rjohara.net

Search:  


This is a web version of a previously published paper. The text is reproduced here in its entirety but no figures are included. Editorial insertions in this web version are enclosed in {braces}. Citations to this paper should refer to one of the printed versions:

O’Hara, Robert J. 1991.
Representations of the natural system in the nineteenth century. Biology and Philosophy, 6(2): 255–274. [Reprinted 1996 as pp. 164–183 in: Picturing Knowledge: Historical and Philosophical Problems Concerning the Use of Art in Science (B.S. Baigrie, ed.). Toronto: University of Toronto Press.]


Representations of the Natural System in the Nineteenth Century1

Robert J. O’Hara
Division of Birds
National Museum of Natural History
Smithsonian Institution
Washington, D.C. 20560, U.S.A.

Abstract

‘The Natural System’ is the abstract notion of the order in living diversity. The richness and complexity of this notion is revealed by the diversity of representations of the Natural System drawn by ornithologists in the Nineteenth Century. These representations varied in overall form from stars, to circles, to maps, to evolutionary trees and cross-sections through trees. They differed in their depiction of affinity, analogy, continuity, directionality, symmetry, reticulation and branching, evolution, and morphological convergence and divergence. Some representations were two-dimensional, and some were three-dimensional; n-dimensional representations were discussed but never illustrated. The study of diagrammatic representations of the Natural System is made difficult by the frequent failure of authors to discuss them in their texts, and by the consequent problem of distinguishing features which carried meaning from arbitrary features and printing conventions which did not. Many of the systematics controversies of the last thirty years have their roots in the conceptual problems which surrounded the Natural System in the late 1800’s, problems which were left unresolved when interest in higher-level systematics declined at the turn of this century. [Keywords: systematics, evolution, history, phylogeny, ornithology, diagrams, representation, natural history, natural system, taxonomy, classification.]

Introduction

Naturalists try to arrange the species, genera, and families in each class, on what is called the Natural System. But what is meant by this system?

—Darwin (1859: 413)

The Natural System—the idea of the order in living diversity—is one of the great theoretical conceptions in the history of science. Although systematists—those who study the Natural System—have not always been able to agree upon ‘what is meant’ by this conception, they generally have agreed that the results of systematic research are best presented diagrammatically. In proposing his ‘mapmaking’ approach to systematics, for example, the British naturalist Hugh Edwin Strickland (1811–1853) observed that

The true order of affinities can only be exhibited (if at all) by a pictorial representation on a surface, and the time may come when our works on natural history may all be illustrated by a series of maps on the plan of those rude sketches which are here exhibited [Strickland, 1841: 192; italics in the original].

Alfred Russel Wallace (1823–1913), a promoter of Strickland’s methods, also wished

that in every systematic work each tribe and family should be illustrated by some such diagram, without which it is often impossible to tell whether two families follow each other because the author thinks them allied, or merely because the exigencies of a consecutive series compels him so to place them [Wallace, 1856: 207].

And indeed the only illustration in Darwin’s Origin was his well-known diagram of an evolutionary tree, illustrating the theoretical structure of the Natural System. ‘The accompanying diagram’, he wrote, ‘will aid us in understanding this rather perplexing subject’ (1859: 116).

In a previous paper (O’Hara, 1988b) I defined three periods in the history of Nineteenth Century systematics. The first of these, the quinarian period (1819–1840), was embodied in the writings of William Sharpe Macleay (1792–1865), Nicholas Aylward Vigors (1787–1840), and William Swainson (1789–1855). Quinarian systematists believed that two sorts of relationship—affinity and analogy—obtained among taxa, that taxa existed in natural groups of five, that circular chains of affinity connected taxa within each group of five, and that relationships of analogy obtained among taxa occupying corresponding positions in different circles of affinity. During the subsequent mapmaking period (1840–1859) Strickland and Wallace reacted against the quinarians, and argued for the exclusion of analogy and symmetry from the domain of systematics. They promoted an empirical approach to systematics which compared the reconstruction of the Natural System to the geographical surveying of an unknown territory. Finally, during the evolutionary period (1859–1901), a variety of systematists explored, to varying levels of depth, the implications that the new doctrine of evolution held for their discipline. In the present essay I will discuss diagrams from all three of these periods, but I will focus less on their temporal development (as I did in my previous paper), and more on the representational elements which vary among them. As Mayr has rightly said (1982: 144), ‘The most important aspect of the history of systematics is that it is, like the history of evolutionary biology, a history of concepts rather than facts.’ I hope that this survey will encourage others to investigate these issues in greater detail, and that it will alter the mind of any who may still believe that the history of systematics is a history of classifications and nomenclatural technicalities.

I have selected ten diagrams to analyze here, and have arranged them chronologically as Figures 1–10; all of these diagrams are ornithological, but none of them appeared in my previous paper, and most are being reproduced here for the first time since the Nineteenth Century. Other studies which have examined systematic diagrams (most of them non-ornithological) include Wilson & Doner (1937), Voss (1952), Greene (1959), Barrett (1960), Stresemann (1975), Winsor (1976), Nelson & Platnick (1981), Stevens (1982, 1984), Reif (1983), and Gaffney (1984).

Elements of the Natural System

The elements of the Natural System that I wish to consider are affinity, analogy, continuity, ‘directedness’ in its various forms, symmetry and predictivity, reticulation, branching, and dimensionality. The methods by which these elements of the Natural System were recognized or discovered by investigators is a fascinating but entirely separate matter, and is beyond the scope of this survey.

Affinity. Affinity is in many ways the core concept underlying the idea of the Natural System, and in the pre-evolutionary systematic literature the term affinity denoted a relationship based on some sort of essential similarity. While the later and somewhat related concept of homology was rarely discussed in the purely systematic literature (homology was a relation that obtained among characters, in contrast to affinity which obtained among taxa), discussions of affinity pervaded that literature. Vigors titled his quinarian study of bird systematics, from which Figure 1 is taken, ‘Observations on the natural affinities that connect the orders and families of birds’ (Vigors, 1824); Strickland published ‘Observations upon the affinities and analogies of organized beings’ (1840), and even included a ‘Scale of degrees of generic affinity’ on his diagram of kingfisher systematics (Strickland, 1841), reproduced in my previous paper (O’Hara, 1988b: 2751). It is easy to understand how Strickland was able to compare the Natural System to a map, because the language of affinity was almost always spatial language: taxa were spoken of as being ‘close’ to one another, or ‘far apart’, or as ‘approaching’ other taxa.

In the evolutionary period affinity came to be viewed by some authors as genealogical relationship on a tree, but the traditional spatial language continued to be used in many cases, and map-like illustrations of the Natural System, depicting affinity in the old sense, often existed side-by-side with tree-like illustrations depicting genealogical affinity (compare Figures 8 and 9, in which the map-like view is represented as a cross-section of the tree). Although it not as popular today as it was in the Nineteenth Century, the term ‘affinity’ is still used by some contemporary systematists (for example Harrison, 1969; McGowan, 1982; Olson, 1987).

Analogy. Affinity was not the only aspect of the Natural System for many Nineteenth Century systematists, however. William Swainson and the other members of the quinarian school, followers of the entomologist Macleay, considered analogy to be equally important:

[W]e shall consider that to be a natural system which endeavours to explain the multifarious relations which one object bears to another, not simply in their direct affinity, by which they follow each other like the links of a vast chain, but in their more remote relations [analogies], whereby they typify or represent other objects totally distinct in structure and organization from themselves [Swainson, 1835: 197; italics in the original].

Figure 2 (from Swainson, 1837) illustrates both the circular affinities of the starling and crow families, and also the analogies between them, analogies which connected every circle of affinity in the quinarian system.

Strickland and his followers in the mapmaking period explicitly denied that analogy had any place in the Natural System (Strickland, 1840, 1841), and did not depict it in any of their systematic maps. Similarities among taxa showing little affinity to one another were undeniable, however, and the acceptance of evolution allowed systematists to in some measure reintroduce the depiction of analogy (under the name of evolutionary convergence) into their systematic diagrams. The hoatzin, for example, a South American bird of the family Opisthocomidae, has features in common with both the galliform and cuculiform birds, and Maximilian Fürbringer (1846–1920) could depict this evolutionarily in 1888 by showing the branch of the Opisthocomidae emerge from the galliform section of his tree, but then continue upward, cross into the top section of the diagram, and end near the cuckoos (Figure 8, upper left). The representation of evolutionary convergence in this manner has extended well into the Twentieth Century (see for example Mayr, 1969: 227).

Continuity and Directedness. Continuity and ‘directedness’ in its various form (progress, advancement, time, evolution) are among the most complex notions connected with the Natural System, and they run into many of the larger currents of Western thought. Lovejoy’s classic work The Great Chain of Being (1936) treats extensively of the notion of continuity in the period preceding the one considered here, and provides important philosophical background, because the systematic diagrams of the Nineteenth Century are in a very real sense the wrack of the Chain of Being (see particularly Stevens, 1982).

In the quinarian period there was a clear belief in the Natural System’s continuity. Vigors complained that

By an oversight of the printer’s, the circles in [Figure 1] were not made to touch each other … and they thus seem to convey an erroneous idea of the series of affinity being incontinuous [1824: 509].

And Swainson, in his Preliminary Discourse (1834: 228–235), discussed at length the ‘law of continuity’. For the evolutionists, beginning with Wallace, continuity was transformed from an abstract philosophical principle into a reflection of the real and physical connections of evolutionary genealogy, and it was manifest not only among the living taxa of the present, but more importantly between living taxa and their ancestors. And in this sense, continuity remains a central element of the Natural System for evolutionary biologists: ‘relationship’ is defined today, at least by most cladistic systematists, as the relative recency of genealogical continuity among taxa which are now reproductively isolated.

I have argued (1988a) that belief in any sort of directedness in the Natural System, apart from that of time itself, is mistaken, and the product of an inappropriately narrative way of viewing the world. Meta-temporal directedness was, however, and continues to be, an important element in many systematic representations. Pre-evolutionary quinarian diagrams by virtue of their circular nature do not exhibit strong directionality, although the arrangement of taxa into upper and lower circles is not likely to have been accidental. Wallace (1856) referred to a ‘main line’ of affinities in his text, and intended the central axis of Figure 5 to represent that main line. It was not until later in the evolutionary period that direction regained some of the prominence it had lost in the partial collapse of the Chain of Being. Evolutionary trees almost invariably were drawn extending up to a crown (Figures 6, 7, and 8), and even when they were not, as in the avian tree (Figure 10) drawn by Richard Bowdler Sharpe (1847–1909), direction was communicated by the left-to-right sequence of the branches. Remnants of the sequence in Figure 10 can be found today in the ordering of taxa in any popular field guide to birds, as well as in many technical handbooks and checklists. The Chain of Being has by no means been unlinked in its entirety.

Symmetry. The quinarians believed that the Natural System was symmetrical and numerically regular (Figures 1, 2, and 4), and for this belief they were widely criticized. They insisted that this numerical regularity was a simple fact of Nature, and not a product of their own preconceptions, but their critics always found these claims difficult to believe (Strickland, 1841; Wallace, 1856). One might expect that symmetrical and regular views of the Natural System would be completely incompatible with the views of evolutionists (they were certainly incompatible with the views of Wallace, for example), but this was not always the case (contra Ghiselin, 1969: 104). At least one of the quinarians’ contemporaries objected to their work because he thought it sounded too evolutionary:

We are told, for example, that ‘the nearest approach of the mammalia to the birds exists, according to Macleay, among the glires, which make several attempts, as it were, to attain the structure of the feathered class’, as plain, strong, and precise terms, as Darwin [Erasmus Darwin!] or Lamarck himself could have used in talking of a jerboa (Dypus, Gmelin) trying to convert its legs into wings, or a porcupine (Hystrix, Brisson) endeavouring to barb its quills with feathers [Rennie, 1833: xli; italics in the original].

A further example of the compatibility of symmetrical views of the Natural System with evolution can been seen in Figure 6, the tree of the animal kingdom published by Graceanna Lewis (1821–1912). Lewis accepted evolution, but she was also heavily influenced by Lorenz Oken’s Naturphilosophie, and opened her study of the natural history of birds (1866) with Oken’s declaration that ‘The animal system is a multifariously constructed temple, with its nave, choir, chapels and towers.’ Lewis’s views of evolution were little shared by her contemporaries, however (Warner, 1979), and later evolutionists did indeed abandon all notions of a symmetrical Natural System. A remarkable example of how completely the notion of symmetry did disappear can be seen in Figure 7, a phylogeny of birds published in 1882 by Anton Reichenow (1847–1941).

Predictivity. If the Natural System was symmetrical and numerical as the quinarians believed it was, then it could also be predictive: whenever we find taxa which appear to exhibit an incorrect number of subgroups, or inappropriate analogical relationships, we know that there must be other taxa in that group which have not yet been discovered. Thus Figure 4, which shows the relationships of the crow family according to Johann Jakob Kaup (1803–1873), was drawn with several empty triangles for taxa which were believed to exist, but which had not yet been found. According to Swainson (1835: 225ff), these ‘gaps’ could be caused either by undiscovered living taxa, or by extinct taxa.

One might expect that the acceptance of evolution would cause the problem of predictivity to disappear, at least for those evolutionists who rejected systematic symmetry and numerical regularity. But evolution is in fact highly predictive with regard to the structure of the Natural System, because it takes the matter of continuity to an extreme: as noted above, evolution converts continuity into a physical, genetic phenomenon. If evolution is true then every ‘gap’ in the Natural System must be filled by extinct taxa, which may yet be discovered. Wallace, in discussing the earliest of the evolutionary diagrams reproduced here, declares it to be

an article of our zoological faith, that all gaps between species, genera, or larger groups are the result of extinction of species during former epochs of the world’s history.… Thus if the space between the Kingfishers and Hornbills [in Figure 5] has been filled up by a natural succession of families, we can see that the change must have been to heavier, larger, and larger-billed-birds, and we see such a change begun already from the Jacamars to the Kingfishers [Wallace, 1856: 206].

In the Origin Darwin was at pains to show how incomplete the geological record was precisely because of the predictions evolutionary theory made about the structure of the Natural System.

Reticulation and Branching. A key element in the quinarian view of systematics, seen in both Figures 1 and 2, was that chains of affinity were circular: they returned on themselves. Strickland rejected the numerical regularity of the quinarian system, but he did not necessarily reject circular reticulation (Strickland, 1841, 1844), and two circular chains of affinity are visible in the upper part of Figure 3, one connecting the six central genera of the Milvinae, the other adjoining the first, and connecting the lower genera of the Milvinae with one genus in the Accipitrinae and one genus in the Aquilinae. Reticulation was abandoned by the evolutionists, beginning with Wallace, but in special cases, namely in those taxa in which hybridization or symbiosis are believed to have played an important evolutionary role, a reticulate Natural System is accepted again today.

Dimensionality. Although printed on a two-dimensional page, many systematic diagrams attempted to represent three-dimensional structure. Figures 6, and 8 and 9, illustrate such three-dimensional systems, and Figure 10 was also, like Figure 8, part of a double view, showing the Natural System not only from the side but also from above (see O’Hara, 1988b: 2758, for the top view). In a remarkable passage I have quoted previously (1988b: 2750), Strickland even wondered whether the ramifications of the Natural System might exist in more than three dimensions:

whether they are so simple as to admit of being correctly depicted on a plane surface, or whether, as is more probable, they assume the form of an irregular solid, it is premature to decide. They may even be of so complicated a nature that they cannot be correctly expressed by terms of space, but are like those algebraical formulae which are beyond the powers of the geometrician to depict [Strickland, 1841].

In the evolutionary period, when affinity could be taken to mean branching genealogical relationship, an interesting conflict was set up between the depiction of the topological connections of the branches in what may be called a ‘graph space’ (like the space of a subway system, within which a traveler must follow the branching of the tracks in order to arrive at a destination), and the deployment of those branches in a two- or three-dimensional cartesian space (like the space of the city itself, under which the subway runs) in which the traditional spatial language of affinity could be used. This conflict was manifest in the many attempts to illustrate both branching trees and also map-like cross-sections through trees, and is particularly apparent in the work of Sharpe, who constructed his map-like views first, and then ‘tested’ those views by suspending evolutionary trees below them (Sharpe, 1891; O’Hara, 1988b). The conflict between cartesian spatial thinking and what I have called ‘tree thinking’ (O’Hara, 1988a) is far from resolved today.

The Problem of Interpretation

While many aspects of the diagrams I have examined here are easy to interpret, others aspects may always exceed our hermeneutic abilities. Because we can no longer directly question the authors of these diagrams we cannot determine in all cases whether a particular element of a diagram was intended by its author to carry meaning, or whether it was simply an arbitrary illustrative or printing feature. In Vigors’ original of Figure 1, for example, the circles are printed in brown ink. Was this an attempt to contrast the real nature of the taxa themselves with the abstract nature of the affinities which connect them, or was it (as I suspect) simply an illustrative accident? In Strickland’s diagrams, what is the significance of the positions of the taxa on the page in relation to the lines connecting them? If all the lines in Figure 3 were erased, but the positions of the taxa on the page were preserved, would Strickland say that the diagram still conveyed the same information? Probably not, considering his apparently conscious inclusion of circular chains of affinity in Figure 3. What if the topological connections were maintained, but the positions of the taxa on the page were altered? This question is perhaps impossible to answer. T.H. Huxley published two evolutionary trees of birds in 1868; one of them (1868a) pointed up, as most of the trees shown here do, but the other (1868b) pointed down, showing ‘descent’ rather than ascent. What, if anything, did Huxley intend to communicate by this difference? From his text one cannot tell. There is a point at which a work moves from the interactive and manipulative domain of its creation, where it can be directly challenged and where it can be explained and revised by its author, into the comparative and observational domain of history and hermeneutics, where an understanding of the work can only be built up with the tools that its author left behind. All the works discussed here have long since passed into the domain of history.

Yet even in the interactive domain of science and philosophy, meaning is teased out of works only to the extent that they are challenged and questioned, and this suggests an interesting project for some contemporary philosopher of science: take a collection of recently published systematic diagrams and interview both their authors as well as a variety of other systematists about what precisely the diagrams communicate. Can branches be moved without changing meaning? If so, in what ways? Do left to right sequences convey meaning? If not (or if so), do both authors and readers understand this? Such an inquiry would not only be a fascinating study in scientific communication and its difficulties, and of the structure of a scientific community, but it might also be a substantive contribution to systematics, to the extent that it would point out areas where improved communication is needed.

Conclusion

The representational richness I have outlined in this essay disappeared around 1900 as interest in the large-scale structure of the Natural System declined, and as skepticism about the possibility of reconstructing phylogeny grew in the experimental atmosphere of the early 1900’s (Coleman, 1971; O’Hara, 1988b; Bowler, 1989). Darwin’s question of what exactly is meant by the Natural System, and along with it the question of what exactly is the purpose of systematics, remained unresolved. These problems smouldered through the Synthesis era of the thirties, forties, and fifties, and in the systematics controversies of the last thirty years we have seen their reemergence, a reemergence which has taken place with only a superficial understanding of their Nineteenth Century history.

Some see the recent systematics controversies as an attempt to free systematics from its entanglement with evolution, and return it to a more empirical, pre-Darwinian form (Brady, 1985). Perhaps the sought-for empiricism is that of Vigors, the author of Figure 1:

Devoted to no school of natural science, and carried away by the dictates of no authority however high, no reputation however imposing, I have come to the investigation of my subject,—and I trust I may here be allowed to know myself,—unseduced by the fascinations of theory, and unfettered by the trammels of system [Vigors, 1824: 513].

Perhaps it is the empiricism of Swainson, the author of Figure 2:

… science is founded upon facts, and upon a cautious process of inductive and analogical reasoning drawn from those facts: it has nothing to do with speculative opinion or metaphysical reasoning [Swainson and Richardson, 1831: xlv–xlvi].

Or perhaps it is the empiricism of their opponent, Strickland, the author of Figure 3, who declared his approach to systematics to be the truest to Nature:

Being a purely inductive process, the details of any branch of natural history may be in this way worked out and depicted without reference to any theoretical assumptions [Strickland, 1844].

The philosophically inclined student of scientific diagrams might well ask today’s empirical systematists whether Figures 1, 2, or 3 could be published in a systematic work today.

I do not share the views of those who would create an theory-free systematics. Indeed, I believe that such is impossible, because in systematics—in any discipline—observation and theory are inextricably intertwined. Meta-systematic beliefs always influence systematists, as the diagrams in this paper show; likewise the notions of systematists act as meta-influences on those in other fields. Far from showing a need to free systematics from evolution, the controversies of the last thirty years illustrate to me that systematics still contains a great many pre-evolutionary concepts and structures, concepts and structures which ought now to be purged. We have only just begun to understand the truly evolutionary answer to Darwin’s question of what is meant by the Natural System. We are only now coming to realize that the Natural System is in fact the branching chronicle of events in evolutionary time, and that the analogy of systematics to classification is mistaken. The task of a systematist in the evolutionary world is not the construction of classes, but the reconstruction of evolutionary history (de Queiroz, 1988; O’Hara, 1988a), and diagrams of the Natural System today are not information retrieval devices, illustrated classifications, or summaries of character distributions: they are representations of history.

Acknowledgements

I am grateful to Peter Taylor for inviting me to take part in the symposium ‘Making Sense of Science Making Diagrams’, and also to P. Ericson, G.C. Mayer, M. Pinna, M. Ruse, P. Taylor, and R.L. Zusi for their comments on various drafts of this manuscript. P.F. Stevens has long encouraged my studies of the history of systematics, and I extend to him my thanks as well.

Note

1. From the symposium ‘Making sense of science making diagrams’, held at the History, Philosophy, and Social Studies of Biology meeting, London, Ontario, 1989.

References

{Figure Captions}

Fig. 1. The circular affinities of the insessorial order of birds, from Vigors (1824).

Fig. 2. Analogies between the starling and crow families, from Swainson (1837). According to Swainson’s ‘law of representation’ the same five ‘primary types’ are analogically represented in each circle of affinity.

Fig. 3. A portion of Strickland’s ‘chart of the natural affinities of the class of birds’, displayed at the 1843 meeting of the British Association for the Advancement of Science, and published after Strickland’s death by his father-in-law, Sir William Jardine (1858).

Fig. 4. ‘Hexenfuss’ of the crow family, from Kaup (1854). Blank triangles stand for taxa not yet discovered. Compare the arrangement of the subfamilies in this diagram to their arrangement in the corvid circle of Swainson (Fig. 2).

Fig. 5. The affinities of the fissirostral birds, one of two diagrams published by Wallace (1856). Note the empty node between the Alcedinidae and Galbulidae.

Fig. 6. The animal kingdom, from Lewis (1866). ‘Trifiling as it may seem, the rising of the germ to meet the warm bosom of the mother, in reality marks the whole distance from the lowest Radiate to the Warm-blooded animals.’

Fig. 7. Reichenow’s ‘Stammbaum’ of the class Aves (1882) as redrawn by Sharpe (1891). Reichenow’s original uses German vernacular names and also numbers keyed to his text, and so acts as an evolutionary table of contents to his book.

Fig. 8. ‘Verticale Ansicht’ of the evolutionary tree of birds, from Fürbringer (1888), as reprinted by Sharpe (1891). Fürbringer also published a view of this tree from the opposite side, so that the branches on the back of the tree could be seen more clearly.

Fig. 9. ‘Horizontale (Planimetrische) Projection’ of Fürbringer’s evolutionary tree (1888) at the upper horizon. The diameter of each circle corresponds to the number of species contained in it. Fürbringer also published cross sections at the lower and middle horizons.

Fig. 10. The phylogenetic tree of birds, from Sharpe (1891). Note how all of the branches come up to the same level. Sharpe also drew a top view of this tree, which is reproduced in O’Hara (1988b: 2758).


© RJO 1995–2014