, Volume 22, Issue 2, pp 231–260 | Cite as

Species are Processes: A Solution to the ‘Species Problem’ via an Extension of Ulanowicz’s Ecological Metaphysics

Invited Paper


The ‘species problem’ in the philosophy of biology concerns the nature of species. Various solutions have been proposed, including arguments that species are sets, classes, natural kinds, individuals, and homeostatic property clusters. These proposals parallel debates in ecology as to the ontology and metaphysics of populations, communities and ecosystems. A new solution—that species are processes—is proposed and defended, based on Robert Ulanowicz’s metaphysics of process ecology. As with ecological systems, species can be understood as emergent, autocatalytic systems with propensities for centripetality and mutuality in the course of dynamically balancing ascendency (order and persistence) and overhead (randomness and change). The species-as-processes perspective accords with the Ulanowicz’s postulates of process ecology and it can be accommodated by existing theories of species—particularly in a reframing of Richard Boyd’s metaphysics such that species are homeostatic process clusters. Rather than contending that process-based metaphysics is the only, best or true account of species, a pluralist-realist approach is advocated based on the pragmatic principles that are reflected in modern view of species and ecology. If species are understood to be comprised of processes and to be emergent processes themselves, there are important implications for the life sciences, including: animal models in medical and environmental studies, conservation biology, extinction, biodiversity, restoration ecology, and evolutionary biology.


Species problem Process philosophy Autocatalysis Centripetality Mutuality Ascendency Overhead Propensity Homeostatic property clusters 

1 Introduction

One of the oldest and most vexing philosophical issues in biology is the nature of species or the ‘species problem’ (Ayala and Arp 2009; Rosenberg and McShea 2007; Hull and Ruse 1998; Ruse 1998; Sober 1993). The problem involves ontology (what exists), metaphysics (what something is), and epistemology (how we know). No entirely satisfactory answer has been developed to the question of “What is a species?” Even the matter of whether one should be a realist (i.e., species exist independently of subjective human interests) or an anti-realist (i.e., species are mind-dependent constructs) remains unresolved. There are those who propose that no single account of the nature of species will suffice, so we must take a pluralistic—but not relativistic—approach to accommodate the varied interests of biologists.

More recently, ecological ontology and metaphysics have become topics of intense philosophical inquiry (Keller and Golley 2000; Dodds 2009; Reiners and Lockwood 2010; Brown et al. 2011). The central questions include: What is a population, community, or ecosystem? Not unexpectedly, realist-antirealist and monist-pluralist debates parallel those regarding species. In this regard, Ulanowicz (2009) has developed a compelling perspective in arguing that processes—rather than material entities—are the fundamental reality of ecology.

The purpose of this paper is to propose a solution to the species problem based on Ulanowicz’s (2009) principles of process ecology. In short, while everyone seems to agree that evolution is a process, nobody has explicitly taken the next conceptual step of arguing that species—the products of evolution—are, themselves, processes.

In what follows, I briefly review the major philosophical theories as to the ontology and metaphysics of species. Then I review Ulanowicz’s process-based ecological metaphysics. Next, I integrate these two conceptual endeavors to assess how both philosophical views of species and Ulanowicz’s perspective on ecology might accommodate species-as-processes. Finally, I consider the broader implications of a process-based metaphysics of species for biology.

2 The Species Problem

2.1 Historical Overview

The ancient Greeks laid the foundation for our conceptions of species that have persisted into modern times (Serafini 2001). In very approximate terms, the Pre-Socratics viewed the world primarily as being either in constant flux (the Milesian school, as exemplified by Heraclitus’s famous saying about being unable to step into the same river twice) or that nature was fundamentally a single, unchanging substance (the Eleatic school, as represented by Zeno and Parmenides). Plato attempted to synthesize these views by contending that the material world fluctuated, but behind this lay an unchanging reality of ideal forms. Aristotle did not so much deny this view as he shifted focus from the ideal to the sensed world. Despite their differences in metaphysical views, the Greek philosophers agreed as to the nature of species. They were natural kinds—classes that existed independently of human interests.

The species-as-kinds metaphysics prevailed through the nineteenth century, although scientists differed with respect to whether species were fixed or changing, as well as whether they were real or not (Glass 1960). For example, Linnaeus originally held that species were fixed, but his subsequent work with hybrids convinced him that species were malleable. While Darwin established the changeability of species, his ontological views were unclear. His wording suggests he was still thinking in terms of natural kinds (in Korey 1984, p. 228):

There is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one; and that, whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved.

Although it appears that Darwin took species to be kinds, this interpretation is not essential to my line of argument. Rather, what is important is that in the twentieth century philosophers of biology have explicitly defended this metaphysical perspective—as well as offering intriguing alternatives.

2.2 Contemporary Theories

2.2.1 Species as Sets or Classes

Kitcher (1984) and Kitts and Kitts (1979) are two of the leading proponents of the view that species are abstract collections of organisms. Kitcher (1984) advances a pluralistic realism grounded in species being understood as sets. His argument is derived from four fundamental theses. First, species are sets of organisms, such that an organism and its species are understood in terms of the member-set relationship. Second, species are sets of organisms related to one another by complicated biological relations, no one of which is ontologically privileged so as to uniquely delimit all species. Third, the necessary heterogeneity in what constitutes species arises from the two main—and legitimately variable—approaches used by scientists to demarcate biological categories (i.e., grouping organisms by structural similarities and by phylogenetic relationships). Fourth, pluralism about species is compatible with realism regarding these biological sets of organisms. Thus an organism can belong to a morphological set which delimits one sort of species and to a phylogenetic set which encompasses another kind of species—and while the placement of the organism into one or the other of these sets (species) reflects human interests, the existence of both is an objective fact of the world.

According to Kitts and Kitts (1979, p. 613), members of a species, “share an essence which distinguishes them from the members of other species and accounts for the fact that they are reproductively isolated from the members of other species.” They hedge as to whether species are technically either sets or classes. The important point is that the view of Kitts and Kitts along with many others (e.g., Gregg 1950; Buck and Hull 1966; Lehman 1967; Ruse 1969, 1971) entails that species are kinds in a way that accords with evolutionary theory. Contrary to Kitcher (1984), Kitts and Kitts advocate a monistic view of species grounded in evolution and dismissing mere morphological properties such as pigmentation. The property which all organisms of a species must share is not merely some manifest feature but an essential, underlying trait which accounts for reproductive isolation. In reply to those who would deny that species are natural kinds because their names cannot be included in scientific laws (e.g., Ghiselin 1974; Hull 1976a, b, 1978), Kitts and Kitts contend that on this basis most classes of material entities would not be natural kinds—a most unfavorable consequence for science.

2.2.2 Species as Individuals

Ghiselin (1966, 1969, 1974) and Hull (1976a, b, 1978) are two of the primary proponents of the alternative, modern view of species (others in or near this camp include Griffiths 1974; Mayr 1976; Sober 1993). Ghiselen holds that although species traditionally have been regarded as classes or universals, they are best understood to be individuals or particular things. Resistance to this perspective might derive from confusing the logical term ‘individual’ with the biological synonym ‘organism’. The better way of understanding the concept is by the analogy that species are to evolutionary theory as firms are to economic theory. From this view, four important features of species follow. First, as with individuals, species do not have intensions or defining properties. Analogically, there are no necessary and sufficient properties that define a firm such as American Motors. Second, because it is not possible to specify characteristics that define a species, one must appeal to ostensive definitions. That is, the only way to define a species is to ‘point’ to that which bears the name (as we must do in identifying American Motors). Third, if species are individuals there cannot be instances of them. For example, the class of manufacturing firms has instances: American Motors is a manufacturing firm. However, we would not say, “the factory in Detroit is an American Motors.” Fourth, while classes have members, individuals have parts. So the constituent organisms of a species—or the people working for American Motors—are parts of these individuals. Ghiselin (1974) extends the economic metaphor, which is pervasive in ecology and evolution, to define species as, “the most extensive units in the natural economy such that reproductive competition occurs among their parts.”

Hull (1978) and Sober (1993) make a somewhat different argument to contend that species are individuals. As with Ghiselen, they recognize that species typically have been conceived as being spatiotemporally unrestricted classes. Hull’s central line of reasoning is that for species to function in the evolutionary process, they must be spatiotemporally localized individuals. He maintains that the result of evolution is that copies of the entities are selected, not sets. Just as individual genes are understood to be historical entities which exist for short periods of time, breeding populations (Sober 1993) and gene lineages (Hull 1978) are historical entities dynamically persisting indefinitely through time.

A third theory might be best classified as version of the species-as-individuals perspective. Ereshefsky (1992) refers to species simply as the “units of selection.” He notes that the founding population is the locus of evolution such that species are not kinds. Rather, species are historical entities which constitute individuals of a sort. But what sort of individuals are they? Ereshefsky (1999) notes that many species lack the integrating force of gene flow (Mishler and Donoghue 1982) so they are not individuated in the traditional, evolutionary sense via interbreeding. Instead, if biologists are correct in their belief that such entities are species, then these entities are perhaps more like organisms. But this is not the entire story. According to Ereshefsky’s account, species are individuals that are held together by processes such as selection, genetic homeostasis, or developmental canalization. As such, a species is not individuated by virtue of causal interactions among its parts or elements—at least a species in which there is little or no gene flow among its organismal parts. This is what we might expect of an organism, but it is possible that some species are not like the interdependent cells that constitute an organism. Asexual species, for example, are individuals by virtue of exogenous forces acting in some common and unifying manner. Such a view can be contrasted with Vrba’s (1995) contention that species should be understood as complex systems—presumably individuals of some sort—with internally derived structure.

2.2.3 Species as Clusters

Boyd’s (1999) position is derived from the central concept that species are ‘homeostatic property clusters’ which takes into account their contingent, historical features and indistinct or vague boundaries. Conceiving of species as homeostatic property clusters entails four elements. First, properties are contingently clustered. That is, in the natural world certain features co-occur in a biologically important number of cases—and these cases amount to species.

Second, the clusters of properties typically arise and persist through metaphorical or literal homeostasis. This stabilizing process may result from some properties maintaining or favoring the presence of others, a relationship which can presumably occur in a unidirectional or reciprocal manner. Homeostasis might also be a biophysical process that holds together a cluster of properties that share some causal relationship to an underlying mechanism. And, of course, a property cluster can be formed from both the interactions among properties and biophysical mechanisms.

Third, the clustering of properties is causally important. That is, the clustered properties should produce important effects. Although Boyd does not fully explicate what makes for an important effect, it seems that this must entail some outcome that is more or less biologically vital to the species’ ability to persist. However, Boyd’s ontology is explicitly pluralistic in that there is an accommodation between how biologists classify natural kinds with regard to varied scientific interests and the causal structures in the world. So, a homeostatic property cluster must reflect some objective causal outcome that pertains to the flourishing of the species and this outcome must be of conceptual importance in human inquiry.

Fourth, homeostasis can be a matter of degree. Particular instantiations of a property cluster may be imperfectly held together, so it is nomologically possible (or actual) that homeostasis is imperfect. As such, we should expect that a given organisms may display some but not all of the properties. This might be result of having only some of the relevant homeostatic mechanisms present or perhaps in a weakened state.

Boyd’s homeostatic property clusters represent a more open-ended solution to the species problem than that of Sterelny (1999), who argues that species are evolutionarily linked metapopulations. In a sense, these are population clusters with homeostasis provided by evolutionary processes. As with Boyd’s concept, there is no necessary requirement for spatial contiguity—a species can be scattered across different ecological communities. Indeed, Sterelny’s evolutionarily linked metapopulations might be thought of as a particular case of Boyd’s homeostatic property clusters. Species-as-clusters sidesteps or subsumes, depending on one’s interpretation, the class-versus-individual debate. Boyd (1999) contends that whether species are classes or individuals, they are natural kinds. From this perspective, class-individual distinction is of little metaphysical importance. What matters is that species are natural kinds whatever further claims one might want to make.

3 The Ecology Problem

3.1 Newton, Darwin and the Third Window

Ulanowicz (2009) makes the case that there have been two major, conceptual advances in the history of science at least insofar as our understanding of biology is concerned. Newton’s worldview entailed five fundamental principles: causal closure (causes are a function of physical forces coupled to particular forms of matter), atomism (material existence is decomposable to elementary units via metaphysical reductionism), reversibility (processes are symmetrical with respect to time so nothing essentially new can arise), determinism (the past and future can be specified with arbitrary precision once we’re given the precise initial conditions), and universalism (the laws of nature apply everywhere, at all times, over all scales). As such, life is just a special and complicated physical phenomenon.

The next conceptual revolution was Darwin’s theory of evolution. The radical implications of chance, contingency, and novelty were initially vitiated by the neo-Darwinists who aped physicists in applying the principles of statistical mechanics to genetics (Ulanowicz 1997). But this stopgap measure failed to recognize that Darwin had described a macroscopic process not a microscale law.

Newtonian and Darwinian concepts were attempts to explain the orderly arrangements of matter using a few linked and invariant laws. Ulanowicz (2009) proposes another way of seeing the natural world—a third window—which provides a new ecological metaphysics centered on process and based on a fundamental dichotomy.

3.2 Irreducible Dichotomy

Ulanowicz (2009) begins with the superficially simple observation that things both persist and change. This deceptively straightforward empirical statement gives rise to a philosophically challenging question. Can a monistic metaphysics, such as that of Newton and Darwin, account for what appears to be fundamentally different features of reality? Ulanowicz (2005, 2009) takes seriously the irreconcilability of stasis and change. In the most basic sense, these two aspects of the world cannot be reduced to common terms because change includes the element of time and stasis does not (Ulanowicz 2009). Ulanowicz suggests that our conceptual difficulty with such a dichotomy may lie in the sense that these two aspects of the world would be at odds, cancelling one another in a way that would undermine the development or persistence of complex features of the world. However, he notes that such a view pertains, at best, to spatiotemporally restricted frames of reference:

“This relationship between the complementary dynamics of organization and chance is akin to a Hegelian dialectic. They remain antagonistic within the immediate domain, but they become mutually dependent over the larger realm” (Ulanowicz 2009, p. 7).

Another expression of this essential dichotomy frames Ulanowicz’s ‘third window’ and pertains to the nature of complexity, which is the condition for biological systems. Complexity decomposes into two terms. One term captures the features of order, efficiency, and coherence (which we associate with persistence), and the other term encompasses the qualities of disorder, inefficiency, and incoherence (which we associate with change). The inclination of earlier scientists has been to conceptually or methodologically marginalize the aleatoric qualities (i.e., the element of chance). Conceptually, the approach has been one of a monistic reconciliation in which chance is taken to be only an epistemic shortcoming, not a metaphysical reality. Methodologically, probabilistic and statistical approaches were used to nullify the role of randomness (e.g., the Maxwell and Boltzmann distribution for idealized gas, the Hardy–Weinberg equilibrium for allele frequencies, and Fisher’s models of phenotypic trait frequencies for population genetics). By driving this aspect of the world to conceptual—even metaphysical—extinction, science promises (but often fails to deliver, sometimes in catastrophic ways) predictability and control (Ulanowicz 2009).

3.3 Ecological Metaphysics

The fundamental dichotomies of stasis/change and order/chance provide the basis for Ulanowicz’s metaphysics, which are expressed in terms of three postulates.

3.3.1 First Postulate

The initial postulate of Ulanowicz’s (2009) ecological metaphysics concerns the nature of change: The operation of any system is vulnerable to disruption by chance events. In particular, Ulanowicz maintains that for complex configurations such as biotic communities, it is implausible to contend that dynamics are purely a function of the behavior of the smallest elements acting in accord with physical laws. He grounds this postulate on two concepts: chance and propensities.

With regard to chance, Ulanowicz argues that this feature is both macroscopic (not merely occurring at the level of gas particles and genes) and ontological (not merely resulting from our epistemic shortcomings). Earlier scientific attempts to apply probability theory to random events relied on there being generic and repeatable happenings which accord with ‘simple chance’. Ulanowicz builds on Elsasser’s (1969, 1981) analysis of ‘complex chance’—events which are unique and therefore not analyzable via probability. Elsasser showed that we can be reasonably certain that any particular combination of just 75 distinct, randomly co-occurring tokens will be unique. With this many or more elements, any particular combination will never recur again by chance. Ulanowicz notes that real ecosystems are most assuredly composed of at least 75 distinguishable individuals (Kolasa and Pickett 1991). Thus, every ecosystem is unrepeatable and any hope of there being ecological laws is dashed. Another crucial result is that chance is macroscopic (at the level of biotic communities) and ontological (even complete knowledge of every biotic community would not allow reliable predictions). Stochasticity is not an external form of interference for living systems that scientists can reasonably idealize away or methodologically avoid; it is an internal and essential feature.

Because unique events are not prone to probabilistic analysis, Ulanowicz (2009) adopts Popper’s (1990) concept of propensities, or the tendencies of complex systems to respond in characteristic ways (cf. Charles Peirce’s ‘habits’; Shapiro 1973). Even with extremely contingent conditions in which chance plays an important role, a system may persist. Propensities can be understood as the constraints that provide internal coherence and prevent a system from disintegrating in response to entirely unique events. A vitally important feature of complex, persistent systems is that their propensities co-occur and may be interdependent. From here, Ulanowicz suggests that the juxtaposition of propensities could serve to counter the ubiquity of chance perturbations. Within this metaphysical framework interacting propensities hold biological systems together.

3.3.2 Second Postulate

Ulanowicz’s (2009) next postulate concerns the nature of persistence: A process, via mediation by other processes, may be capable of influencing itself. This concept is derived from Bateson’s (1972) research on process networks which demonstrated how a later element in a causal chain can affect an antecedent element via feedback. As a consequence, it is possible for causal action to occur at the level of observation—a denial of atomistic reductionism and causal closure. As such, Ulanowicz (2009, p. 75) maintains that, “The processes, as a union, make a palpable contribution toward the creation of their constituent elements.” This second postulate entails five features: autocatalysis, centripetality, autonomy, mutuality and ascendency/overhead.

Autocatalysis occurs when all consecutive links in a system have a propensity for positive feedback. Such arrangements are subject to natural selection in an interesting way. Because the processes are interdependent, if one changes and impedes another, there is system-level selection against this modification. Conversely, if a change enhances some linked process, then selection will favor the new feature. In addition, the distinction between endogenous and exogenous selection becomes blurred. For example, if some element of the system changes and thereby gives rise to a modified process which allows the use of a resource that in turn causes a linked process within the system to operate more efficiently, then it’s not clear whether this is better understood as an external or internal dynamic (the resource was external but the advantage was via an internal relationship).

Natural selection acting on autocatalytic systems gives rise to centripetality. Ulanowicz (2009) takes this to be the physical pulling of matter and energy into the autocatalytic orbit, although it also can be understood as a dynamic that draws a network of processes toward a stable configuration. Central tendency accounts, at least in large part, for the persistence of autocatalytic forms. Indeed, such systems generally outlive their constituent elements (e.g., bodies outlive cells and ecosystems outlive organisms).

From the interaction of selection pressure, centripetality and long life emerges autonomy. That is, an autocatalytic system is not merely the sum of its parts. The behavior of a biotic community cannot be understood by reducing the system to its constituent organisms. But for Ulanowicz, emergence is not only epistemic. Ontological emergence results from the network of feedbacks. Because causal action can occur at the level of the system itself (not merely via its parts, as with Newtonian causal closure), feedback processes can give rise to truly novel phenomena.

Taken together, autocatalysis and selection result in mutuality, rather than competition, being ontologically primary. Only a system that first exists via a network of cooperatively reinforcing processes has the persistence needed to acquire resources which may become in short supply such that competition follows.

The final feature of Ulanowicz’s (2009) second postulate is ascendency/overhead. Ascendency is the capacity of a system to order itself and its environment. To prevail in the face of disturbances and competition, an ecological system must have both size and organization. These features can be determined by ecologists using measures of ‘total system throughput’ and ‘information’. Counteracting ascendency is the complementary feature of overhead. This term refers to the degree of disorganization or the incapacity of the system to self-organize and make maximally efficient use of resources.

The crucial aspect of the tension between ascendency and overhead is the trade-off between performance and reliability. Systems, whether biological or engineered, cannot simultaneously achieve high performance and low risk. Supersonic aircraft have far less stability than biplanes, so a power failure in the latter is less likely be catastrophic. In terms of living systems, highly specialized organisms make efficient use of resources but their populations are prone to crash if particular conditions are not met, while generalist species are not efficient but their populations are resilient when resources are unstable.

3.3.3 Third Postulate

The final postulate of Ulanowicz’s (2009) ecological metaphysics is: Systems differ from one another according to their history, some of which is recorded in their material configuration. He maintains that coherent systems function as a unit with feedback that often establishes a particular orientation in their development. Once a trajectory is established, the system no longer responds blindly to chance events. It demonstrates a preference for changing in ways that are in accord with its internal pattern of established processes and resists changes contrary to its historical path. In this sense, living systems can be said to have a telos; their development is directional—not by virtue of some exogenously imposed purpose but through their endogenous canalization. The past events and responses that strengthen or weaken linkages in an ecological network are local and hence unique. As such, “history can serve as a criterion of identity” (Ulanowicz 2009, p. 70).

3.4 Process Ecology

Ulanowicz’s concept of process ecology is the synthetic result of his postulates and represents a particular formulation of the broader framework of process philosophy. A brief review of this larger context—largely adapted from Rescher (2008)— both situates Ulanowicz’s view and serves as a basis for my argument that species can be fruitfully and validly understood to be processes.

Ever since Aristotle, western philosophers have accepted that things are real. Of course, everyone recognizes that objects are not static. Iron rusts, bodies age, and wood decays. But these changes are secondarily imposed on the underlying substance. Process philosophy inverts this commonplace perception. What is ultimately real is the process of oxidation manifest in iron, the process of maturation instantiated in bodies, and the process of decomposition evidenced in wood. Thus, both productive and destructive processes characterize the essential nature of existence. In short, a thing is what it does; its doings constitute its reality.

Processes can be understood as a series of complex, sequentially structured events. And the core project of process philosophy is to gain an understanding of these modes of change, rather than things. For example, a wave is not the mound of water. Rather it is the passage of energy through a liquid which is manifested as a temporal change in the surface of the water. As with all natural phenomena, we can best understand waves not in terms of their material properties (e.g., height and weight) but in terms of their immaterial processes (e.g., contingent originations and novel actions).

Process philosophy pertains to ontology, metaphysics, and epistemology. Processes exist and the elements of change and time, as evidenced in matter, are how we know about reality. So process philosophy does not deny the existence of objects (e.g., rocks, squirrels, and ponds). Rather, such apparently fixed entities are enduring patterns of relative stability (Rescher 2008)—processes unfolding slowly enough, relative to the temporal frame of human experience, to appear unchanging. But an object is ultimately the instantiation of a dynamic process, however gradual the change might be. In an important sense, we know of something’s existence by virtue of its interactions. A hypothetical entity that did not participate in processes would indiscernible, a view formulated by Reiners and Lockwood (2010) as afficit ergo est (it affects, therefore it is).

With this background, we can turn to Ulanowicz’s (2009) formulation of ecology in which process underlies both change and persistence. In accord with the postulates of process philosophy (Rescher 2008), Ulanowicz maintains that process is ontologically foundational to ecology. Second, he holds that the essential features of ecological processes—change and time—are metaphysically fundamental. Next, process, change, and time provide the best epistemic framework for understanding of the major elements of ecology (e.g., food webs, communities, and populations). Finally, the cardinal questions of ecology pertain to the nature of contingency, emergence, novelty, and change.

Ulanowicz (2009, p. 29) is keenly interested in the autocatalytically self-organized networks of ecology and so defines process as, “the interaction of random events upon a configuration of constraints that results in a nonrandom but indeterminate outcome.” In particular, he maintains that there are three important features of the kinds of processes that he’s defined. First, while they are subject to stochastic inputs, their outputs become less random with time. Second, these processes are self-structuring in that their state at a given time affects their future state. Third, the unfolding nature of the process is progressively constrained by the actual history of events.

Ulanowicz (2009) advocates a shift in ecological ontology, metaphysics, and epistemology away from objects to configurations of processes. As with other process philosophers, he holds that material substance is necessary to our empirical engagement with the world. In macro-scale living systems, however, causal processes most often occur at levels at or above the phenomenon of interest (a bottom-up causal account may provide an explanation for the dynamics of micro-scale physical systems).

Most process philosophers have emphasized constructive change, particularly in their optimistic framing of individual and social dynamics. Ulanowicz’s (1997, 2009) concepts of ascendency and overhead suggest a more balanced view of productive and destructive events than is typical of process philosophy. Even so, his account of ecological process as a progressively ordered network of interactions accords with a kind of optimism. Indeed, Ulanowicz (2009) asserts that ecological change is directional, but his view of telos is one of an historical canalization of change reflecting how a system has self-organized in an environmental context, rather than a pre-existing or externally determined purpose toward which a system develops.

4 Ecology and Evolution: Species as Processes

4.1 The Process Thesis

The solution that I wish to offer to the ‘species problem’ is a conceptual extension of Ulanowicz’s process ecology. That is, a philosophically sound and scientifically useful answer to the long-standing question of “What are species?” is that they are processes. I have previously suggested that a species is what it does (Lockwood 2004a, b). However, my earlier proposal was not explicated in any detail. Rather, it offered as an explanation for why we think that a species that can no longer engage in what we take to be its essential actions is extinct.

In particular, I was considering the possibility that some remnant of the Rocky Mountain locust, Melanoplus spretus, could be found surviving in a remote habitat. A defining feature of being a locust is the process of phase variation in which the solitary form transforms into the migratory form (involving morphological, physiological and behavioral changes) which assembles into cohesive swarms (Lockwood 2004a). No individuals of M. spretus have been found for more than a century, and any surviving pockets of would not have exhibited the essential activities of the ancestral stock in more than 100 generations such that even the genetic capacity for phase transformation and swarming would likely have been lost (Lockwood 2010). As such, one might hold that the Rocky Mountain locust, understood as in terms of its defining processes, is extinct whatever might be the case with regard to a remnant population.

I am proposing that species are processes, and they are comprised of processes. This metaphysical view can be derived from the framework that Ulanowicz (2009) has erected for process ecology. Borrowing from Ulanowicz, a species is a network of self-organizing processes that emerges from the interactions of its constituent processes (organisms are also best understood as autocatalytic processes). Moreover, processes may provide the foundational ontology uniting ecology and evolution. The process of natural selection accounts for the formation of networks of ecological processes (Ulanowicz 2009), so it is not implausible to contend that evolution also gives rise to the biophysical processes held together via centripetality that we recognize as species.

Ulanowicz (1995, 2009) and Zorach and Ulanowicz (2003) comes close to my contention that species are processes. In arguing for his ecological metaphysics he presents many flow diagrams in which organisms interact as autocatalytic systems, but none of these examples are based on a single species. However, one can imagine how different life stages, sexes, phases, morphs, castes, or other variants of a species could exhibit the features that Ulanowicz attributes to multi-species, ecological networks. He opens the door (perhaps the window?) to such a conceptualization:

“Under process ecology, some types [which I take to include individuals of a species] become linked in more or less autocatalytic fashion, so that changes in any one of the group conform to a degree with (are rendered meaningful by) the actions of other members of the participating ensemble” (Ulanowicz 2009, p. 134).

In arguing that species are individuals, Ghiselin (1974) suggested that he, “aims to show that such a position has attractive qualities from the point of view of logic and biology alike, and that it is perhaps not so radical as one might think.” The same may be said of my view that species are processes.

4.2 Reframing Species Ontologies

My first approach to arguing for the plausibility of species-as-processes is to assess whether this view can be accommodated within or adapted to the existing conceptualizations of species. As such, the challenge becomes one of how processes might be consistent with the metaphysics of sets/classes, individuals, and clusters.

4.2.1 Species and Processes as Sets or Classes

Kitcher (1984) argues that species are sets, and it would seem to be a category error to propose that sets can be processes. But there is perhaps no better reason to specify that they are things or abstract objects. It may be just as reasonable to say that a set is an abstract process. Of course, this would be most plausible if the sets are intensional (i.e., being able to change as their membership depends on some sort of rule) rather than extensional (i.e., being fixed as their membership is enumerated, although one might conceive of a set as being a process of indefinite duration which is indistinguishable from being an object). The preferable concept might be that of abstract individuals (which can be either objects or processes, as we will see). Although sets might be metaphorically understood as processes in some sense, there are more direct connections between Kitcher’s view of species and Ulanowicz’s view of ecological systems.

Two of Kitcher’s theses (see Sect. 2.2.1) have aspects that pertain to a process-based metaphysics. First, he contends that species are sets of organisms, and this does not preclude organisms being processes. By this account, a species could be a set of processes. Kitcher (1984) further maintains that the organisms within a species are related to “complicated, biologically interesting relations” (p. 309). In this regard, processes could constitute such relations. For example, organisms might share in the processes of sexual reproduction, cooperative defense, or mutual grooming.

Second, Kitcher (1984) argues that the species category cannot be unified because biologists demarcate species in two legitimate ways. Both criteria can be interpreted as processes. One approach is to group organisms with regard to structural similarities. From a process-based metaphysics, a structure is the physical manifestation of an underlying process. Anatomical structures do something, and what they do is fundamental. Rather than the classical view of operari sequitur esse (function follows upon being), a process-based biology holds esse sequitur operari (form follows function). As such, taxonomists identify sets of organisms based on shared processes. The other approach used to group organisms into species is by reference to phylogenetic relationships. And, of course, phylogeny is the evolutionary development of an organism (or parts of an organism) which is unambiguously an account of process.

Kitts and Kitts (1979) regard species as natural kinds, and it seems that processes can also be natural kinds (at least as reasonably as they can be objects). Tornadoes, avalanches, and solar flares are processes that are quite plausibly natural kinds. Kitts and Kitts further specify that the natural kinds that are species have members with essential attributes in common. By this account a species could also be a process, as processes can be natural kinds that share essential features.

Kitts and Kitts (1979) also contend that the essence that distinguishes members of a species does so in virtue of its accounting for reproductive isolation from members of other species. This further conforms with a process-based metaphysics as it is not sufficient that some property be merely different (e.g., pigmentation) but it must restrict or prevent the flow of genes (a process) between species.

Finally, in arguing against the metaphysics of species as individuals, Kitts and Kitts (1979) point out that just like complex wholes have interacting parts (e.g., automobiles and organisms), complex classes have interacting members (e.g., football teams and species). Whichever concept one chooses, it seems that process (or ‘interaction’) is fundamental, and the emergent whole can readily be understood as a process itself.

So, as with Kitcher, it seems that Kitts and Kitts allow that processes could comprise membership in a species and even be what creates and maintains a species. But whether a set or class can be a process (or for that matter an object, except in a metaphorical sense) is not entirely clear. At the very least, if species are sets or classes, it is no less plausible to conceive of them as processes than it is to refer to them as objects.

4.2.2 Species and Processes as Individuals

Sober (1993) provides an explicit analysis of what it is to be an individual. Although he frames the analysis in terms of objects, there seems to be no reason why it can’t apply equally well to processes. What makes two things (or processes) part of the same individual is not their similarity (cells and gestures from identical twins are strikingly similar but not parts of a single individual) but causal interaction with each other in characteristic ways (muscle cells and blood cells—as well as gestures and speech patterns—interact within an individual). From this perspective, not only can processes that causally interact comprise an individual (consider Ulanowicz’s autocatalytic systems), but an individual can be a process. Hence, if species are individuals they can also be processes.

Hull (1978) doesn’t specify the criteria for individuation, but he does allow the possibility that species are processes. Hull maintains that significant evolutionary change takes place through a series of successive species so it is a species lineage that evolves. And a lineage can surely be understood in terms of time and change such that it is a process. He also asserts that organisms belong to a particular species not because they possess any essential traits but because they are parts of a ‘genealogical nexus’. It appears that species are analogous to snapshots in time or still frames from a movie reel. Moreover, Hull (1978) contends that the existence of asexual species demonstrates that gene flow can’t be the only mechanism that maintains evolutionary unity. There must be some other processes, and Ulanowicz’s (2009) self-organization and centripetality would seem to be ideal candidates.

Finally, in a kind of reductionistic move as to the nature of species, Hull (1978) argues that genes are of central importance as they are “the entities which are passed on in reproduction”. But, this understanding of genes as physical objects is metaphysically problematical (Rheinberger and Müller-Wille 2009). An arguably more viable concept holds that genes are complex, autocatalytic systems providing the flow of information; the material instantiation is secondary to process. This view accords with Ulanowicz’s (2009, p. 154) contention that medical research has myopically focused on genes:

“If one demotes genomes to the status of (passive) material causes, one does not thereby abrogate the necessity to include them in analysis of pathology, The insufficiency of material cause as full explanation, however, motivates the investigator to continue the search for efficient agency among the enzymatic/proteomic system of reactions that interface with and interpret the genomes.”

So a species may be a process comprised of organisms which are individual processes, whose ontogenies are controlled by genes which are processes interacting with biochemical pathways. From this perspective life appears to be processes ‘all the way down’ (Bickhardt and Campbell 1999).

For Ghiselin (1974), species are individuals but the central question for biologists is: Individual whats? Given the above arguments, it seems entirely plausible to answer: Individual processes. Ghiselin argues that the process of competition is vital to understanding the nature of species (and firms; recall his economic analogy) and that biologists fail to appreciate the full significance of competition. According to Ghiselin, “A firm is a firm because it forms a closed system of a given kind. It can compete with craftsmen and firms outside itself, and is characterized by a particular kind of internal organization.” For him, species comprise the most extensive units in the natural economy such that reproductive competition occurs among their parts. But for Ulanowicz (2009), cooperation is metaphysically prior; without positive feedback among organisms, there are no processes holding together the system. Nothing in either account—competition or cooperation—precludes the emergent system itself from constituting a process. In fact, a firm or species competes by virtue of what it does, not the matter from which it is formed.

This process-based metaphysics of species in context of Ghiselin’s emphasis on competitive interactions reflects a long-standing ecological principle. Gause’s principle of competitive exclusion holds two species competing for the same resources cannot coexist over the long term, providing that other ecological factors are constant (Townsend et al. 2003). A species may be an individual, but it is not made of material substances. Legs, wings, eyes, muscles, and neurons do not compete—all of these objects are present at the moment of death but there is no competition (or course, two objects cannot occupy precisely the same space at the same time, but this is a trivial physical consideration). Biological competition (or cooperation) entails that physical entities engage in some process. Or that organisms and species are processes—networks of mutually reinforcing processes that persist by acquiring resources for which other autocatalytic systems compete until one or the other prevails (Ulanowicz 2009).

Ereshefsky (1999) describes species as ‘historical entities’, a kind of object-process hybrid. But he also allows that at least some species may be full-blown processes. Asexual species, for which there is no gene flow among organisms, are not individuated by virtue of causal interactions among their parts (Sober’s analysis of individuals notwithstanding). Rather, processes such as selection, genetic homeostasis, and developmental canalization (all of which are echoed in Ulanowicz’s postulates) are what cause organisms to belong to a species. If so, he concludes that, “many species…are merely aggregates of processes” (Ereshefsky 1999, p. 288). The use of ‘merely’ suggests the conventional bias toward material metaphysics, but Ereshefsky seems forced to concede that the individuals that we call species may be just what Ulanowicz (2009) described as historical systems of autocatalytic processes, held together by centripetality and exhibiting a dynamic balance between ascendency and overhead.

4.2.3 Species and Processes as Clusters

Boyd’s (1999) formulation of species as homeostatic property clusters would seem to permit, if not require, that such clusters could constitute processes. Indeed, an account of homeostasis has much to gain by reference to Ulanowicz’s (2009) process ecology. Boyd’s theory entails that properties co-occur via homeostasis, but he provides little explication of the nature of this phenomenon other than an allusion to some underlying processes (e.g., gene exchange, reproductive isolation, and developmental constraints) and, most importantly, a contention that, “the presence of some properties tends to favor the presence of others” (Boyd 1999, p. 143). So what is missing from Boyd’s otherwise compelling conceptualization of species is an account of what holds a cluster together. Ulanowicz (2009) provides two process-based explanations.

First, autocatalysis is a means by which properties (or more to the point, processes) favor others via a positive feedback system. In fact, Ulanowicz (2009) contends that in response to chance perturbations, “autocatalysis acts in a homeostatic fashion to restore [a system]” (p. 77). In a particularly intriguing move, he also makes the case that, “The processes, as a union, make a palpable contribution toward the creation of their constituent elements.” (p. 75). If we take this union of processes to constitute a species, then a species is engaged in the making of its constituent processes which are, of course, organisms.

Second, in light of the uniqueness of even modestly complex ecological systems embedded in an historical context, Ulanowicz rejects the existence of laws and forces as descriptions of the world and embraces Popper’s (1990) concept of propensities. These tendencies “provide a more appropriate glue for holding the world together” (Ulanowicz 2009, p. 11)—and perhaps for holding together Boyd’s clusters of properties. Indeed, Boyd even refers to propensities although he doesn’t explicitly propose that they are the homeostatic glue: “Perhaps the generalizations speak of causal powers and propensities rather than of determinate effects so that it is the causal sustenance of propensities rather than the causation of effects that is relevant” (Boyd 1999, p. 147). Perhaps then, Boyd’s view could accommodate the possibility that because the properties of species are not law-like forces a more viable framing of his metaphysics is that species are homeostatic propensity clusters. Indeed, Ulanowicz (2009) comes very close to just such a proposal: “[A]gencies in living systems are more a matter of configurations of propensities rather than physical forces or their attendant objects” (p. 117).

To push the conceptual merger of Boyd and Ulanowicz to its logical outcome, I suggest taking the final step such that we conceive of species as homeostatic process clusters (properties and propensities being surrogates for the underlying foundation of processes). At least for biologists, properties matter because they have effects. They allow species and their members or parts (organisms) to cooperate and compete. A property is selected in the course of evolution because it is a process, it is part of a process (e.g., horns used in fighting), or it is the material manifestation of some underlying process (e.g., breasts via sexual maturation).

Consider the following case that would seem to unite the metaphysical perspectives of process ecology and property clusters (Fig. 1). There is a creature with a genetic modification that results in sharp, retractable claws. The claws make capturing rabbits efficient, so selection favors this property (i.e., the process of grasping prey). The claws also provide traction, so there is a positive influence on the animal’s speed. Next, the creature’s fur becomes sleek to improve the aerodynamics of sprinting. So selection favors this property (i.e., the process of sprinting). It turns out that sleek fur matches the grassland texture and thereby provides an element of crypsis. Finally, the animal’s posture is modified so that it blends into the background and crawls on its belly to avoid detection (i.e., the process of sneaking). But of course, retractable claws aid in quiet movements that rabbits can’t easily hear. So for this imaginary but biologically plausible creature each of the three processes or properties provides a positive feedback to another. The species is, in this simplified case, an autocatalytic network of processes which might also be viewed as a homeostatic cluster of properties.
Fig. 1

Homeostatic cluster of mutually supporting physical features and biological processes within a hypothetical predator

Boyd (1999) makes the case that species are natural kinds but that they lack a necessary and sufficient, intrinsic, unchanging essence. He refers to species as historical entities, not unlike the process-object hybrid of Ereshefsky (1999). I presume the Boyd uses the language of material objects (i.e., entities) because the metaphysics of material objects is familiar and traditional. However, he makes it clear that, “the causally sustained regularities need not be eternal, exceptionless or spatiotemporally universal” (Boyd 1999, p. 152). They are, in this sense, a series of complex, sequentially structured events—a process. Within a generation, the properties comprising a species interact to hold the cluster together via positive feedback. And the generations themselves can be viewed as sequential events. Once again, it is processes all the way down.

Finally, Boyd (1999) focuses on the nature of evolutionary tendencies toward stasis. As with many others concerned with giving accounts of species, he is understandably concerned with how species persist. The homeostatic processes that he considers (as with the centripetality of Ulanowicz 2009) describe a process in which species would inexorably evolve a set of fixed properties. This seems to be another way of expressing what Ulanowicz refers to as ascendency. What Boyd is missing is an account of novelty or change other than canalization. Here, Ulanowicz’s notion of macroscopic chance and disorder (i.e., system overhead) provides a means by which new species might arise and existing species might develop novel properties.

The problem of accounting for both persistence and change in species was addressed, albeit in somewhat different terms, by another cluster-based theory. Sterelny (1999) discusses the tension between two propensities. On the one hand local selection factors operating on insular (but not wholly isolated) populations would tend to foster novelty and divergence from the more widely distributed species. On the other hand, gene flow from the movement of organisms into and out of the local population (i.e., what has been termed Mayr’s Brake) would tend to dilute the unique features. This evolutionary tension within a species is expressed by Ulanowicz (2009) in terms of the change-stasis processes of his three postulates.

4.3 Reframing Ulanowicz’s Postulates

Having proposed how the existing theories regarding the nature of species can accommodate process-based ontology, I turn to the reciprocal endeavor. That is, I want to argue that Ulanowicz’s postulates of process ecology can be framed in terms not only of ecological systems but of species as well.

4.3.1 First Postulate

The functioning of any system is prone to disruption by chance events, and Ulanowicz (2009) applied this postulate to networks of ecological processes comprising mutualistic communities (e.g., aquatic plants, periphyton and zooplankton). However, this vulnerability seems to pertain to any process or complex configuration of processes. So it stands to reason that the fundamental feature of change applies to species.

Ulanowicz further contends that ecological dynamics are not reducible to the smallest elements of a system according to exceptionless, eternal, and ahistorical laws because these systems are fundamentally non-identical (via Elsasser’s (1969, 1981) argument regarding the number of distinct tokens forming a unique system). If real biotic communities are composed of at least 75 individuals, then a similar case can surely be made for species. So it follows that species are perturbed by both microscopic (e.g., genetic mutations) and macroscopic (e.g., interspecific competition) disturbances.

4.3.2 Second Postulate

Ulanowicz’s (2009) next postulate concerns the capacity of a process to influence itself via the mediation of other processes, and this is the principle means through which systems persist. In broad terms, it seems plausible that a species could engage in interactions with other processes that would provide positive feedback either internally (i.e., among the members or parts of the species) or externally (i.e., via other species as described by Ulanowicz in his accounts of ecological systems).

Autocatalysis is perhaps the most directly relevant aspect of Ulanowicz’s second postulate to the species-as-process metaphysics. In the case of a simple, three-step cycle each element fosters the next through some process. Of course, many instances of autocatalysis in nature involve two steps, such as we commonly see with biological mutualisms. In any case, we can understand the elements as processes, although our affinity for objects typically leads us to conceive of entities connected by processes rather than a pair or triad of processes. If a species is what it does, then each of the steps in the autocatalytic cycle could be understood as a process. For example, eating might foster moving (via energetic reserves) which in turn could enhance reproduction (via mate finding) which could facilitate eating (via provision for offspring) (Fig. 2).
Fig. 2

Autocatalytic cycle operating within a single, hypothetical species

We might also frame this process-based notion of a species in terms that are more familiar to the materialist metaphysics of most biologists. As Ulanowicz notes, autocatalysis can exert selection pressure on its own dynamic constituents. Although he proposes no such examples, there is nothing in process ecology that precludes these constituents from being conspecific. Various forms manifesting different processes may provide autocatalytic systems comprising a species. In some species, developmental stages interact in mutually beneficial ways (e.g., eggs, larvae, pupae, and adults in holometabolous insects; eggs, juveniles and adults in birds) (Townsend et al. 2003). Animals from insects to primates form networks of social processes (e.g., workers, soldiers, and queens in ants; the hierarchy of dominant, subordinate, and juvenile males and females in baboon troops) (Wilson 1980). Some species manifest different reproductive phases to respond to changing conditions (e.g., aphids may be parthenogenic in the summer when genetic variation could impede the exploitation of stable resources, but the insects switch to sexual reproduction when genetic variation in the overwintering eggs provides a hedge against uncertain conditions in the following year (Matthews and Matthews 2009). Species also possess an impressive variety of genders that engage in different processes which form positive feedback systems. For example, red deer have one female and two male genders; bluegill sunfish have one female and three male genders with distinct social and reproductive behaviours; the white-throated sparrow has two male and two female genders with one pair of morphs exhibiting greater aggression and territorial defense and the other providing greater parental care; and the side-blotched lizard has five genders which vary in terms of their social interactions and reproductive processes (Roughgarden 2004).

Ulanowicz’s second postulate also gives rise to mutuality so that cooperation is metaphysically prior to competition, and such an ordering accords with species being processes. That is to say, the coherence of processes that comprise a species must be sufficient to initially ensure persistence such that the species can compete with other species. This is not to say that individual organisms within a species invariably cooperate, as it is evident that intraspecific competition for food, mates, and territory may be intense. However, at least in the sexually reproducing species, there are no competitors if the previous generation did not cooperate with respect to the fertilization of eggs. That males typically compete with one another to fertilize a female is incontrovertible, but the foundation of this competition is the prior fact that neither a male nor a female can reproduce on its own. Even within asexually reproducing organisms, cooperation is common (e.g., microbial mats, slime mold sporangia, and lichens).

The tendency toward centripetality that Ulanowicz (2009) ascribes to ecological systems plays an important role in sustaining and canalizing the cluster of processes that comprise a species, as noted earlier. And centripetality along with selection and long life (which are also features commonly ascribed to species) gives rise to a system’s autonomy. In this light, a species is not only a network of processes but the species is, itself, an emergent process (at least in the sense of Wimsatt 2000). If so, then we ought to be able to at least gesture toward properties of species that are unique or not manifested by the lower order processes of the constituent organisms. Most fundamentally, species (or interbreeding populations or lineages which variously stand in for species) evolve and organisms do not. In addition, species have indefinite existences while organisms have limited life expectances. And along with many other ecological processes, species engage in epidemics and outbreaks which exhibit non-linearities that are not explicable in terms of individual actions (Lockwood and Lockwood 1997) and which manifest features that are not reducible to individuals (e.g., forming swarms, swamping predators and overwhelming host defenses). The classic ecological process of colonization is also a feature of species. Single organisms can be colonists but only insofar as an emergent process of colonization is underway.

The final aspect of Ulanowicz’s second postulate that can be fruitfully reframed in terms of a process-based metaphysics of species is the ascendency/overhead relationship. As noted above in terms of Boyd’s homeostatic property clusters, ascendency accounts both for the tendency of species comprised of many organisms to persist and for the propensity of species to achieve increasingly efficient use of resources via specialization. Conversely, overhead accounts for the plasticity and inefficiency that prevents species from becoming so canalized that they are unable to respond to chance events (e.g., parallel and redundant processes, stochastic spatial distributions and erratic temporal dynamics).

4.3.3 Third Postulate

Ulanowicz’s (2009) ecological metaphysics includes the premise that systems differ from one another as a function of their history which can persist in the form of a pattern of interactions. This aspect of process ecology is readily reframed in terms of species insofar as process philosophy does not obviate material existence, so the ‘material configuration’ can be taken to be the system of organisms and their underlying processes. The configuration of genders, social castes, morphs, and developmental stages within a species all represent historical events stored as physical entities. It is within the historical context of process ecology that Ulanowicz comes closest to explicitly recognizing that species are processes when he notes that, “Furthermore, the mode of recording doesn’t even have to imprint upon a persistent object. History can endure as well through time as a very stable configuration of processes which reestablishes itself whenever the system is disturbed [emphasis added]” (Ulanowicz 2009, p. 69). This would seem to be a fine definition of a species: an historically derived, highly stable configuration of processes that resist exogenously imposed change. If Ulanowicz is right and the trajectory of a system through time is a valid basis for identification, then taxonomy can be understood not as a matter of picking out a necessary and sufficient set of traits but as finding the historical events that have persisted as material configurations.

5 Pluralist, Perspectivist, and Pragmatist Views of Process

Having considered the ways in which existing theories as to the nature of species and Ulanowicz’s framework of process ecology might accommodate the species-as-processes metaphysics, I now turn to a less direct, but perhaps no less important, implication. In particular, I’ll explore how my proposal accords with the notion of species pluralism—a view that is orthogonal to the class-versus-individual-versus-cluster debate.

Ulanowicz (2009) walks a fine line between the claim that process ecology is the single, true metaphysics of nature and that process-based inquiries are just another perspective on the world. He asserts that we must acknowledge “the true order of nature” (p. 145) as if process ecology corresponded uniquely with an objective, mind-independent reality. He goes on to argue that, “One goal of this presentation has been to restore the priority of process over objects in how we see the natural world” (p. 164). From these contentions, one might well conclude that Ulanwicz asserts a kind of absolutism, but he softens his rhetoric by saying, “we are better off in a complex world diverting some attention from objects toward configurations of processes as the key agencies that impart pattern to the world.” (p. 164). And he explicitly offers a domain-specific view of knowledge in reference the viability of Newtonian physics within a set of assumptions that pertain to simple, physical systems. But his endorsement of pluralism seems a bit half-hearted, a compromise with those who might otherwise wholly reject process ecology:

“Should the reader be reluctant to make a clean break with historical foundations, I would hope that he or she would at least entertain the feasibility of viewing phenomena through multiple windows in order to obtain a ‘stereoscopic vision’ that might provide new depth to our understanding of nature.” (p. 10).

With regard to species-as-processes, it would seem that Ulanowicz’s framework would accommodate this metaphysics as a possibility worthy of pursuit by philosophers and scientists. In this regard, his view seems very close to that of Reiners and Lockwood (2010) who developed a philosophy of ecology termed ‘constrained perspectivism’. This adaptation of American pragmatism held that incommensurable views of the natural world within ecology represent partial truths of an objective reality derived from the different, legitimate interests of scientists. Hence, the way the world actually is constrains the veracity of our perspectival claims.

We know if a belief is true by acting accordingly and empirically determining if the results satisfy our interests. The same would hold for the various perspectives regarding species, and Ulanowicz (2009) contends that, “even microbes can be individuated although it serves our interests to ignore some differences and group them into classes, but one should always remain aware that…the members within any such class could always be subdivided when and if necessary” (p. 50). So it would appear that his view of species entails that we carve reality in ways that accord with our interests—and this would clearly entail pluralism. But his pluralism stops short of relativistic anti-realism in much the same way as Reiners and Lockwood’s constrained perspectivism. In the end, there are processes (and objects) out there and when we are wrong about their kinds, the world ‘pushes back’ (e.g., expecting elephants to pollinate wildflowers or attempting to interbreed lions and zebras will empirically fail).

Treating species as processes would also seem to fit readily into Boyd’s (1999) framework. His position is that there is an accommodation between how biologists classify clusters of properties into natural kinds with regard to varied scientific interests and the causal structures in the world. Natural kinds, including species, are discipline relative. For example, a species or other taxonomic grouping for the purposes of a botanist may be very different than a species for the purposes of a horticulturalist. Lilies understood from the perspective of a gardener exclude onions and garlic which are most assuredly lilies in the domain of the plant taxonomist. But the aesthetic interests of the gardener are satisfied through a different classification.

While natural kinds are discipline relative, Boyd is a realist about species. That is, one can be mistaken about membership in classes if a particular individuation is contrary to the causal structure of the world (e.g., the horticulturalist plants a flower garden with garlic bulbs). As with the principles of constrained perspectivism, accommodation is not an ‘anything goes’ venture. Many—but not just any—mind-independent causal relationships exist, and biologists pick out those which accord with their varied lines of inquiry.

While Quine (1969) maintained that classificatory schemes contribute to the development of projectable hypotheses, Boyd goes a step further in that natural kinds are taken to be the accommodation of inferential practices to actual causal structures. A true species then is one which allows the scientist who proposes the species to make generalizations that are appropriately related to the causal structure embedded in the taxonomic classification. In a sense, this reflects the modern pragmatist’s contention that the truth is the compliment we pay to ideas that work—or William James’ notion that, “You can say [of an idea] that ‘it is useful because it is true’ or that ‘it is true because it is useful.’ Both these phrases mean exactly the same thing” (in Reiners and Lockwood 2010). So if framing species as processes results in inductions that lead to inferential practices which fit into existing worldviews and satisfy our interests, then Boyd would presumably endorse this metaphysical view.

In this regard, it is worth noting that for the pragmatists, the constrained perspectivists, and Boyd, knowledge itself is a kind of process insofar as what we know depends on our acting upon a belief in a way that empirically tests its fit with the objective world. As such, the acquisition of knowledge in science, as with the acquisition of adaptive traits in biological lineages, is an evolutionary process. In Boyd’s (1999) terms: “Natural kinds are features not of the world outside our practice but of the ways in which that practice engages with the rest of the world,” (p. 174) a view that seems to resonate with a process-based understanding of both what species are and how we know them. His contention that no natural kinds exist independently of scientific practices suggests that a homeostatic process cluster includes the process of human knowing.

Kitcher (1984) has a rather similar understanding of species pluralism, and it would seem that his view also permits a process-based metaphysics. He holds that species correspond to objective features of nature but different aspects of the world correspond to various research interests. So the correspondence on which the truth of species depends is twofold—fitting a mind-independent aspect of reality and fitting into human interests. As such, there are many relations that could be used to delimit species and none of these has a privileged status as long as the different perspectives meet the criteria of integrating objective features of the world with the subjective desires of scientists.

Somewhat oddly, Kitcher appears to exclude the perspective that species are individuals making his pluralism more restricted than that of Boyd who denies neither classes nor individuals. Presumably Kitcher draws the line because he believes the species-as-individuals concept fails the standard of corresponding with an objective feature of the natural world, although such is not at all clear from his analysis. With respect to a metaphysics in which species are delimited based on processes, Kitcher’s pluralism would be permissive in that it is plausible to contend that sets or classes could have processes as their members. However, processes might be excluded from being extensional sets (which are unchanging) but they might be permitted as classes if one takes these to be intensional sets in which membership can change.

Ereshefsky (1992, 1999) makes a further proposal with regard to pluralism, contending that the single term ‘species’ fails to express the range of uses. He notes that there are at least a dozen species concepts within biology, and much of the confusion lies in having one word to represent highly divergent ways of dividing reality. Along with Grant (1981) who called for using prefixes to specify which of the legitimate interests or domains are reflected in a particular classification (e.g., ‘biospecies’ for those groups based on interbreeding), Ereshefsky proposes designating ‘phyllospecies’ for groups based on phylogenetic relationships and ‘ecospecies’ for groups based on ecological features. In this context, it would seem that ‘actiospecies’ might be a term that Ereshefsky could endorse for groups based on processes—or perhaps for species-as-processes.

6 Some Implications of Species-as-Processes for Biology

The search for viable solutions to the species problem is justified in at least three ways. First, conceptual clarity is worth pursuing insofar as knowledge is plausibly an intrinsic good. Next, as cogently argued by Zagzebski (2008), the claim that someone cares about something (e.g., conservation biologists care about species) entails that those individuals seek to know about the subject of their caring. Finally, ontological commitments in science lead to particular research programs and obstruct or preclude other lines of inquiry. The first two justifications are addressed in virtue of my having engaged this subject, so I’d like to conclude with how the species-as-process perspective might pertain to some of the pressing issues in biology.

6.1 Processes in Applied Biological Sciences

A rich and diverse set of biological fields implicitly treat species as processes, and it would seem that making this metaphysical view explicit could provide a powerful conceptual context for further progress. Medical and toxicological studies are often conducted on organisms that are taken to represent humans in important ways (Shayne 2006). Animals are ‘models’ insofar as their processes align with those of interest to scientists. Mice, dogs, monkeys and other species are chosen in large part because they resemble us with regard to sensitivities to toxins, responses to drugs, proclivities for cancers, etc. What matters is not the material form of the subject of study but what it does in terms of biochemistry, physiology, or behavior. This is why alternatives to animal testing (e.g., cell and tissue cultures) are being developed; they manifest the processes of interest to researchers (Johns Hopkins Center for Alternatives to Animal Testing 2011).

Along similar lines, ecologists and conservation biologists often make use of surrogate species, which are chosen for a variety of reasons including: because they have a cluster of processes that comprise a sensitivity to environmental change (e.g., generation time and metabolic rate), or because what they do is considered representative of other, less well-known or observable, species in their guild, or because their dynamics reflect some more difficult to detect process in the ecosystem (Caro and O’Doherty 1999; Wiens et al. 2008). In any of these cases, the species can be productively understood as the manifestation of some process of interest to the scientist.

The field of restoration ecology (along with the allied processes of rehabilitation, reclamation, and remediation) is concerned with, “the restoration of the fundamental processes by which ecosystems work” (Bradshaw 2000, p. 9). In this sense, the goal can be understood as one of achieving a particular cluster of processes via the actions of species. Ulanowicz’s (2009) process ecology is entirely consistent with this scientific endeavor. But what about a process-based metaphysics of species? In that all of the ecological processes are to some (often very significant) extent a function of the species that are present, restoration ecologists might benefit from adopting a process-based metaphysics. It is clear that restoration can fail if species are perceived as objects or puzzle pieces; ecologists have found that merely putting species back into a system often fails due to the so-called Humpty Dumpty Effect (Pimm 1991). The order in which species are added to a system matters because of what they do. Their interactions with other species are vital, not their being objects of a particular kind.

Conservation biologists who advocate rewilding (including the extreme version of Pleistocene rewilding which entails introducing relatives of large mammals that went extinct 12,000 years ago (Donlan et al. 2006)) often do so with an implicit view that what matters about species is their ecological roles or processes. For example, the wolf is often termed a ‘keystone predator’ such that the rationale for its reintroduction is framed in terms of process (Wilmers 2004). And the opposite end of the environmental management spectrum also depends on an implicit metaphysics of species-as-process. Gnotobiotic (Cairns 1988) or no-analog ecosystems (Parker et al. 2010) are novel, artificial networks of species specifically chosen to perform some function such as toxic waste cleanup.

6.2 Biodiversity and Extinction

The concept of biodiversity in its most simple form is taken to mean the number of species in some spatial context (Wilson 1988; Maclaurin and Sterelny 2008). However, few biologists take this object-counting definition to be the whole story. Rather biodiversity has become enmeshed with discussions of ecological services, ecosystem functions, and evolutionary potentials (Maclaurin and Sterelny 2008). In some ways, this is a return to a process-based understanding of biodiversity that was rooted in studies of the diversity-stability relationship via trophic processes (McCann 2000). As such, the view that species are processes—or at least are elements of more complex processes involving physical and chemical dynamics—has considerable, unrealized potential in this field. Making process-based ontology explicit could have important uses in framing questions about biodiversity.

Ulanowicz (2009) applied measures of ascendency to locating vulnerabilities in complex, dynamic systems. Sensitivity analysis reveals the species, times, and places of greatest vulnerability and allows predictions of the conditions under which a system is most prone to disturbance. If species are autocatalytic processes, then the analytical methods that have been used for ecological networks could be extended to our understanding of the conditions that are likely to precede a species’ extinction.

If a species is what it does, there are at least two important results with respect to our understanding of extinction. First, most species are manifest by a large number of processes—only a portion of which interest biologists. And given a pluralistic view of species, there is no ontologically privileged cluster of processes. As such, an extinction event in conventional terms (i.e., the disappearance of all of the organisms comprising a species) would not be recognized from the perspective of process philosophy if some other organisms performed the same functions. Of course, Gause’s principle of competitive exclusion would mean that no two species are fungible with respect to all of their processes.

Second, a process-based metaphysics leads to the possibility that a species is extinct before all of the organisms which are members or parts have disappeared. It is certainly plausible that whatever emergent processes comprise a species may cease to function when some of its component organisms still persist. I proposed this interpretation with respect to the extinction of the Rocky Mountain locust—along with a possible implication for our understanding of organismal death (Lockwood 2004a, p. 138):

If we choose to describe the locust as a process, there is no doubt that this species was extinct in the late 1800s [the last individual was collected in 1902]. That is, its ecological role and biological activities ceased well before the last, corporeal manifestation disappeared. This notion of life-as-process might seem unusual in a society in which material existence is primary. But such a perception informs our deepest understanding of life. Indeed, life-as-process underlies our notion of euthanasia. When a loved one is simply a body, devoid of the capacity to care, respond, or relate ever again in a way that we can recognize as being “them,” we understand that they are gone even before they are dead.

Of course, process philosophy does not unambiguously support the practice of euthanasia, as one might well argue that rather than using the higher cognitive processes as a basis for what it is to be a person, more biologically fundamental or conceptually holistic processes may be better criteria. In any case, the point is that processes may be a sound basis for deciding when life—whether that of a conventional organism or a species—has ended.

6.3 Biological Concepts

If a process-based metaphysics of species is adopted, the tools that have been developed for process ecology could be productively employed in understanding species of interest (e.g., pests, biological control agents, wildlife, and endangered species). For example, ecologists often represent ecosystems as networks of flows (either energy or matter) in the form of diagrams or graphs showing the connections between the nodes (Ulanowicz 2009). These are further refined in the form of digraphs which reveal the direction of the flows and finally by weighted digraphs showing how much energy or matter flows from one node to another. Perhaps the most familiar of these is the classic food web diagram. But if species are networks of processes, then the same modeling and analytical approaches could be used (e.g., a weighted digraph showing the flow of energy or matter between developmental stages, sexes, phases, morphs, castes, or social ranks). In this way, a deeper—or at least different—understanding of species could be achieved.

If species are processes, then the biologist might well ask whether they are alive. This question has not been raised in more conventional metaphysics, although one might presume that classes and sets are not living while individuals might be alive. Perhaps the claim that species are alive is not so radical with regard to ecological thinking. Most famously and controversially, Frederic Clements (1936) drew a metaphysical parallel between organisms and ecosystems—a concept that connected to the ancient past and continues to resonate in modern times (e.g., Lovelock 2000). Although he skipped over species, it would seem that odd, although logically possible, that organisms and ecosystems are alive but species are inanimate. Ulanowicz (2009) maintains that only those systems that manifest the dynamics and history described in his three postulates are alive. Insofar as it would seem that species can incorporate these features, it follows that species are alive. However, Ulanowicz (2001) seems wary of this possibility and proposes that ensemble living systems (which could well include species, although he doesn’t explicitly say so) exhibiting highly flexible biological behaviors could be called “organic systems”. Although Ulanowicz sidesteps whether these systems of biological processes are themselves alive, thereby avoiding the enmity of ecologists opposed to the organismal metaphor of ecosystems, his metaphysical framework is entirely consistent with species (and ecosystems) being alive.

And finally, process ecology and a process-based metaphysics of species could provide a conceptual bridge between the major realms of biology: ecology and evolution. Ulanowicz (2009) contends that process ecology differs from conventional evolutionary theory on three points: in process ecology selection is endogenous (autocatalysis being the primary agency of selection), while in Darwinian evolutionary theory selection is exogenous. Next, in process ecology systems exhibit a preferred directionality via the canalization of autocatalysis in the absence of perturbation, while most neo-Darwinists flatly reject directionality of any kind. And finally, in process ecology mutuality is fundamental and competition is derivative, while evolutionists hold the inverse view. However, if process philosophy pertains to both ecological systems (as Ulanowicz argues) and species (as I contend), then species—the meta-processes that are vital to both scientific realms—could provide the common ground for a new dialogue between ecologists and evolutionists.


  1. Ayala FJ, Arp R (2009) Contemporary debates in philosophy of biology. Wiley-Blackwell, New YorkCrossRefGoogle Scholar
  2. Bateson G (1972) Steps to an ecology of mind. Ballantine Books, New YorkGoogle Scholar
  3. Bickhardt MH, Campbell DT (1999) Emergence. In: Anderson PB, Emmeche C, Finnemann NO, Christensen PV (eds) Downward causation. Aarhus University Press, AarhusGoogle Scholar
  4. Boyd R (1999) Homeostasis, species, and higher taxa. In: Wilson RA (ed) Species: new interdisciplinary essays. MIT Press, CambridgeGoogle Scholar
  5. Bradshaw AD (2000) What do we mean by restoration? In: Urbanska KM, Webb NR, Edwards PJ (eds) Restoration ecology and sustainable development. Cambridge University Press, CambridgeGoogle Scholar
  6. Brown B, de Laplante K, Peacock K (2011) Philosophy of ecology. Elsevier (North Holland), New YorkGoogle Scholar
  7. Buck RC, Hull DL (1966) The logical structure of the Linnaean hierarchy. Syst Zool 15:97–111CrossRefGoogle Scholar
  8. Cairns J (1988) Increasing diversity by restoring damaged ecosystems. In: Wilson EO (ed) Biodiversity. National Academy Press, Washington, DCGoogle Scholar
  9. Caro TM, O’Doherty G (1999) On the use of surrogate species in conservation biology. Conserv Biol 13:805–814CrossRefGoogle Scholar
  10. Clements FE (1936) Nature and the structure of climax. J Ecol 24:252–284CrossRefGoogle Scholar
  11. Dodds W (2009) Laws, theories and patterns in ecology. University of California Press, BerkeleyCrossRefGoogle Scholar
  12. Donlan CJ, Berger J, Bock CE, Bock JH, Burney DA, Estes JA, Foreman D, Martin PS, Roemer GW, Smith FA, Soule ME, Greene HW (2006) Pleistocene rewilding: an optimistic agenda for twenty-first century conservation. Am Nat 168:660–681CrossRefGoogle Scholar
  13. Elsasser WM (1969) A causal phenomena in physics and biology: a case for reconstruction. Am Sci 57:465–523Google Scholar
  14. Elsasser WM (1981) A form of logic suited for biology? In: Rosen R (ed) Progress in theoretical biology, vol 6. Academic Press, New YorkGoogle Scholar
  15. Ereshefsky M (1992) The units of evolution: essays on the nature of species. MIT Press, CambridgeGoogle Scholar
  16. Ereshefsky M (1999) Species and the Linnaean hierarchy. In: Wilson RA (ed) Species: new interdisciplinary essays. MIT Press, CambridgeGoogle Scholar
  17. Ghiselin MT (1966) On psychologism in the logic of taxonomic controversies. Syst Zool 15:207–215CrossRefGoogle Scholar
  18. Ghiselin MT (1969) The triumph of the Darwinian method. University of California Press, BerkeleyGoogle Scholar
  19. Ghiselin MT (1974) A radical solution to the species problem. Syst Zool 23:536–544CrossRefGoogle Scholar
  20. Glass B (1960) Eighteenth-century concepts of the origin of species. Proc Am Philos Soc 104:227–234Google Scholar
  21. Grant V (1981) Plant speciation. Columbia University Press, New YorkGoogle Scholar
  22. Gregg JR (1950) Taxonomy, language and reality. Am Nat 84:419–435CrossRefGoogle Scholar
  23. Griffiths GCD (1974) On the foundations of biological systematics. Acta Biotheor 23:85–131CrossRefGoogle Scholar
  24. Hull DL (1976a) Are species really individuals? Syst Zool 25:174–191CrossRefGoogle Scholar
  25. Hull DL (1976b) The ontological status of biological species. In: Butts R, Hintikka J (eds) Boston studies in the philosophy of science, vol 32. D. Reidel, DordrechtGoogle Scholar
  26. Hull DL (1978) A matter of individuality. Philos Sci 45:335–360CrossRefGoogle Scholar
  27. Hull DL, Ruse M (1998) The philosophy of biology. Oxford University Press, New YorkGoogle Scholar
  28. Johns Hopkins Center for Alternatives to Animal Testing (2011) Accessed 27 April 2011
  29. Keller DR, Golley FB (2000) The philosophy of ecology: from science to synthesis. University of Georgia Press, AthensGoogle Scholar
  30. Kitcher P (1984) Species. Philos Sci 51:308–333CrossRefGoogle Scholar
  31. Kitts DB, Kitts DJ (1979) Biological species as natural kinds. Philos Sci 46:613–622CrossRefGoogle Scholar
  32. Kolasa J, Pickett STA (1991) Ecological heterogeneity. Springer, New YorkCrossRefGoogle Scholar
  33. Korey K (1984) The essential Darwin. Little, Brown & Company, BostonGoogle Scholar
  34. Lehman H (1967) Are biological species real? Philos Sci 34:157–167CrossRefGoogle Scholar
  35. Lockwood JA (2004a) Locust: the devastating rise and mysterious disappearance of the insect that shaped the American frontier. Basic Books, New YorkGoogle Scholar
  36. Lockwood JA (2004b) Tales from a species’ crypt. Conserv Pract 5:20–27CrossRefGoogle Scholar
  37. Lockwood JA (2010) The fate of the Rocky Mountain locust, Melanoplus spretus Walsh: implications for conservation biology. Terr Arthropod Rev 3:129–160CrossRefGoogle Scholar
  38. Lockwood DR, Lockwood JA (1997) Evidence of self-organized criticality in insect populations. Complexity 2:49–58CrossRefGoogle Scholar
  39. Lovelock J (2000) Gaia: a new look at life on earth. Oxford University Press, New YorkGoogle Scholar
  40. Maclaurin J, Sterelny K (2008) What is biodiversity?. University of Chicago Press, ChicagoGoogle Scholar
  41. Matthews RW, Matthews JR (2009) Insect behavior. Springer, New YorkGoogle Scholar
  42. Mayr E (1976) Is the species a class or an individual? Syst Zool 25:192CrossRefGoogle Scholar
  43. McCann KS (2000) The diversity-stability debate. Nature 405:228–233CrossRefGoogle Scholar
  44. Mishler B, Donoghue M (1982) Species concepts: a case for pluralism. Syst Zool 31:491–503CrossRefGoogle Scholar
  45. Parker KA, Seabrook-Davison M, Ewen JG (2010) Opportunities for nonnative ecological replacements in ecosystem restoration. Restor Ecol 18:269–273CrossRefGoogle Scholar
  46. Pimm SL (1991) The balance of nature? Ecological issues in the conservation of species and communities. University of Chicago Press, ChicagoGoogle Scholar
  47. Popper KR (1990) A world of propensities. Thoemmes, BristolGoogle Scholar
  48. Quine WVO (1969) Natural kinds. In: Quine WVO (ed) Ontological relativity and other essays. Columbia University Press, New YorkGoogle Scholar
  49. Reiners WA, Lockwood JA (2010) Philosophical foundations for the practices of ecology. Oxford University Press, New YorkGoogle Scholar
  50. Rescher N (2008) Process philosophy. Stanford encyclopedia of philosophy. Accessed 26 April 2011
  51. Rheinberger H-J, Müller-Wille S (2009) Genes. Stanford encyclopedia of philosophy. Accessed 26 April 2011
  52. Rosenberg A, McShea DW (2007) Philosophy of biology: a contemporary introduction. Routledge, New YorkGoogle Scholar
  53. Roughgarden J (2004) Evolution’s rainbow: diversity, gender, and sexuality in nature and people. University of California Press, BerkeleyGoogle Scholar
  54. Ruse M (1969) Definitions of species in biology. Brit J Philos Sci 20:97–119CrossRefGoogle Scholar
  55. Ruse M (1971) The species problem: a reply to Hull. Brit J Philos Sci 22:369–371CrossRefGoogle Scholar
  56. Ruse M (1998) Philosophy of biology. Prometheus, New YorkGoogle Scholar
  57. Serafini A (2001) The epic history of biology. Basic Books, New YorkGoogle Scholar
  58. Shapiro G (1973) Habit and meaning in Peirce’s pragmatism. Trans Charles S Peirce Soc 9:24–40Google Scholar
  59. Shayne CG (2006) Animal models in toxicology. Informa Healthcare, LondonGoogle Scholar
  60. Sober E (1993) Philosophy of biology. Westview, BoulderGoogle Scholar
  61. Sterelny K (1999) Species as ecological mosaics. In: Wilson RA (ed) Species: new interdisciplinary essays. MIT Press, CambridgeGoogle Scholar
  62. Townsend CR, Begon M, Harper JL (2003) Essentials of ecology. Blackwell, MaldenGoogle Scholar
  63. Ulanowicz RE (1995) Ecosystem trophic foundations. In: Patterson BC, Jorgensen SE (eds) Complex ecology: the part-whole relation in ecosystems. Prentice-Hall, Englewood Cliffs, NJGoogle Scholar
  64. Ulanowicz RE (1997) Ecology, the ascendent perspective. Columbia University Press, New YorkGoogle Scholar
  65. Ulanowicz RE (2001) The organic in ecology. Ludus Vitales 9:183–204Google Scholar
  66. Ulanowicz RE (2005) Fluctuations and order in ecosystem dynamics. Emerg Complex Organ 7:14–20Google Scholar
  67. Ulanowicz RE (2009) A third window: natural life beyond Newton and Darwin. Templeton Foundation Press, West ConshohockenGoogle Scholar
  68. Vrba ES (1995) Species as habitat-specific complex systems. In: Lambert DM, Spencer HG (eds) Speciation and the recognition concept: theory and applications. Johns Hopkins University Press, BaltimoreGoogle Scholar
  69. Wiens JA, Hayward GD, Holthausen RS, Wisdom MJ (2008) Using surrogate species and groups for conservation planning and management. BioScience 58:241–252CrossRefGoogle Scholar
  70. Wilmers CC (2004) The role of wolves as a keystone species: Examining the ecological effects and conservation implications of a reintroduced top predator on the scavenger guild, Yellowstone National Park. US Environmental Protection Agency website Accessed 27 April 2011
  71. Wilson EO (1980) Sociobiology. Harvard University Press, CambridgeGoogle Scholar
  72. Wilson EO (1988) Biodiversity. National Academy Press, Washington, DCGoogle Scholar
  73. Wimsatt WC (2000) Emergence as non-aggregativity and the biases of reductionism(s). Found Sci 5:269–297CrossRefGoogle Scholar
  74. Zagzebski LT (2008) On epistemology. Wadsworth, FlorenceGoogle Scholar
  75. Zorach AC, Ulanowicz RE (2003) Quantifying the complexity of flow networks: how many roles are there? Complexity 8:68–76CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V. 2011

Authors and Affiliations

  1. 1.Department of PhilosophyUniversity of WyomingLaramieUSA

Personalised recommendations