiz，1857）。（3）另外还有些人把“自然的”看作是“经验的”而和“人为的” （即完全功利主义的）相对映。按这一概念自然分类将符合John Stuart Mill的要求： “当某种科学分类较之其他分类方法能提出更多、更重要的归类方案时科学分类的目的就充分地达到了”。阿丹森基本上就是按照这种观点进行植物分类的。这是由Bauhin倡导的一种传统，虽然Morison和瑞并不是十分热心地支持这一传统，但Magnol则是毫无保留地支持它。（4）1859年以后，“自然的”这个词用来表述分类系统时指的是“具有共同起源”。达尔文以后的自然分类系统是，其中一切分类单位是由一个共同祖先的后代所组成。 上面列出了“自然的”这个词的几种不同涵义，但并没有毫无遗漏地阐述这段时期中所提出的各种分类的概念基础。对自然（界）和谐或计划的探索还受到其他一些概念的影响（其中有一些在前面曾经部分地涉及到），特别是在连续的几个时期中一直广泛流行的下述三种概念。 几百年来完善（性）等级（scale of Perfection）似乎是将秩序引进多样性的唯一可行的方案。布鲁门巴赫（Blumenbach）和其他很多学者一样认为自然阶梯是自然系统的牢固基础，自然系统可以“将自然物体按照它们最大最多样的相似性加以安排，把相似的归在一起，把彼此不相似的挪开。”拉马克，特别是在他的早期着作中，也表示了类似看法。自然阶梯的观念在植物学家之中并不普遍，因为在植物中趋向干完善性的倾向很少发现，除了从藻类及其他隐花植物进化到显花植物而外。因此林奈将分类比作一幅地图，其中每个国家和其他几个国家毗邻衔接。 有人认为从最不完善的原子到最完善的生物、人之间有一连续序列。随着对多样性的认识逐步深入，这种看法也日益受到怀疑。林奈这时已不再坚持无机世界与有机世界之间的任何连续性。虽然他还不时地提到自然发生。人们特别注意对所谓的“植形动物” （珊瑚，珊瑚虫等等）的研究。它们果真是动植物之间的中间型吗？如果不是，那么它们究竟是植物还是动物？当Trembley于1740年发现绿色水螅（Chlorohvdraviridissima）确实是具有叶绿素的动物，并具有前此一直认为只限于植物的强大再生能力时在学术界引起了极大的震动。随后不久，Trembler又证明珊瑚和苔藓虫也是动物，并不是动物和植物的中间型。拉马克在动物相似性不同系列中所承认的许多分岔也和单一的完善等级相矛盾。 当居维叶（1812）强调动物只有四个不同的门（一个不多一个不少）而且在它们之间绝对没有任何联系时它（自然阶梯）遭到了最后的致命一击。居维叶之后虽然有时在较小类别的成员之间仍然可以察知相似性，但是作为“增长中的完善性” （growingperfection）排列原则却已不再适用。而且认为在不同的类别之间彼此有联系的主张也越来越不能令人信服。看来曾经一度以自然阶梯为标志的有机界的统一崩溃得越彻底，人们对生命世界脚理解也就越透彻。当人们认识到一维的线性指导原则不适用时就开始探寻多维的方案（multidimensional scheme）。 将某类生物安排在完善等级的什么位置上由它和相邻的较不完善与较为完善的同类生物和相似性（affinity）决定。现代生物学家很难想像进化论以前的着作中的相似性究竟指的是什么；或许指的就是类似性（similarity）。但是过去认为这种类似性反映了某种因果关系，例如表现在自然阶梯中或者像阿伽西所说的造物主的创造蓝图。 看来自然阶梯之所以令人难以理解是由于有两种类似性，一种是真正的相似性，另一种是由Schelling，Oken等所命名的同功（analogy）。企鹅与鸭或海雀有关，是由于真正的相似性，而企鹅和水生哺乳动物（如鲸）有关则是由于同功。鹰和鹦鹉及鸽子显示相似性，而和哺乳动物中的食肉动物则是同功。虽然自然哲学派的某些思想显得希奇古怪，但是他们将“相关”划分为相似与同功却对生物学随后的历史很重要。RichardOwen就是在这个基础上发展了他的同源（homology）和同功概念。自此以后这两种概念在比较解剖学上就占有主导地位，特别是这两个词按进化论思想再下定义之后。 但是在进化论提出之前怎样才能将相似和同功的思想转化到分类系统中去？在这里自然哲学派借用了毕达哥拉斯学派的思想，没有什么比数字能更好地表示自然的和谐与规律。昆虫学家W·S·MacLeavy（1819）选用了5这个数字，虽然随后也有人试用过3和4，但5这个数字还是最通用，被称为五元系统或五元论（quinarian srstem）。 MacLeay认为所有的分类单位都排列在圆上，每个圆有五个单位，相邻的圆彼此接触（“密切”）。同一圆上的分类单位显示相似（性），和其他的圆的关系则表示同功，因此，分类学家的任务之一就是探寻这种交叉关系。 虽然这类古怪的方案遭到严肃博物学家的苛刻批评，但是五元论者却值得同情。他们试图探索自然界多样性的规律，而在进化论之前数字方案似乎是最现成的。甚至赫胥黎也一度热衷于五元论并多番试图将无脊椎动物的高级分类单位排在圆上或平行的行列上（Winsor，1976b）。19世纪40年代和50年代五元论在英国十分流行，连达尔文也曾认真考虑过。因为如果生物真的是按五的倍数排列，那就表明这种排列是由超人设计的。 因而由自然选择所形成的多样化就须排除。但是用不着深入研究，达尔文就发现分类学上的多样性和任何算八字的方案是不相容的，特别是他对藤壶的研究并没有为五元论提供任何证据。 即使反对五元论的人也不能不承认有几种类似性。除了相似和同功以外，有时还可以发现仅仅“外表类似”和某些其他种类的类似。在所有这些类似性之中相似是最令人费解的，然而当时广泛地认为“它（相似性）是造物主所据以创造万物的生命法则的直接结果”（Sttickland，1846：356）。这就是阿伽西为什么把相似性当作是有造物主存在的最强有力证据之一的原因。 自然阶梯、自然哲学派的方案以及毕达哥拉斯式的算八字方案的相继失败对分类学产生了十分严重的后果。在“物种起源”出版以前的五十年中，绝大多数分类学家完全回避作理论上的探讨。当采用上行分类原则时也只是满足于将表面上类似的种和属；归类到一起的简单实用主义的做法。 在这个阶段中很少概念性的进展。即使居维叶也只是在他的晚期着作中重复他二十年以前所阐述的原则。在植物学方面情况也好不了多少。A·P·德坎多尔（1778－1841）的《基本原理》（Theorie elementaire，1813），虽然遭到反对，仍然坚持传统本质论的从原因推及结果的先验方法。但是实际上分类学者本人也没有注意到，用单一性状确立高级分类单位的方法正在被用联合性状的方法代替。上行分类变成理所当然的。 “从底层”开始（分类）大大促进了专业化的发展（Lindroth，1973）。 采用新办法的结果，发现很多前此认定的分类单位实际是很混杂、很不均一的。例如Meckel（1821）和Leuckart（1848）证明居维叶按关键性状“辐射对称一两侧对称” 区分的“辐射对称动物”（Radiata）是由两个非常不同的门：棘皮动物和腔肠动物拌合在一起的非自然集群。从门一直到属前此已确认的高级单位的每个层次如果发现是非自然集群就重新审查并划分成更均一的分类单位。到了1859年动物的大部分分类单位都被重新划分过，特别是对按结构和生物学性状一致的种的归类。 当发现某些过于相信外表类似的现象后，这种不受学说约束的单纯实用主义态度受到一定影响。当然早就知道毛虫和蝴蝶是同一动物，但是随着对分类的兴趣增高人们不禁会问，依据毛虫的分类是否会和根据蝴蝶（是由毛虫变态所形成的）的分类相同？19世纪前半期发现很多无脊椎动物都具有这样的变态现象；对绝大多数海洋座生动物类这是一种正常现象。从一开始藤壶就分在软体动物或其下属的介壳动物中。当J．V．Thompson于1826年5月8日观察到附着在玻璃容器底部的甲壳动物的动史变成了藤壶（Winsor，1969）时引起一时轰动。进一步研究后确凿无疑地证明藤壶是座生的甲壳动物。而且Thompson以及其它研究海洋生物的学者还发现许多浮游生物只不过是一些无脊椎动物的幼虫阶段，甚至自由生活的甲壳动物也可以经由几个幼虫阶段的变态（无节幼虫期，妇状幼虫期，介形幼虫期）。 各种生物既可按主要功能（居维叶）又可根据具体的构成计划（冯贝尔及自然哲学派主张的Baunlan）的决定来分类的思想由于19世纪前半期的两项发现而弄得混乱不堪。 第一个是发现了涤虫、吸虫及其他内寄生虫的复杂生活史，虽然来自同一遗传型，但其世代交替的不同阶段，例如囊尾蚴和涤虫，尾蚴和吸虫，却如此不同以致令人怀疑完全按外表类似的方法进行分类是否正确，另一个更加惊人的发现是Adelbert vonChammiso（1819）发现萨尔帕（salps，一种被囊类动物）以及Michael Sars（1838－1846）和J．J．SteenstruP（1842）发现腔肠动物都具有正常的世代交替。单一的物种具有如此不同的自由活动世代和座生世代在此以前一直将之安排在完全不同的分类单位中（Winsor，1976b；Churchill．1979）。这种情形在植物界中也并不少见，其中隐花植物的配子体和孢子体通常都是完全不相同的。 所幸的是，这些多少没有弄清楚的发现并没有引起另一番形而上学的推测而只是促使分类学家加倍努力将“相关”生物的“自然”类别归类到一起。这种努力几乎自动地产生了从属阶元的分类，现在通常称之为林奈的等级结构。下面介绍分类学说中等级结构的涵义及有关问题。 4．11等级结构分类（Hierarchical classlfications） 绝大多数分类，无论是对无生物还是对生物，都是等级结构分类。既有高级、低级阶元，也有高等、低等级别。“等级结构”这个词的用法有时常含糊不清。等级结构有两类：排它性等级结构（exclusive hierarchy）和内涵性等级结构（inclusivehierarchy）。军阶，如士官、尉官、校官、将官，是排它性等级结构的典型例子。比较低级的军阶并不是高级军阶的再分而成的部分，因此校官并不是将官的再分部分。从十六世纪到十八世纪支配分类学的“自然阶梯”也是排它性等级结构。每一个完善性层次是由等级结构中的较低一级的层次上升而成的，但并不包含较低一级层次。 从切查皮诺到局维叶所提倡的功能等级结构也是排它性等级结构的例子。在功能等级结构中生长占有最高级别，生殖为次高级别，这并不意味着生殖是生长的再分部分。 现代分类（学）阶元的等级结构是典型的内涵性等级结构。这可以用具体例子说明，似狗的物种，如北美的土狼，非洲的胡狼等都归类为犬属（Canis），各种似犬的属和似狐的属合并为大科（Canidae）。犬科连同熊科，猫科，鼬科以及其他有关的科联合成食肉目（Carnivora）。纲，亚门，门，界是这一等级结构连续的较高级分类单位。 每一较高级的分类单位包含较低级的、从属的分类单位。 就理论上来说，内涵性分类等级结构既可由下行分类（逻辑分类）产生，也可以由上行分类（合成分类）形成。然而从历史上来说，逻辑分类从来没有导致明确的分类等级结构，因为每一层次都是分别处理，每个“种”（按逻辑分类规定）在下一个较低级等级结构层次又成了“属”。悌字列弗及林奈都以属为中心，在很大程度上把属稳定了下来，他们只确认了两个较高级阶元并且对之未予重视（见前）。 林奈首先合理地运用了较高级阶元，从总体来说也是一贯的。但是他的思想还是过多地受逻辑分类原则的支配，因而不能制定出完全一致的包含一切生物的内涵性等级结构。直到1795年至1815年这二十年之间才真正发生了变化（Winsor，1976b：2—3）。 这种概念性变化主要应归功于巴黎博物馆的一些着名分类学家。但是不同的学者采纳这种新思维方式的程度也各有不同。例如拉马克对较高级分类单位的安排仍然代表了一种严格的排它性等级结构，即使在较低级阶元层次上也包含有合成分类（也就是内涵性）因素。居维叶的四个门并没有等级结构关系，如果有，也是排它性的。但是在较低级阶元上也可发现有某种内涵性等级结构倾向。 按合成法程序建立内涵性等级结构的方法很重要，这有很多理由。很明显，除非有了分类阶元的内涵性等级结构，像达尔文1859年提出的共同起源学说是不会有任何人想起的。更直接有关的是，新观点提出了建立内涵性等级结构应根据什么原则的实际问题和概念性问题。这是特别棘手的问题，因为绝大多数分类学家的思想仍然受到（如果不是完全受支配的话）“自然阶梯”这种排它性等级结构思想的影响。 早在十七世纪及其后关于高级阶元的“真实性”问题，即高级阶元是否实际存在的问题。一直有着争议。本质论者像林奈坚持认为最低限度以其本质为特征的属是“真实的”。以布丰为首的具有唯名论倾向的分类学家则旗鼓相当地坚持只有个体存在，最低限度像属，科，钢等高级阶元只是约定俗成的称谓。十七世纪时没有任何两位植物学家作出相同的分类这一事实，似乎肯定地支持了唯名论者的观点。为了区别抽象的意念和具体的真实（实际），布丰曾为解决这个问题打下了基础，但是争论仍然又持续了两百多年。 为什么这场争论历时这样久原因主要是术语上的混乱。“阶元”（category）这个词在使