1. Introduction
Authors promoting the modern synthesis forcefully banned essentialism from evolutionary theory, calling purportedly essentialist arguments “typological” and their proponents “typologists.” Mayr (Reference Mayr1982, 38) judged nineteenth-century philosophers to have been unable to understand Darwin's (Reference Darwin1859) theory of natural selection because none of them were allegedly able to shed “the strait jacket of Plato's essentialism.” Zangerl's (Reference Zangerl1948) claim that the concepts of “type” or “Bauplan” are empirically rooted in developmental and comparative anatomical studies angered G. G. Simpson to such a degree that “he never spoke to the man again” (Hull Reference Hull1988, 69). Mayr (Reference Mayr1982, 340) judged Linnaeus guilty of “essentialistic thinking” because of his insistence “on the constancy and fixity of species.” De Queiroz (Reference De Queiroz1988) accordingly asserted that the Darwinian revolution would remain incomplete for as long as the Linnean system of taxonomy would remain in use. Hull (Reference Hull1965) concluded that essentialism was responsible for two thousand years of stasis in systematics.
Essentialism in biology turns out to be an ideologically loaded concept that may be tainted not only by a misrepresentation of Linnaeus's thoughts on species (Winsor Reference Winsor2006) but also by a misinterpretation of Aristotle's biological writings. Here again, Mayr's (Reference Mayr1982, 306) account of Aristotle's essentialistic understanding of species contrasts with Lennox's (Reference Lennox1980) interpretation, according to which Aristotle distinguished species “by the more and the less.” But the distorted interpretation of historical sources is not the only source of confusion that renders the discussion of essentialism in biology difficult. There also prevails a controversial discussion of what counts as essentialism among philosophers of science. A strong version of modern essentialism (Ellis Reference Ellis2001, Reference Ellis2002) in science requires essences to be universal intrinsic properties, in virtue of which objects obey universal laws of nature. Such objects will be considered members of a natural kind, taking part in an identity relation that is grounded in the relevant essential property(ies) they all share. Species—filiae temporis for the elder Linnaeus; evolving lineages for Darwin—cannot possibly instantiate universal intrinsic properties: they cannot instantiate “true” natural kinds, but at best “cluster” or “variable” kinds (Ellis Reference Ellis2002, 12, 28). Other philosophers have claimed that species can be of a kind, the so-called homeostatic property cluster natural kind (Boyd Reference Boyd1991), but to call such a cluster natural kind an essentialistic concept (Boyd Reference Boyd and Wilson1999) requires a “refurbished” version of essentialism (Winsor Reference Winsor2006, 3). Finally, there is yet another kind of essentialism that has risen to prominence in biology, the so-called origin essentialism grounded in historical essences that are relational properties (Griffiths Reference Griffiths and Wilson1999; Millikan Reference Millikan1999; Okasha Reference Okasha2002; LaPorte Reference LaPorte2004). In spite of the furor of the architects of the modern synthesis against typological thinking, essentialism seems to raise its head again in modern biology, albeit in variously modified versions. In this article, I propose to investigate the new essentialism in biology in three different contexts: the monophyly of taxa, the bar coding of species, and the search for laws of development.
2. Essentialism and Monophyly
There is some ambiguity to the concept of monophyly as it is currently used (Sober Reference Sober, Keller and Lloyd1992). Hennig's definition (i) of a monophyletic taxon is “a group of species descended from a single (‘stem’) species, and which includes all species descended from this stem species.” He went on to consider an alternative definition (ii): “A monophyletic group is a group of species in which every species is more closely related to every other species than to any species that is classified outside this group” (Hennig Reference Hennig1966, 73). Although Hennig (Reference Hennig1966) rejected ii, it is this latter definition of monophyly that has become entrenched in phylogeography (Avise Reference Avise2000). To avoid confusion of monophyly i based on uniqueness of ancestry, and monophyly ii based on exclusivity of relationship (de Queiroz and Donoghue Reference De Queiroz and Donoghue1990; Baum Reference Baum1992), Graybeal distinguished monophyly (i) from exclusivity (ii): “Exclusivity but not monophyly is a criterion by which groups of organisms at the species level form systems” (Reference Graybeal1995, 241).
Whereas exclusivity (ii; “reciprocal monophyly” in Avise Reference Avise2000) can be a potential property of species or populations, it is less clear whether monophyly (i) can be a property of species. On Hennig's (Reference Hennig1966) definition i, it cannot, for Hennig intended the concept to apply to groups of species, not to singular species and their parts (Hennig Reference Hennig1950, 40). The major concern is that species cannot be monophyletic and ancestral to other species (Sober Reference Sober, Keller and Lloyd1992). This follows from Farris's (Reference Farris1974) disambiguation of the notion of monophyly relative to the concepts of paraphyly (a group that includes some, but not all the descendants of an ancestor) and polyphyly (a group that includes descendants of more than one ancestor) by the requirement that the ancestor be included in the monophyletic taxon that contains all and only its descendants. Sober (Reference Sober, Keller and Lloyd1992, 205) gave this definition a set theoretical expression: let xT stand for “species x is a member of T,” and xAz for “x is an ancestor of species z.” Then, “T is monophyletic if and only if there exists a species sT and for any other species d, if sAd, then dT.”
The set theoretical approach to classification that most impressed Hennig (Reference Hennig and Hannemann1957, Reference Hennig1966) was that of Woodger (Reference Woodger1952) and Gregg (Reference Gregg1954). Let the stem-species A give rise to two daughter species, B and C. Following Gregg's (Reference Gregg1954, 13) notation, the relation w = “to be an immediate descendant of” is symbolized by a w-arrow running from A to B, and a w-arrow running from A to C. These relations can also be expressed as ordered pairs: w(A; B), and w(A; C). Stipulating strictly bifurcating speciation events, as Hennig (Reference Hennig and Hannemann1957, Reference Hennig1966) did, the relation w marks out a set that contains two ordered pairs as its members: w = {(A; B), (A; C)}. The upshot is that the monophyletic taxon {(A; B), (A; C)} marked out by the relation w includes the ancestor (A) and all of its descendants (B, C). The relation w is multiply instantiated as it is shared by A-B and by A-C: it is a “one-many” relation.
As was realized by Hennig (Reference Hennig1966, 17), Woodger's (Reference Woodger1952) definition of a general concept of hierarchy was based on ordered one-many relations. Multiply instantiable intrinsic properties are the classical universal properties that function in uncompromising essentialist accounts of nature (e.g., Ellis Reference Ellis2001, Reference Ellis2002). With Hennig (Reference Hennig and Hannemann1957, Reference Hennig1966), it is multiply instantiated (one-many) relations that become the essential property of monophyletic taxa. Monophyletic taxa are individuated in space and time through their unique evolutionary origin. The essentialism underlying the concept of monophyly is thus a historical essentialism, one based on relational (extrinsic) properties. “From the point of view of a phylogenetic taxonomy … being descended from a particular ancestral population may seem to be the essential property of a particular taxon” (Dupré Reference Dupré2002, 43). Some take a relational property such as uniqueness of common ancestry to mark out individuals subject to the part-whole relation (Ereshefsky Reference Ereshefsky2007); others take uniqueness of common ancestry to constitute a historical essence marking out natural kinds (Griffiths Reference Griffiths and Wilson1999; Millikan Reference Millikan1999; Okasha Reference Okasha2002; LaPorte Reference LaPorte2004).
3. Species Essentialism and the Bar Coding Initiative
The question arises whether species can be subject to “origin essentialism,” even if the concept of monophyly cannot apply to them. A monophyletic taxon is marked out by (one-many) phylogenetic relations, whereas Hennig (Reference Hennig1950, 40; Reference Hennig1966) considered (sexually reproducing) species to form reticulated systems tied together by (many-many) tokogenetic relations.
On Hennig's (Reference Hennig1950, Reference Hennig1966) account, a species is individuated by its origin through a dichotomous lineage-splitting event (claogenesis) and goes extinct when its own lineage splits again (or through terminal extinction). Defending Ghiselin's (Reference Ghiselin1974) thesis that “species are individuals,” Hull (Reference Hull1976; see also Hull Reference Hull1988, 496) applied Kripke's (Reference Kripke1972/2002) analysis of the semantics of proper names to species names. A species name then rigidly refers to the lineage of an individual species throughout any (anagenetic) change the species might undergo during its existence. Such an account of the species as a numerically self-identical and spatiotemporally located lineage invites origin essentialism about species (Griffiths Reference Griffiths and Wilson1999; Millikan Reference Millikan1999; Okasha Reference Okasha2002; LaPorte Reference LaPorte2004): “Darwinism has done nothing to stop the resurgence in recent years of essentialist ideas, stemming from considerations in the philosophy of language” (Okasha Reference Okasha and Newton-Smith2006, 73).
Günther (Reference Günther1956, 45) praised the “significance and fertility” of Hennig's (Reference Hennig1950) “definition of speciation as a strictly dichotomous process of lineage splitting,” but as any perusal of biology textbooks will reveal, there are countless species, both vertebrate and invertebrate but particularly also plant species, that are of (sometimes multiple) hybrid origin; there are species that form through symbiosis (lichens); and there are species, both prokaryotic and eukaryotic, that are subject to (sometimes) extensive horizontal gene transfer. Indeed, it is well documented that the bacterial and prokaryotic “root system” of the tree of life has a reticulate rather than hierarchical structure (Doolittle Reference Doolittle1999, Reference Doolittle2009; Doolittle and Bapteste Reference Doolittle and Bapteste2007; Bapteste and Boucher Reference Bapteste and Boucher2008). Finally, species cannot only undergo speciation but sometimes also undergo fusion, that is, “despeciation” (Grant et al. Reference Grant, Grant, Markert, Keller and Petren2004, 1596).
Origin essentialism cannot capture the intricate dynamics of speciation, despeciation, and species organization at different levels of biological complexity. Nonetheless, species essentialism has enjoyed a renaissance with the “bar coding initiative” that endorses Mayr's nondimensional species concept (“the species concept of the local naturalist” [Mayr Reference Mayr1963, 17]). The bar coding initiative was originally introduced as a means of rapid identification of known species, but ambitions quickly grew, as bar coding was declared an efficient tool to discover hitherto unknown (cryptic) species (Savolainen et al. Reference Savolainen, Cowan, Vogler, Roderick and Lane2005). The essentialism underlying species identification by means of a DNA bar code results from the claim that phylogenetic background knowledge is not needed to identify gene species that are based on nothing but bar coding gaps (discontinuities of DNA variation among contemporaneous populations). For this reason, Wheeler (Reference Wheeler2004, 576) identified DNA bar coding as a “theoretically vacuous technology” (see also Hamilton and Wheeler Reference Hamilton and Wheeler2008).
Consider LaPorte's (Reference LaPorte2004, 5) theoretically vacuous account of species discovery based on species essentialism, with species names functioning as rigid designators: “Biologists coin new species terms by providing a sample, called a ‘type specimen.’ The newly coined term refers to the species instantiated by the type specimen. … Two biologists in the Amazonian rain forest today coin new species terms. Each collects a type specimen, specifies a term, and says, ‘Let this term name the biological species exemplified in this organism’ (the speaker points). … The type specimen may consist of only a fragment of a skeleton preserved in a fossil”—or, indeed, of a DNA fragment. Compare this passage with what Mallet and Willmott (Reference Mallet and Willmott2003, 53) say about species: “The DNA sample and its sequence readout would act simultaneously as a key part of the type specimen … and as a kind of label for the taxon to which the specimen is deemed to belong. … In principle, a single DNA sequence is no different from a single unique name for every taxon.” For Tautz et al. (Reference Tautz, Arctander, Minelli, Thomas and Vogler2003, 71), “if DNA sequences serve as the main reference [in DNA taxonomy], name changes become less of a problem, because the sequence will always provide the link to the previously used name. … This would then be akin to the situation for humans, where names are used for communication, but passport numbers or social security numbers are used for unequivocal identification.” Hanna and Harrison (Reference Hanna and Harrison2004; see also Evans Reference Evans1982) root rigid designation of proper names in a socially grounded name-tracking network that, according to Hull (Reference Hull1988, 497), should look “suspiciously familiar” to biologists, as it corresponds to the “type specimen method.” “Population thinking” denied the type specimen any particular role other than serving as voucher of a species that is, in reality, constituted by a population subject to variation. With bar coding, however, the typical DNA sequence becomes the underlying (intrinsic) essence of a species in the sense of Mayr's (Reference Mayr1963, Reference Mayr1982) nondimensional species concept, that is, a contemporaneous time slice through a locally restricted population. Rubinoff, Cameron, and Will (Reference Rubinoff, Cameron and Will2006, 2) conclude that bar coding is based on a “phenetic, essentialist” approach to species, whereas Dupré (Reference Dupré2006, 155) identified what he called “the bar code theory of genetics” with a new essentialism about biological kinds: “Feed a DNA sample into the sequencer and the machine will print out Canis domesticus.” The theoretical vacuity of this approach was recognized by Hacking (Reference Hacking2007, 234): “Projects are afoot to identify genetic ‘barcodes’ for recognizing species, but these will be contingent markers, not essences that explain why the Toco Toucan has an enormous bill” (see also Herbert and Gregory Reference Herbert and Gregory2005).
4. Essentialism and Developmental Biology
In the introduction to his Phylogenetic Systematics, Hennig (Reference Hennig1950, 2; Reference Hennig1966, 2) took up the neo-Kantian distinction of nomothetic versus idiographic sciences, arguing against the German biologist and biophilosopher Max Hartmann, who had claimed that among natural sciences, only physics is a truly nomothetic science, a “pure science of law.” On Hartmann's account, biology would qualify as an idiographic, mostly descriptive science. Following Tschulok (Reference Tschulok1910), Hennig (Reference Hennig1950, Reference Hennig1966) rejected the relevance of the idiographic/nomothetic distinction, for “description and explanation do not divide the different sciences from each other, but in each science and in each partial discipline they are interwoven in an inseparable way” (Hennig Reference Hennig1966, 3). However, the idiographic/nomothetic distinction has recently surfaced again as theoretically relevant for evolutionary developmental biology (Jenner Reference Jenner, Minelli and Fusco2008).
Participants in a meeting held in July 2008 in Altenberg, near Vienna, with the agenda to expand the modern synthesis of evolutionary theory were reported to claim that “body plans with interior spaces, segments, appendages, and multiple layers of tissue are inevitable. That's heresy for the modern synthesis but inescapable if you incorporate physics into the picture” (Pennisi Reference Pennisi2008, 197). According to Minelli and Fusco (Reference Minelli and Fusco2008, xviii), “evo-devo” seeks to disentangle “the search for ‘laws,’ or general principles, from the study of historically contingent and unique events.” Müller (Reference Müller, Minelli and Fusco2008, 19) finds the “reach of evolutionary theory significantly expanded by … a causal-mechanistic approach toward the explanation of phenotypic change in evolution.” Jenner (Reference Jenner, Minelli and Fusco2008, 101) calls “the discovery of law-like regularities or generalities” the nomothetic aspect of evo-devo, the “description of the unique and historically contingent particulars” its idiographic component. Embedding evo-devo in the modern synthesis comes, for some, at the cost of “a fundamental tension … between mutability and stability” (Schwenk and Wagner Reference Schwenk, Wagner, Pigliucci and Preston2003, 390). This echoes Darwin (Reference Darwin1859, 206), who claimed that “all organic beings have been formed on two great laws—Unity of Type and the Conditions of Existence.”
There had been earlier claims that the modern synthesis is incomplete because it remains silent on developmental biology: “[Deterministic] constraints to [biological] form thus arise as natural necessity rather than from natural selection” (Ho and Saunders Reference Ho and Saunders1984, 9–10). According to Darwin (Reference Darwin1859), organisms are fundamentally variable in all of their parts in ever so slight degrees, and yet there are gaps in morphospace, most notably those that separate animal phyla. Some of these animal phyla seem to be characterized by phylotypic stages, such as the gastrula of metazoans, the nauplius stage of crustaceans, or the pharyngula stage of vertebrates. Stable developmental stages preserved through geological time create a puzzle for Darwinian evolutionary theory, based as it is on fundamental variation, natural selection, and continuity of transformation. Some have therefore criticized the notion of phylotypic stage as a typological, or essentialistic, concept and have replaced it with the notion of a “phylotypic period” (Richardson Reference Richardson1995; Richardson et al. Reference Richardson, Minelli, Coates and Hanken1998). However, the discontinuities in morphospace persist, as development seems subject to canalization and constraint. Wimsatt (Reference Wimsatt and Bechtel1986) introduced the concept of generative entrenchment to explain how evolutionary contingency coupled with laws of physics (e.g., laws that govern the surface tension of cells, epithelia) could become fixed and conserved in the developmental trajectory of organisms. Recognizing the tension between developmental stability and phenotypic plasticity, Walsh (Reference Walsh2006, 445) argued in favor of an evolutionary essentialism: “Evolutionary biology must accept the need for the statistical theory of selection embodied in the modern synthesis to be augmented by an essentialist conception of the natures of individual organisms.”
Stability and plasticity of development are most plausibly explained as a consequence of developmental modularity. Modularity of development is a consequence of the compartmentalization (West-Eberhardt Reference West-Eberhard2003; parcellation in Wagner and Altenberg Reference Wagner and Altenberg1996) of the organism into developmental modules, which themselves are conceptualized either as local processes such as morphogenetic fields, or as morphogenetic modules such as incipient structures in anatomy (Raff Reference Raff1996, 326). Pursuing a causal-mechanistic approach to the understanding of developmental processes, Wagner (Reference Wagner2001) was among the first to conceptualize developmental modules as “homeostatic property cluster natural kinds” sensu Boyd (Reference Boyd1991, Reference Boyd and Wilson1999; see also Wagner Reference Wagner1996; further references in Rieppel Reference Rieppel2005). Characterized by more or less stable sets of homeostatic dispositional properties, such natural kinds can accommodate the contingency of variation and change, especially if dissociable developmental modules can be units of natural selection (Brandon Reference Brandon1999; Winther Reference Winther2001). Wagner (Reference Wagner2001, 8) considered Boyd's (Reference Boyd1991, Reference Boyd and Wilson1999) homeostatic property cluster natural kinds “an important liberalization to the classical notion of natural kinds which was and is often expressed in fairly absolute terms.”
5. Conclusion
The architects of the modern synthesis of evolutionary theory forcefully rejected essentialism as incompatible with the theory of natural selection and with the seamless continuity of evolutionary change (e.g., Mayr Reference Mayr1982). Modern developmental biologists argue that a return to essentialism is necessary to complete the modern synthesis. Hennig (Reference Hennig1950, Reference Hennig1966) individuated monophyletic taxa through the uniqueness of their common ancestry, thereby laying the groundwork for origin essentialism. The (controversial) application of Hennig's (Reference Hennig1950, Reference Hennig1966) concept of monophyly to species, combined with the adoption of new developments in the philosophy of language (Kripke's analysis of the semantics of proper names as rigid designators) resulted in a new, historical essentialism about species. Whereas the discussion of such historical essentialism about species remained largely restricted to the domain of philosophy of biology, modern conservation efforts resurrected an essentialism about species through the DNA bar coding initiative, which adopts Mayr's (Reference Mayr1963, Reference Mayr1982) nondimensional species concept.
An uncompromising essentialistic account of nature invokes natural kinds marked out by distinctive universal (intrinsic) properties that are necessary and sufficient for membership in the kind. Essentialism, if there is any in historical (evolutionary) biology, cannot be of this strong kind (Armstrong Reference Armstrong1997; Ellis Reference Ellis2001, Reference Ellis2002) but must be of a “liberated” (Wagner Reference Wagner2001) or “refurbished” (Winsor Reference Winsor2006) kind, as are historical essences, or an essentialism associated with cluster concepts of natural kinds. But cluster concepts are subject to some indeterminacy, and species, as well as the monophyletic taxa that they constitute, have irreducibly fuzzy boundaries. According to the Linnean scheme of classification, any organism must belong to some species and cannot belong to more than one species. Hennig (Reference Hennig1950, 311) acknowledged that speciation is a gradual process stretching through time and space, thus imparting some indeterminacy on species concepts. If monophyletic taxa originate through speciation (the disruption of tokogenetic relationships and their replacement by phylogenetic relations), then the same indeterminacy must in principle apply to the concept of monophyly (Hennig Reference Hennig1950, 311).
Time, as Whitehead (Reference Whitehead1920) put it, obtains from the passing of nature, but for the uncompromising essentialist, the basic (hidden) structure of the universe remains the same forever. If science got it right, water is H2O at all times and everywhere. Acids act acidic through the donation of a hydrogen ion to another compound at all possible world states. Sir John Herschel, who famously ridiculed Darwin's theory of natural selection as “the Law of the higgledy-piggledy,” proclaimed: “All [my] endeavors have a common feature: they can be interpreted as an attempt to annihilate time, as a search for the constant amidst change” (Schweber Reference Schweber1989, 3). Universal (intrinsic) properties don't change, nor do the laws associated with them. This contrasts with an evolving world of living beings: “There is no holding nature still and looking at it” (Whitehead Reference Whitehead1920, 14). Evolving entities, such as developmental modules, or species, are processual entities. Such processual systems are marked out by transient properties that cluster historically in theoretically relevant ways, with fluid and shifting transitions between them (Keller, Boyd, and Wheeler Reference Keller, Boyd and Wheeler2003; Rieppel Reference Rieppel2009). If evolving entities are processual systems, then an uncompromising essentialism cannot apply to them, unless one accepts with Rensch (Reference Rensch1959) a lawful determination of evolutionary processes, such that “most of the basic qualities of animals and plants … originated of necessity” (Rensch Reference Rensch1959, 75). However, it seems inconceivable that even the most nomothetically inclined evolutionary developmental biologists would be willing to adopt, or be capable of defending on empirical grounds, Rensch's (Reference Rensch1959) thoroughly deterministic worldview (see the discussion in Rieppel Reference Rieppel2007).
Some might justifiably argue that any “essentialism in degrees” robs essentialism of its essence. If so, there is no room for essentialism in evolutionary biology. Indeed, whether cluster concepts of natural kinds, applied to “developmental modules” or to species and higher taxa, should be considered as essentialistic in some novel sense of the term, or liberated entirely from essentialism, seems largely to be a matter of philosophical debate. The same applies to the adoption, or rejection, of historical (relational) essences.
