This is a revised version of a paper read in part at the XV International Congress of the History of Science, Edinburgh, 18 August 1977. I wish to acknowledge the support of National Science Foundation (grant GS 33728) for initial stages of the research, and a University of Notre Dame O'Brien Fund grant for the later stages. I also wish to express my appreciation of several criticisms and comments on the original draft of the paper by Michael Crowe, Thomas L. Hankins, Jonathan Hodge, Rodney Kilcup, Camille Limoges, Ernst Mayr, and Timothy Lenoir. I would like to thank Professor Everett Mendelsohn for help in gaining access to needed materials at Harvard University. I have also been helped in some difficult translation problems by Dr Fritz Marti, although the responsibility for all translations is my own.

¹ Darwin, C., The origin of species, reprint of 1st edn., Harmondsworth & Baltimore, 1968, p. 427.Google Scholar

² I am adopting the definition of ‘historicism’ as given by Meinecke, F., Historism (trans, by Anderson, J. E.), New York, 1972, pp. lv–lvi.Google Scholar This interprets the primary change in historical consciousness at the end of the eighteenth century as involving the recognition of specifying forces and principles acting in time, in opposition to universally acting general laws. While Meinecke restricted this to human history, I would suggest that a similar transformation is evident in late Enlightenment natural history. I would not like to be thought to be identifying historicism simply with genetic explanation.

³ Garden, George, ‘A discourse concerning the modern theory of generation’, Philosophical transactions of the Royal Society, 1691, 17, 476–7 (published 1693).Google Scholar

⁴ This claim must, of course, be modified when talking of the later writings of Bonnet and Haller. Bonnet's mature discussion of the problem in his Contemplation de la nature (1764–5), for example, treated the pre-existence theory in terms of preformed germs, which develop in the appropriate time and circumstance. This is very close to the position from which Kant initially began in 1775 with his theory of the preexistent Keimen (see below, p. 129). This allowed at least a limited degree of historical change and development, while still ensuring the essential unity of the species.

⁵ In its more sophisticated versions, climate and food could effect the precise development of the ‘germs’, if not their character.

⁶ Valuable general remarks on this crisis of historical knowledge can be found in Reill, P. H., The German Enlightenment and the rise of historicism, Berkeley, 1975Google Scholar, chapter I. Reill makes no attempt to discuss this issue in terms of the natural sciences, but the claims of the ‘historical’ sciences were subject to the same epistemological difficulties.

⁷ I have explored this at length in the introduction to Lyon, J. and Sloan, P. R., From natural history to the history of nature: the initial response to Buffon'sGoogle Scholar Histoire naturelle (MS in preparation). An excellent discussion of certain of these issues is to be found in Baker, K. M., Condorcet: from natural philosophy to social mathematics, Chicago, 1975, chapters II and III.Google Scholar For a valuable discussion of the Newtonian criticism of ‘world building’ I have profited from Kubrin, D. C., ‘Providence and the mechanical philosophy’, Cornell University PhD thesis, 1968, chapter XI.Google Scholar

⁸ Pluche, Abbé Noel, The history of the heavens (tr. by de Freval, J. B.), London, 1740, ii, 279.Google Scholar In many places this work defended Newton against the Cartesians on the grounds of the epistemological modesty of Newtonian science, which Pluche saw as never claiming to reach beyond the level of experience. See especially ii, 215.

⁹ Ibid., pp. 286–7.

¹⁰ Fontenelle, , Histoire de l'académie royale des sciences (1733)Google Scholar, in Potenz, H. (ed.), Pages choisies des grands ecrivains: Fontenelle, Paris, 1909, p. 143.Google Scholar

¹¹ In response to some perceptive criticisms made by Jonathan Hodge of the original version of this paper, I wish to make clear my claim that the problem confronting historical science in the early eighteenth century was primarily epistemological in character; I do not deny that numerous attempts were made before Buffon to construct historical cosmologies and geologies. In the wake of the crisis in historical knowledge created in part by Locke, Bayle, and Newton, the problem was not the simple assertion of such claims, but their tenability. Buffon's great significance, as I see it, lies in his willingness to confront the epistemological problem.

¹² The tendency to conflate Newton's and Locke's positions on space and time is to some extent evident in the Leibniz-Clarke controversy, and the conflation was explicit in later thinkers looking back at this. See, for example, du Châtelet, M. Gabriel, Institutions de physique, Paris, 1740, p. 97.Google Scholar

¹³ Alexander, H. G. (ed.), The Leibniz-Clarke Correspondence, New York, 1956, fourth letter, p. 39.Google Scholar

¹⁴ Ibid., pp. 63, 75.

¹⁵ Ibid.

¹⁶ Ibid., fifth letter, p. 90.

¹⁷ Koyré, A., From the closed world to the infinite universe, Baltimore, 1956, p. 275.Google Scholar

¹⁸ Leibniz's remarkable version of the preexistence theory denied not only the origin of the organism in historical time, but also its death. See Weiner, P. P. (ed.) Leibniz: selections, New York, 1951, pp. 110, 195.Google Scholar

¹⁹ Wolff's importance in fashioning the insights of Leibniz into a comprehensive programme of metaphysical, theological, and scientific inquiry is in need of greater emphasis. This work included the delimitation of numerous disciplines, including cosmology, hydrology, phytology, psychology, physiology, and minerology, and the composition of treatises at the textbook level in some of these subjects. Wolff's general programme is spelled out in his Preliminary discourse on philosophy in general (tr. by Blackwell, R. B.), Indianapolis, 1963.Google Scholar

²⁰ Wolff, , Philosophia prima sive ontologia (1736)Google Scholar, in Gesammelte Werke: Lateinische Schriften (ed. by Ecole, J. and Arndt, H. W.), Hildesheim, 1962, iii, 443, 445.Google Scholar

²¹ See Champs, Jean Des, Cours abrégé de la philosophie wolffienne en forme de lettres, Amsterdam & Leipzig, 1743, i, 160.Google Scholar This work is a systematic précis of Wolff's philosophy, and was prepared with his approval.

²² Ibid., pp. 236–7.

²³ Ibid., p. 242.

²⁴ See Hanks, Lesley, Buffon avant l'histoire naturclle, Paris, 1966, p. 127.Google Scholar Buffon's deep admiration for Newton is undeniable.

²⁵ See Wade, I. O., Voltaire and Candide, Princeton, 1959, chapter IIIGoogle Scholar; idem., Voltaire and Madame du Châtelet, New York, 1967, chapter IGoogle Scholar; Barber, W. H., Leibniz in France, Oxford, 1955Google Scholar; idem., ‘Mme du Châtelet and Leibnizianism: the genesis of the Institutions de physique’, in Barber, W. H. et al. (eds.), The age of Enlightenment: studies presented to Theodore Besterman, Edinburgh, 1967Google Scholar; Iltis, Carolyn Merchant, ‘Madame du Châtelet's metaphysics and mechanics’, Studies in the history and philosophy of science, 1977, 8, 29–48.CrossRefGoogle Scholar I particularly wish to thank Professor Merchant for allowing me to have access to this manuscript before publication, and for leading me to additional references in this material.

²⁶ French translations of Wolff's metaphysics and logic had been sent to Cirey by Frederick the Great in 1736 (see Barber, , ‘Madame du Châtelet’, op. cit. (25), p. 215Google Scholar; and Wade, , Voltaire and Candide, op. cit. (25), p. 35)Google Scholar. The interest in Wolff's philosophy generated by these works was apparently responsible for du Châtelet's bringing Wolff's disciple, Samuel König, to Cirey to act as household tutor in March 1739, a post he held until December that year.

²⁷ See especially Barber, ibid. It is more correct to see du Châtelet's work as an exposition of Wolffianism, rather than Leibnizianism. Des Champs, for example, pointing this out, indicated that much of her discussion seems to be no more than a translation of Wolff's Ontologia (Deschamps, , op. cit. (21), i, ‘Preface’).Google Scholar

²⁸ Reviews noted that this work had finally made clear and explicit the principles in the work of Leibniz and Wolff that were of value. See Mercure de France, 06 1741, pp. 1274–310.Google Scholar

²⁹ Most notably a foundation for natural necessity. Valuable historical remarks on this are to be found in Tonelli, G., ‘La nécessité des lois de la nature au XVIIIe siecle et chez Kant en 1762’, Revue d'histoire des sciences et leurs applications, 1959, 12, 225–41.CrossRefGoogle Scholar

³⁰ du Châtelet, Gabrielle Emile, Institutions de physique, Paris, 1740, p. 119.Google Scholar

³¹ Ibid., p. 105.

³² Des Champs, , op. cit. (21)Google Scholar; de Vattel, Emer, Defense du système leibnitien, Leiden, 1741.Google Scholar

³³ La méthode des fluxions et des suites infinies, Paris, 1740.Google Scholar In the historical preface to this work, Buffon sided with Newton against Leibniz on the discovery of the calculus. This preface, apparently written by 1738 (see Hanks, , op. cit. (24), p. 108)Google Scholar, elicited a defence of Leibniz from his friend Gabriel Cramer, who was the editor both of some of Wolff's mathematical works, and of the Leibniz-Johann Bernoulli correspondence. I have been unable to determine the contents of Cramer's defence of Leibniz as sent to Buffon in 1741, but by 1744 Buffon wrote that he now agreed with Cramer ‘on many things’. See letter of 4 April 1744 in Weil, F. (ed.), ‘La correspondence Buffon-Cramer’, Revue d'histoire des sciences et leur applications, 1961, 14, 124.Google Scholar

³⁴ Buffon was a close friend of Pierre de Maupertuis during the late 1730s and early 1740s when Maupertuis was in the company of Wolffians like Samuel König. Buffon had also apparently visited Cirey for an extended period in November 1738 (see letter of 4 November 1738 in: Hanks, , op. cit. (24), p. 259)Google Scholar. By early 1740 Buffon mentioned that he was personally acquainted with König, which suggests some kind of contact between them in 1739 when König was in France. See the letter to Jalabert, Jean, 11 01 1740Google Scholar, in Ritter, E. (ed.), ‘Lettres de Buffon et de Maupertuis adressées à Jalabert’, Revue d'histoire littéraire de la France, 1901, 8, 652.Google Scholar

³⁵ Piveteau, J. (ed.), Oeuvres philosophiques de Buffon, Paris, 1954, p. 26.Google Scholar

³⁶ Ibid., p. 11. I have discussed this point in some detail in my ‘The Buffon-Linnaeus controversy’, Isis, 1976, 67, 356–75.Google Scholar

³⁷ At this point I should clarify what may appear to be confused definitions of category and taxon in my analysis, as Professor Ernst Mayr suggested in comments on my original paper. In terms of the distinctions initially made by Gregg, J. R., ‘Taxonomy, language and reality’, American naturalist, 1950, 84, pp. 419–35CrossRefGoogle Scholar, and by Buck, R. and Hull, D. L., ‘The logic of the Linnean hierarchy’, Systemic zoology, 1966, 15, 97–111CrossRefGoogle Scholar, I am concerned with different logical types, identifying taxa as the classes of individual organisms (denoted by the proper Latin names), and categories as the classes of taxa (denoted by the names of the seven levels of the Linnaean hierarchy). In these terms the interbreeding criterion is a proposal for a definition of a category (e.g. species), which would demand that all taxa at the species level satisfy this condition for class membership. Taxa at species level would comprise the fertilely-interbreeding (rather than simply the morphologically-similar) group of individuals to which the taxon name may correctly be applied. Apart from problems of anachronism, the difficulty in applying this analysis to Buffon's discussion of species is that Buffon proposed that the biological species as a taxon be conceived not as a class of individuals (however defined), but as a concrete, historical lineage, linked in time by the materal bond of generation; he claimed that this could be the basis for a taxonomic system of temporal connexions, rather than of relations of classes.

³⁸ Buffon, , ‘Comparison des animaux & des vegetaux’, Histoire naturelle, 1749, iiGoogle Scholar, in Piveteau, , op. cit. (35), p. 236.Google Scholar

³⁹ Buffon, , ‘L'asne’, Histoire naturelle, 1753, ivGoogle Scholar, in ibid., pp. 355–6.

⁴⁰ This criterion was often cited in the eighteenth century, with reference to the works of Ray, John. See, for example, his Wisdom of God manifested in the works of the creation, 5th edn., London, 1709, pp. 22, 342–3.Google Scholar

⁴¹ It should be noted that Buffon's theory of the moule intérieure conforms to the epistemological premises I have suggested as primary in Buffon's mature thought, for he would seem to have been claiming that Newtonian forces cannot be ‘abstract’, but must also have a ‘concrete’ manifestation in nature. This is evident in his controversy in 1748 with Clairaut over the law of attraction. Buffon there argued that particular, physical, causes should be sought to make Newton's inverse-square law fit the observed situations, rather than introducing ‘abstract’ mathematical terms which ‘give no more than an arbitrary abstraction, instead of representing reality to us’ (‘Reflection sur la loi de l'attraction’, in Cuvier, G., ed., Oeuvres complètes de Buffon, Paris, 1835, i, 343)Google Scholar. The theory of the moule intérieure did precisely this by giving specific, physical, causes for each species, and thereby making immanent the principle of universal attraction.

⁴² Buffon, , ‘L'asne’, in Piveteau, op. cit. (35), p. 386.Google Scholar

⁴³ I have discussed this point previously in my ‘The Buffon-Linnaeus controversy’, loc. cit. (36), q.v.

⁴⁴ Buffon's preference for genealogical and biological relations over those of morphological resemblance is a critical difference from virtually all of his important contemporaries, including even his collaborator Daubenton (see below, p. 121). This point formed much of the substance of the article on the ass, in that in terms of morphological similarity, the ass and horse would appear to be more closely related than many of the varieties of dogs. Buffon was also explicit on this point in the Histoire naturelle articles ‘La chevre’ (1755), and ‘L'isatis’ (1765).

⁴⁵ See Linnaeus, , Philosophia botanica, Stockholm, 1751, aphorisms 155–61.Google Scholar

⁴⁶ Buffon, , ‘Histoire naturelle de l'homme’ (1749)Google Scholar, in Piveteau, , op cit. (35), p. 313.Google Scholar

⁴⁷ Ibid., p. 195. Compare also with statements as early as 1749 in ibid., p. 313.

⁴⁸ ‘Le mouflon et les autres brebis’, Histoire naturelle, Paris, 1764, xi, 369.Google Scholar

⁴⁹ ‘Le lion’ (1761), in Piveteau, , op. cit. (35), p. 378.Google Scholar

⁵⁰ Ibid., pp. 401, 408. The designation ‘Famille’ in the taxonomic literature had previously been fairly consistently used as a French synonym of Linnaeus's Ordo. See, for example, Klein, J. T., Système naturel du règne animal par classes, families ou ordres, genres et espèces, Paris, 1754.Google Scholar

⁵¹ See quotation from ‘Le mouflon’, above, p. 119 at n. 48.

⁵² ‘De la nature, seconde vue’ (1765), in Piveteau, , op. cit. (35), pp. 35–6.Google Scholar

⁵³ ‘De la dégénération des animaux’, in ibid., p. 401.

⁵⁴ ‘Oiseaux qui ne peuvent voler’, in ibid., p. 417. Buffon went on to remark that this ‘grand view’ was itself ‘abstract’, and needed to be made ‘concrete’ by showing the actual connexions of the birds with the quadrupeds.

⁵⁵ I have previously discussed this development in Ray, John's taxonomic philosophy in my ‘John Locke, John Ray, and the problem of the natural system’, Journal of the history of biology, 1972, 5, 1–53.Google Scholar

⁵⁶ Linnaeus, , op. cit. (45), aphorisms 162–3, 206.Google Scholar The use of the habitus had first been popularized by Bauhin, and Linnaeus had given it a prominence from his earliest works. Linnaeus's somewhat ambiguous conclusions on this have been well summarized by Larson, James L., Reason and experience: the representation of natural order in the work of Carl von Linné, Berkeley, 1971, pp. 62–5.Google Scholar See also Stafleu, F. A., Linnaeus and the Linnaeans: the spreading of their ideas in systematic botany, 1735–1789, Utrecht, 1971, pp. 66–73.Google Scholar This also includes a translation of some of the critical passages from the works of Linnaeus on this matter. Daudin's classic study, De Linné à Jussieu, Paris, 1926Google Scholar, shows that the French taxonomists, under Bernard de Jussieu's lead, were intent on following the Linnaean approach to the natural system through the use of the habitus, rather than using the sexual system based on a single part.

⁵⁷ See, for example, Adanson, M., Familles des plantes, Paris, 1763, i, pp. clv–vi, clix, clxx.CrossRefGoogle Scholar Adanson had also affirmed this principle in his earlier writings, probably deriving it from Bernard de Jussieu or Linnaeus. Other influential examples of the attempt to use the habitus or ‘port’ as a foundation for the ‘natural system’ can be seen in Jussieu, A. L.'s highly influential ‘Examen de la famille de des renoncles’, Mémoires de l'Académie royale des sciences, 1773, especially pp. 237–9.Google Scholar In this, Jussieu also insisted on the importance of the differential weighting of characters. Condorcet, in his summary of Jussieu's paper, considered it revolutionary, and responsible for inserting a new principle into taxonomy: Histoire de l'Académie royale des sciences, 1777, pp. 34–6.Google Scholar This drew an indignant response from Adanson, who claimed he had clearly delimited this principle as early as 1759, and felt that Jussieu had taken it from him without acknowledgment. See Adanson, M., ‘Eclairissement sur la méthode naturelle de m. de Jussieu, comparé a mes families des plantes en 1774’Google Scholar, Adanson MS. A D 315, Hunt Botanical Library, Pittsburgh (probably written in 1780).

⁵⁸ Adanson, , Familles des plantes, op. cit. (57), p. cccxxiv.Google Scholar

⁵⁹ Farber, P. L., ‘Buffon and Daubenton: divergent traditions within the Histoire naturelle’, Isis, 1975, 66, 63–74.CrossRefGoogle ScholarPubMed Farber has noted well the contrast between Daubenton's ahistorical and comparative-morphological point of view, and Buffon's dynamic and historical approach to natural history. Some explanation for this difference in perspective between two apparent collaborators can perhaps be given by considering Buffon's solitary work habits. The picture revealed by Buffon's correspondence, as well as by the interview with Herault de Seychelles late in Buffon's life, suggests that Buffon was doing his scientific work almost exclusively at Montbard, while in Paris he carried on only administrative work. The extant correspondence between Buffon and Daubenton, dating mainly from the 1760s, indicates almost no direct contact between them, and suggests that they were simply sending articles for the Histoire naturelle independently to the publisher, and without consultation on the contents. See Michaut, G., ‘Buffon administrateur et homme d'affaires: lettres inédits’, Annales de l'université de Paris, 01–02 1931, pp. 15–36.Google Scholar On this divergence, see also St Hilaire, Isidore Geoffroy, Histoire naturelle générale des règnes organiques, Paris, 1859, ii, 398–9.Google Scholar

⁶⁰ Daubenton, , ‘Botanique’, Encyclopédie, ou dictionnaire raisonné des sciences, des arts et des métiers, Paris, 1751, ii, 341.Google Scholar

⁶¹ I am following the summaries of all these systems in Adanson, , Familles des plantes, op. cit. (57), i, ‘Preface’.Google Scholar

⁶² These systems, unlike the Linnaean artificial system, or those of Tournefort and Rivinus, all used characters other than reproductive parts.

⁶³ For remarks on the complex influence of Linnaeus in Germany, see especially Stafleu, op. cit. (56), chapter VIII. Stafleu's analysis needs to be revised somewhat, however, in light of the fact that many also saw Linnaeus, as advocating a ‘natural system’ based on the habitus. See below, p. 123.

⁶⁴ This is despite what might have been expected in view of the strong Wolffian tradition in Germany. Only Kant, of the figures I discuss, seems to have been sufficiently in contact with the intricacies of Wolffian metaphysics, and epistemology to have perceived the critical point.

⁶⁵ See, for example, the reviews in the Göttingische Anzeigen von gelehrten Sachen, 1750, pp. 51–4, 59–62, 283–7; 1754, pp. 188–91.Google Scholar It is particularly surprising that the last review ignored all significant comment on the theses of the unity of type, the unity of historical origin, and the problem of species, which are so prominent in several of the articles in the volume reviewed.

⁶⁶ The first volume was translated by Zink, and the second two by Abraham Kästner, according to the short notice in ibid., 1751, p. 3. I have been unable to determine the subsequent translators of this original German edition. Kästner also seems to have been responsible for the notes to all of the first three volumes. The subsequent translation by F. H. W. Martini et al., Berlin, 1771–1808, was, like most foreign editions, inexpensive. It was based primarily, it seems, on the 1769 Paris edition, with the original order of exposition altered. I wish to acknowledge the assistance of Professor Shirley Roe in determining more about the translation of the original edition.

⁶⁷ von Haller, Albrect, ‘Vorrede’, to Buffon, Allgemeine Historie der Natur, Hamburg & Leipzig, 1750, i, p. xv.Google Scholar

⁶⁸ Kästner, A., in note to ‘Erzahlung der methodischen Eintheilungen der vierfüssigen Thiere’Google Scholar, in ibid., 1754, ii, p. 78n.

⁶⁹ Kästner, note to ‘Von der Art die Historie der Natur zu erlernen und abzuhandeln’;, ibid., 1750, i, p. 38n.

⁷⁰ Ibid., pp. 36n–37n.

⁷¹ See review in Allgemeine deutsche Bibliothek, 1771, 15, 589–92.Google Scholar

⁷² Linnaeus, , Systema naturae, 1st edn., Leyden, 1735.Google Scholar The classification of the Anthropomorpha given in this edition remained essentially unaltered through to the ninth edition of 1756.

⁷³ Buffon made this point with particular force in the article on the ass in 1753. See the strong endorsement of Buffon's conclusions in Diderot's article ‘Humain espèce’, Encyclopédie, ou dictionnaire raisonné …, Neuchâtel, 1765, viii, 344–8.Google Scholar

⁷⁴ This included a split of genus Homo into H. nocturnus (night men) and H. diurnus (including the H. sapiens of the 10th edn.), in the twelfth edition of the Systema of 1766. The thirteenth edition split Homo into two species, sapiens and monstrosus, with sapiens including a mute, fourfooted, and hairy ‘wild man’ on the same level as the other traditional varieties. On Linnaeus' complex struggle with the classification of man, see Broberg, G., Homo sapiens L., Uppsala, 1975 (English summary).Google Scholar

⁷⁵ In Piveteau, (ed.), op. cit. (35), p. 390.Google Scholar

⁷⁶ Home, Henry, Six sketches on the history of man … with an appendix concerning the propagation of animals, Philadelphia, 1776Google Scholar (first published, 1774), sketch I.

⁷⁷ A close familiarity with the ‘Preuves de la théorie de la terre’, of the first volume of the Histoire naturelle, is evident through much of Kant's discussion, and he used Buffon's calculations to support his own cosmological theory. See Kant, Immanuel, Frühschriften (ed. by Klaus, G.), Berlin, 1961, i, 99.Google Scholar Kant does not seem to have owned a copy of Buffon, but then his entire library apparently consisted of only around 200 books, and for financial reasons he could probably not have purchased the sumptuous Leipzig edition of Buffon. On Kant's library, see Warda, A., Immanuel Kants Bücher, Berlin, 1922.Google Scholar

⁷⁸ Lulof, Johan, Einleitung zu der mathematischen und physikalischchen Kenntniss da Erdkugel (tr. by Kästner, A.), Göttingen & Leipzig, 1755Google Scholar; Varenius, Bernard, Geographia generalis, Cambridge, 1681.Google Scholar See Kant's 1757 prospectus for the course in Frühschriften, op. cit. (77), i, 284.Google Scholar

⁷⁹ Ibid., p. 290.

⁸⁰ On the complex history of this text, see the summary remarks in May, J. A., Kant's conception of geography, Toronto, 1970, pp. 74–5.Google Scholar May is drawing on Adickes' extensive collation of the manuscripts. Of central importance is Adickes' conclusion that the first fifty-two sections of this work date from Kant's post-1770 period, while the second half date from the 1760s.

⁸¹ Buffon had, however, begun in 1770 to treat the birds by main genera in the Histoire naturelle des oiseaux.

⁸² Unold, Johann, Die ethnologischen und anthropographischen Anschauung bei I. Kant und J. Reinh. Forster, Leipzig, 1886, p. 19.Google Scholar However, I have been unable to locate direct evidence proving Kant's knowledge of this work of Kames.

⁸³ Blumenbach, J. F., ‘On the natural varieties of mankind’ (1st edn., 1776), in Bendysche, T. (ed. and tr.), The anthropological treatises of Johann Friedrich Blumenbach, London, 1865, p. 98.Google Scholar

⁸⁴ Kant had formally separated anthropology from geography by 1773, and had also separated out a metaphysics of nature from the metaphysics of sense. See letter to Marcus Herz written towards the end of 1773, in Kant, Immanuel, Briefwechsel (ed. by Schöndörffer, O.), Hamburg, 1972, p. 115.Google Scholar

⁸⁵ This distinction was, I would suggest, implicit in Buffon's separation of the ‘abstract’ and ‘physical’ approaches to natural history, although it was never made with the clarity Kant developed. Kant's use of the term Naturgeschichte, in the technical sense of a genetic history of nature, was first suggested in his early Allgemeine Naturgeschichte und Theorie des Himmels of 1755, which gave a genetic account of the origin of the solar system. The more explicit distinction of Naturgeschichte from Naturbeschreibung seems to have been contemporaneous with Kant's separation of space and time, as particularly manifest in the separation of geography from history in the opening lecture to the Physische Geographie, which apparently dates from around 1775. See May, , op. cit. (80), p. 72.Google Scholar

⁸⁶ This is evident in his earliest work, where he argued that extension and space (and presumably time) have no existence unless there is a connection of substances with one another through living forces; Kant, , ‘Thoughts on the true estimation of living forces’, in: Kant's inaugural dissertation and early writings on space (tr. by John Handyside), Chicago, 1929, p. 10.Google Scholar See also Al-Azm, Sadik J., Kant's theory of time, New York, 1967, pp. 11–14.Google Scholar

⁸⁷ This is most evident in his ‘Von dem ersten Grunde des Unterschiedes der Gegenden im Raume’, in Frühschriften, op. cit. (77), ii, 356Google Scholar, where he argued that absolute and primordial space is not an object of sensation, but a ground-concept making possible the relationships of corporeal objects. The complex connexion of this concept to the relation and succession of objects constituting Leibnizian time and space is the subject of a complicated discussion in the Critique of pure reason, especially at A 177 ff. By relating time and space as a priori forms of intuition to the empirical coexistence and succession of objects, Kant would seem to have justified carrying on empirical enquiries into the natural world in both a historical and a descriptive way, while still remaining consistent with the ‘critical’ reform. Such a synthesis would, in my view, have made possible a continuation by Kant of the ‘historical’ approach to the issue of taxonomic relationships, despite his break with Leibniz on the space-time question, which I have seen as a fundamental component of Buffon's thought.

⁸⁸ Kant, , Physische Geographie (ed. by F. T. Rink, 1802)Google Scholar, in Kant, Immanuel, Sämtliche Werke (ed. by Gedan, P.), Leipzig, 1905, ix, 5–18.Google Scholar A full translation of this introduction is to be found in May, op. cit. (80).

⁸⁹ Ibid, p. 11.

⁹⁰ Ibid.

⁹¹ See Critique of pure reason, A 34.

⁹² Kant, , op. cit. (88), p. 14.Google Scholar

⁹³ Kant's technical uses of Rasse, Stamm and Abartung would all appear to have been derived from Buffon, as the German translations of the French race, souche and dégénération and his examples closely resemble those appearing in Buffon's article, ‘dégénération des animaux’ of 1766, which had appeared in German in the Leipzig edition of Buffon in 1772. Kant was later accused of importing into taxonomy concepts derived from the French. See below, note 116.

⁹⁴ Kant, , op. cit. (88), p. 14Google Scholar: ‘Wahre Philosophie aber ist es, die Verschiedenheit und Mannigfaltigkeit einer Sache durch all Zeiten zu verfolgen’.

⁹⁵ Kant, , ‘Von der verschiedenen Racen der Menschen’, in: Kants Werke, Berlin, 1912, ii, 429.Google Scholar The ambiguity in the German Gattung complicates the exact translation of these passages. Unlike the mainstream of German zoologists, Kant seems generally to have used Gattung as the translation of the Latin genus, or French genre. A consistent practice was established in the Leipzig translation of Buffon, which was followed out systematically by Blumenbach, for example, in rendering genre as Geschlecht, and espèce as Gattung. Kant argued in 1785, however, that in the sense in which the terms were used in Naturgeschihte, there was no distinction to be made between Art and Gattung. See quotation below, p. 131.

⁹⁶ See quotation above, p. 117.

⁹⁷ Kant, , loc. cit. (95).Google Scholar

⁹⁸ It is the epistemological primacy of a ‘physical’ and ‘historical’ meaning of taxonomic groups that I consider to be the critical connexion between Buffon and Kant, and a point which was missed by all other contemporaries of Buffon I have yet encountered. The critical difference which might possibly separate Buffon and Kant would seem to be related to Kant's rejection of the primacy of Leibnizian space and time. Whereas Buffon had emphasized the ‘physical’ reality of the species as constituted by the temporal succession of beings, Kant rather emphasized the grounding of the species in an underlying law, manifest in reproductive fertility. This is indeed related to Buffon's concept of the moule intérieure but, if my analysis is correct, Buffon had subordinated this fertility criterion to historical succession.

⁹⁹ Kant, , op. cit. (95), p. 440.Google Scholar

¹⁰⁰ Ibid., p. 430.

¹⁰¹ Ibid., p. 440.

¹⁰² Kant was conversant with Buffon's theory of the moule intérieure, but considered it untenable (see Frühschriften, op. cit. (77), ii, 79–80)Google Scholar. There are many similarities between Kant's theory of generation and Charles Bonnet's germ theory, as developed in his Considérations sur les corps organisés, 1762, section XXVIII.

¹⁰³ See the review in Allgemeine deutsche Bibliothek, 1777, 35, 330–6.Google Scholar No mention is made of Kant's distinction between Naturgeschichte and Naturbeschreibung, nor is the significance of Kant's ‘physical’ interpretation of the concept of race noted.

¹⁰⁴ Blumenbach gave only a passing reference to Kant's 1777 paper in 1781, in the second edition of his De generis humani varietate, and then only to note that Kant had recognized four ‘varieties’ of men, which he contrasted with the varieties recognized by Linnaeus, Goldsmith, Erxleben, and Hunter. It seems that Blumenbach only attended to Kant's arguments more closely after the Kant-Foster controversy in 1786.

¹⁰⁵ This is not to deny that there may have been other areas of agreement between Kant and Blumenbach, but only to note the lack of any evident impact of Kant's arguments on Blumenbach's actual practice in taxonomy.

¹⁰⁶ Daudin's suggestion that Blumenbach's taxonomy showed some close similarities to the influential principles of Antoine Laurent de Jussieu which were published in 1773 and 1774 would be chronologically consistent. See Daudin, , op. cit. (56), p. 219.Google Scholar

¹⁰⁷ Blumenbach, , Handbuch der Naturgeschichte, Göttingen, 1779Google Scholar, ‘Vorrede’, pp. [iii-iv]. This primarily ‘morphological’ concept of Naturgeschichte was also noted by Blumenbach's chief biographer and student, Marx: ‘Natural History, not the description of nature, was the aim he placed before him. With Bacon he considered that as the first subject of philosophy …, and showed also how the inner properties, relations, and attributes of the individual were connected with each other and their connection and position to the whole. With this view he busied himself actively on organic and also on animal nature’: The anthropological treatises of J. F. Blumenbach, op. cit. (83), p. 10.Google Scholar Thus, while Blumenbach's concept of Naturgeschichte was not purely descriptive, it was not really equivalent to Kant's concept. In this my conclusions would disagree somewhat with those of Timothy Lenoir in his forthcoming ‘Johann Friedrich Blumenbach's research programme for natural history and the biological thought of the Naturphilosophen’ (unpublished MS; personal communication).

¹⁰⁸ Blumenbach, Handbuch, ibid., pp. 56–57. This passage was only slightly altered in subsequent editions of the Handbuch.

¹⁰⁹ Of particular note here is the examination of Kant's ideas of 1778 by the pioneer zoogeographer, Eberhard Zimmermann (1743–1810), of the Collegium Carolinum at Braunschweig. Zimmermann was closely acquainted with Buffon's theory of degenerative change but, in a unique reading of the consequences of this theory, argued that inasmuch as the interbreeding criterion no longer seemed to hold, it was reasonable to conclude that from a common Stammvater the process of degeneration could proceed in time to the point that the different lineages had to be recognized as forming distinct species [Arten]. See his Geographische Geschichte des Menschen und der allgemeine verbreiteten vierfüssigen Thiere, Leipzig, 1778, iGoogle Scholar, ‘Vorrede’. Zimmerman's arguments in particular seem to have forced Kant to re-examine the topic. See Kant's letter to Engel, Jacob, 4 07 1779Google Scholar in Kant, , Briefwechsel, op. cit. (84), p. 188.Google Scholar

¹¹⁰ See Kant's letter to Breitkopf, Johann, 1 04 1778Google Scholar, in ibid., p. 170.

¹¹¹ Kant, , ‘Bestimmung des Begriffs einer Menschenrasse’, in: Immanucl Kants Werke (ed. by Cassirer, E.), Berlin, 1929, xxiv.Google Scholar This originally appeared in the Berlinische Monatschrift, 1785, 6.Google Scholar

¹¹² Ibid., p. 324.

¹¹³ For biographical details see especially the article by Dove, A. in Allgemeine deutsche Biographie, Leipzig, 1878, vii, 173–81.Google Scholar Forster was a friend of Blumenbach's and co-editor with Lichtenberg of the Göttingisches Magazin der Wissenschaften und Literatur from 1780 to 1785, and then, during his stay in Kassel, he edited the Hessische Beiträge zur Gelehrsamkeit und Kunst. Forster was also the translator of the sixth volume of the Berlin edition of Buffon's Histoire naturelle.

¹¹⁴ Deitrich Tiedemann, a co-editor with Forster of the Hessische Beiträge, gave a detailed discussion of some of the groundwork of Kant's philosophy in Hessische Beiträge, 1785, 1, 113–30; 233–48; 464–74.Google Scholar Kant found this highly incompetent. See his letter to Bering, Johann, 7 04 1786Google Scholar, in Kant, , Briefwechsel, op. cit. (84), p. 291.Google Scholar

¹¹⁵ An example of Forster's concrete taxonomic practice in 1785 can be discerned in his ‘Beschreibung des Brodbaums’, Hessische Beitrage, 1785, 1, 384–400.Google Scholar In this he most closely followed Linnaeus's sexual system.

¹¹⁶ Forster, , ‘Noch etwas über Menschenrassen’, Teutsche Merkur, 1786, 56, 73, 79.Google Scholar

¹¹⁷ Ibid., pp. 79–80.

¹¹⁸ Kant, , ‘Ueber den Gebrauch teleologischer Prinzipien in der Philosophie’, in: Kants Werke, op. cit. (111), iv, 491.Google Scholar This first appeared in the Teutsche Merkur, 1788, 61, 36–52; 123–36.Google Scholar

¹¹⁹ Ibid., p. 491.

¹²⁰ Ibid., pp. 489–90. I am also interpolating here some of the general discussion found in the Critique of pure reason, especially A 644–68.

¹²¹ Kant, , op. cit. (118), p. 493.Google Scholar

¹²² Kant, , Critique of pure reason, A 660, modified from Norman Kemp Smith's translation, London, 1963, p. 543.Google Scholar Kant's general argument on this point in the Critique was that while the unification of organisms into natural species by interbreeding relations gave their unification under the categories of the understanding, their historical unification by derivation from distant common ancestors could only be a unifying and regulative idea of pure reason. As such, it could have a regulative, but never a constitutive function.

¹²³ Kant, , op. cit. (118), p. 493.Google Scholar

¹²⁴ Ibid., pp. 494–5.

¹²⁵ Many of the complex relations of Kant and other members of the German scientific community have been perceptively discussed by Timothy Lenoir in his ‘Generational factors in the origin of Romantische Naturphilosophie’, Journal of the history of biology, 1978, 11, 57–100.Google Scholar He is currently working on a more detailed discussion of the relationship of Kant and Blumenbach.

¹²⁶ Blumenbach, J. F., ‘Uber Menschen-Rassen und Schweine-Rassen’, Voigts Magasin für das Neueste aus der Physik und Naturgeschichte, 1789, 6, 6–7.Google Scholar

¹²⁷ Thus in his Beyträge zur Naturgeschichte, Göttingen, 1790, pp. 38–9Google Scholar, Blumenbach was to argue that the degree of morphological difference between some varieties of dogs was so great that he was willing to presume different stem-species for them, even though they could all interbreed fertilely. For either Buffon or Kant, this would have been a serious confusion.

¹²⁸ See Blumenbach's discussion of the species concept first inserted into the third edition of his Handbuch der Naturgeschichte, Göttingen, 1788Google Scholar, ‘Vorrede’, p. vii, where he stated that at least in the ordinary case, animals of the same species (Gattung) could interbreed with one another. Blumenbach's increasingly morphological definition of taxonomic groups was, however, evident in the revisions of this ever-changing preface in subsequent editions. In the fifth edition of 1797, Blumenbach inserted the claim that the fundamental definition to be given of the category genus (Geschlecht) is ‘the similarity of different species [Gattungen] of things’.

¹²⁹ Banks, as well as Georg Forster, had accompanied Cook on his second voyage around the world.

¹³⁰ Blumenbach, to Banks, , in Anthropological treatises, op. cit. (83), pp. 151–2.Google Scholar

¹³¹ Ibid., p. 153. This classification was slightly revised in the fifth and subsequent editions of the Handbuch der Naturgeschichte, with the Glires and Ferae placed under the Class Palmata.

¹³² Ibid., pp. 188–9.

¹³³ Blumenbach and Kant exchanged books in the 1790s, and Kant made use of the Bildungstrieb theory for aspects of his Kritik der Urteilskraft of 1790. See notes 105 and 125 above.

¹³⁴ By the French, for example, Blumenbach was often seen as no more than a slavish follower of Linnaean systematics. As Pierre Flourens wrote: ‘…everything belonging to method was neglected by Blumenbach; he confined himself to following Linnaeus; he adopted from him almost all his divisions, with whatever advantage they had, and also with all their defects, their narrowness of study, and their caprice’. Flourens, ‘Life of Blumenbach’, in: Blumenbach, , Anthropological treatises, op. cit. (83), p. 53.Google Scholar

¹³⁵ Girtanner had studied medicine under Blumenbach at Göttingen from 1780 to 1782, and then on two separate journeys abroad (1784–87; 1788–90) had studied medicine and chemistry at Edinburgh. The best accessible biography of Girtanner is by Wegelin, Carl, ‘Dr. med. Christoph Girtanner (1760–1800)’, Gesnerus, 1957, 14, 141–63.Google Scholar

¹³⁶ Kant's own knowledge of Girtanner's work seems to have been rather indirect; only on one occasion did he mention Girtanner's application of his philosophy in a brief, laudatory mention in his Anthropologie in pragmatischer Hinsicht (1798)Google Scholar, in: Kants gesammelte Schriften, Berlin, 1917, vii, 320.Google Scholar Apparently Kant owned only Girtanner's book on antiphlogistic chemistry. See Warda, A., Kants Bücher, op. cit. (77).Google Scholar

¹³⁷ Kant commended Reinhold's timely exposition of some of his philosophical principles in this respect, and it was apparently to Reinhold, as one of the editors of the Teutsche Merkur, that he sent his response to Forster in late 1787. See letters of 28 and 31 December 1787, and reply of 19 January 1788 in Kant, , Briefwechsel, op. cit. (84), pp. 334, 337.Google Scholar Under Reinhold's direction, Jena was a centre of intensive study of Kantianism. See Reinhold's letter to Kant of 21 January 1793, in ibid., p. 624. In his subsequent collection and amplification of these letters, Reinhold made much more explicit the programmatic character of Kantianism. See Reinhold, K., Briefe über die kantische Philosophie, Jena, 1790.CrossRefGoogle Scholar He also expounded on Kant's philosophy in his ‘Ueber das bisherige Schicksale der kantischen Philosophie’, Jena, 1789Google Scholar, which also appeared in an abbreviated form in the Teutsche Merkur for 1789.

¹³⁸ See the letter to Kant of 9 October 1789 from Johann Jachmann (in Edinburgh) in: Kants gesammelte Schriften, Berlin, 1900, xi, 92.Google Scholar

¹³⁹ Jachmann seems to have been particularly close to Kant in this period, and was recommended by him in 1788 for the foreign medical scholarship at Edinburgh: ibid., x, 511. Jachmann is also reported to have given lectures on Kant's philosophy at Edinburgh during his stay. See the article on Jachmann, Reinhold B., Johann's older brother and Kant's first biographer, in Neue deutsche Biographie, 1974, x, 214.Google Scholar

¹⁴⁰ Letters from Jachmann, to Kant, , 9 10 1789Google Scholar and 14 October 1790, in loc. cit. (138), xi, 92, 204.

¹⁴¹ Girtanner, Christoph, Ueber das kantische Prinzip für die Naturgeschichte, Göttingen, 1796, ‘Vorrede’, p. ii.Google Scholar

¹⁴² Ibid., pp. 3–4. Girtanner's use of Gattung, unlike Kant's, consistently meant species, thus following Blumenbach's conventions (see above, note 95). In the light of Kant's claim that the distinction of Art and Gattung in Naturgeschichte was unimportant, this terminological ambiguity would not seem critical.

¹⁴³ Ibid., p. 3.

¹⁴⁴ Ibid., p. 4.

¹⁴⁵ Ibid., p. 54.

¹⁴⁶ Ibid., pp. 282–3. Drawing to a large degree on the studies of Pierre Simon Pallas on the fauna and flora of Siberia, Buffon had placed the origin of man in east Asia, between 40 and 55 degrees of latitude, from which there was migration and racialization as the earth gradually cooled. See Époques de la nature, in Piveteau, (ed.), Oeuvres philosophiques, op. cit. (35), pp. 188 ff.Google Scholar Zimmerman, however, was less inclined to allow such unitary centres of origin. See Forster, , ‘Etwas ueber …’, p. 157.Google Scholar Girtanner displays close familiarity with the work of both Buffon and Zimmermann.

¹⁴⁷ See Kant's own synthesis of these ideas in his Kritik der Urteilskraft, para. 81. Girtanner cited this work often.

¹⁴⁸ Girtanner, , op. cit. (141), p. 345.Google Scholar A detailed exploration of Girtanner's use of Blumenbach's Bildungstrieb theory is to be found in Lenoir, , op. cit. (107).Google Scholar

¹⁴⁹ Girtanner termed the work only an ‘–erste Versuch einer philosophischen Naturgeschichte’. See ibid., p. 416.

¹⁵⁰ Unsigned review Göttingische Anzeigen von gelehrten Sachen, 24 10 1796, 171, 1705–12.Google Scholar

¹⁵¹ In 1793 Girtanner solicited, via Johann Erhard, Kant's opinion of his treatise on antiphlogistic chemistry. See Erhard, to Kant, , 17 01 1793Google Scholar, in Kant, , Briefwechsel, op. cit. (84), p. 622.Google Scholar Kant apparently made no reply. See also above, note 136.

¹⁵² Blumenbach, , Handbuch der Naturgeschichte, 6th edn., Göttingen, 1799, p. 24n.Google Scholar The reference was simply added without comment to a note previously referring to Kant's 1788 paper.

¹⁵³ Ibid.,

¹⁵⁴ The Leipzig philologist and anthropologist, Johann Gruber, complained in his German translation of the third edition of Blumenbach's De varietate, that naturalists (presumably including Blumenbach), refused to accept Kant's new divisions of organisms in terms of stemrelations. See Blumenbach, , Über die natürlichen Verschiedenheiten im Menschengeschlechte (tr. by Gruber, J. G.), Leipzig, 1798, ‘Appendix’, pp. 259–61.Google Scholar

¹⁵⁵ See Blumenbach, , Anthropological treatises, op. cit. (83), pp. 190–1.Google Scholar

¹⁵⁶ Mayr, E., ‘Illiger and the biological species concept’, Journal of the history of biology, 1968, 1, 163–78.CrossRefGoogle Scholar This is primarily a translation, with an introductory commentary, of Illiger's preface, ‘Einige Gedanken über die Begriffe: Art und Gattung in der Naturgeschichte’, to his Versuch einer systematischen vollständigen Terminologie für das Thierreich und Pflanzenreich, Helmstadt, 1800, pp. xxv–xlvi.Google Scholar For biographical details, see Mayr, ibid., and Stresemann, E., Ornithology from Aristotle to the present (tr. by H. J. & Epstein, C.), Cambridge, Mass., 1975, chapter VI.Google Scholar

¹⁵⁷ Illiger, ibid., p. xxv.

¹⁵⁸ Illiger's text (ibid.) opens almost identically to Girtanner's (op. cit. (141)). For example, Girtanner: ‘Die Natur is der Inbegriff von allem, was nach bestimmten Gesetzen existirt’ (p. 1); Illiger: ‘Natur, Natura ist der Inbegriff von allem, was nach bestimmten Gesetzen existirt’ (p. 2). Another important example: Girtanner: ‘Die Naturbeschreibung (Physiographie) ist die Kenntniss der naturlichen Dinge, wie sie jetzt sind’ (p. 1); Illiger: ‘Naturbeschreibung, Physiographia, die Kenntnis der natürlichen Körper unsers Erdbodens, wie sie jetzt sind’ (p. 2). Other examples could readily be given. These definitions are themselves reworkings by Girtanner of the opening discussion of Kant's 1788 paper. Illiger's discussion of the division of species into subordinate groups is taken almost verbatim from Girtanner (compare Illiger, , op. cit., pp. 6–7Google Scholar with Girtanner, , op. cit., p. 8).Google Scholar

¹⁵⁹ Illiger: ‘[Art, Species] ist der Inbegriff aller Individuen, welche fruchtbare Junge miteinander zeugen’ (op. cit. (156), p. xxvi); Girtanner: ‘Alle Thiere, oder Pflanzen, die mit einander fruchtbare Junge zeugen, gehören zu Einer physischen Gattung’ (op. cit. (141), p. 4). The fundamental difference lies in Illiger's use of Art rather than Gattung in his discussion, and the absence of the adjective physical.

¹⁶⁰ ‘In eine Terminologie der Naturbeschreibung gehört nicht die Naturgeschichte der Klassen …’, Ibid., p. xxii.

¹⁶¹ Ibid., pp. xxiii–xxv.

¹⁶² See Lenoir, , op. cit. (107).Google Scholar

¹⁶³ Illiger, , op. cit. (156), pp. xxxiii, xxv.Google Scholar See also text, p. 108.

¹⁶⁴ Ibid., p. xlvi.

¹⁶⁵ Ibid., pp. 107–8.

¹⁶⁶ This linkage was also recognized by later commentators. See especially St Hilaire, Isidore Geoffroy, op. cit. (59), ii, 399. I wish to thank Professor Camille Limoges for directing me to this very valuable discussion in his comments on the original draft of this paper.Google Scholar

¹⁶⁷ Illiger was accused by early reviewers of his work of confusing in important respects some issues and terminology. His use of Art and Gattung for genus and species, for example, was viewed as a departure from the standard practice, instituted especially by Blumenbach. He was also accused of neglecting the domain of Naturgeschichte in the text of the work. See the unsigned review in Allgemeine Literature-Zeitung, 8 05 1800, 131, 306–7.Google Scholar

¹⁶⁸ The pervasive conclusion that the key to a ‘natural’ system was to be found in some kind of affinity of organisms, analogous to chemical affinity, was predominant in both French and German taxonomy in the later eighteenth and early nineteenth centuries. See, for example, Jussieu, A. L., op. cit. (57)Google Scholar; d'Azyr, Felix Vicq, ‘Système anatomique: quadrupèdes’, in Encyclopédie méthodique, Paris, 1792, ii, p. iiiGoogle Scholar; Hermann, Johann, Tabula affinitatum animalium, Argentorati, 1783Google Scholar; I would also see similar organizing concepts at work in the taxonomy of Blumenbach and Cuvier.

¹⁶⁹ Forty years later, the botanist and theoretician of natural history Anton F. Spring noted that the distinctions and genealogical concepts of Kant and Girtanner had found their way in to several handbooks, but remained ‘…dennoch für das System selbst ohne Anwendung’. Spring, A. F., Ueber die naturhistorischen Begriffe von Gattung, Art und Abort, und über die Ursachen der Abartungen in den organischen Reichen, Leipsig, 1838, p. 76.Google Scholar