2.1 Ecology as the study of systems or combinations of species and local abiotic factors

Let us begin by contrasting the classic and species+ views regarding ecology’s taxonomy, or the collection of kinds studied in ecology.Footnote ⁵ Informally, a science’s taxonomy picks out what that science is about. The taxonomy of particle physics includes quarks and neutrinos. Neuroscience’s taxonomy includes neurons, myelin, and spinal cords. The classic theorist views ecology as the study of ecological systems as wholes, while the species+ theorist sees ecology as the study of the interactions of particular species and local abiotic factors. We can use the classic and species+ views regarding ecology’s taxonomy to understand the clash between Clementsian and Gleasonian conceptions of ecological systems and philosophical discussions about the metaphysics of ecological systems.

Let us spell out the classic and species+ positions on ecology’s taxonomy. For the classic theorist, ecological systems are tightly integrated, well-defined units which change in causally robust ways. Ecologists study these systems via pan-system generalizations: generalizations about system-level properties which hold across all ecological systems. For example, a classic theorist may look for generalizations about the universal developmental trajectory of ecological systems, from bare mineral soil to climax state. Frederic Clements (Reference Clements1916, Reference Clements1936; but see Eliot Reference Eliot2007).

In contrast, the species+ view is that ecological systems are nothing more than conventionally designated aggregates of species and local abiotic factors. These species and abiotic factors are in the same area at the same time because of their individual histories and/or physiological needs, not features of the ecological system as a whole. Accordingly, ecologists explain phenomena as the product of interactions between local factors. So, the taxonomy of ecology includes species and local abiotic factors, but not ecological systems. Historically, this view was associated with Henry Gleason, who argued that an ecological system “is not an organism, scarcely even a vegetative unit, but merely a coincidence” (Reference Gleason1926, 16, original italics).

Now for an example: why do coastal live oaks (Quercus agrifolia) live in a particular foothill of California’s Coast Range? The classic theorist explains this phenomenon as an instance of pan-system generalizations. For example, every ecological system develops according to a fixed trajectory, from bare mineral soil to climax state. Our foothill of interest has reached its climax state. The classic theorist then supplies an account of the species composition of an ecological system’s climax state, likely referring to abiotic factors like climate. With this, they have explained why our foothill of interest contains coastal live oaks.

The species+ theorist explains the presence of coastal live oaks in terms of the interactions between the biotic and abiotic residents of that particular foothill. Coastal live oaks need at least X inches of rain to survive, and the foothill in fact receives enough rain. Also, scrub jays live nearby, and they tend to cache and forget acorns, effectively planting oak trees. The confluence of generalizations about local species and abiotic factors explains the presence of coastal live oaks in this foothill.

The clash between the classic and species+ views of ecology’s taxonomy was central to the development of twentieth-century ecology.Footnote ⁶ The classic view was popular during the first half of the twentieth century, but was eventually abandoned due to contrary scientific evidence (Hagen Reference Hagen1992, chapter 1). Contrary to the classic view, ecologists have not found joints in nature for delineating ecological systems. Rather, they rely heavily on pragmatic considerations (Fitzsimmons Reference Fitzsimmons1999). Ecologists have struggled to find pan-system generalizations, such as a Bauplan for ecological system development. Instead, many ecological systems change in disorderly ways (Drury & Nisbett Reference Drury and Nisbet1973). In some ecological systems, species composition is largely driven by individual species’ physiological needs and local abiotic conditions, rather than system-level features (Davis Reference Davis, West, Shugart and Botkin1981; Gleason Reference Gleason1926; Whittaker Reference Whittaker1975). As a result, ecologists have largely moved away from the classic view and towards the species+ view (Simberloff Reference Simberloff1980).

The classic and species+ views have similarly shaped philosophical discussions of ecology’s taxonomy. Philosophers of ecology have discussed the taxonomic status of ecological systems by way of their metaphysics: do ecological systems exist over and above their parts (Lean Reference Lean2018; Odenbaugh Reference Odenbaugh2007, Reference Odenbaugh and Hazlett2010; Sterelny Reference Sterelny2001, Reference Sterelny2006)? To count as a single entity, ecological systems must possess stable boundaries, internal structure, and system-level properties (Lean Reference Lean2018). Philosophers have concurred with ecologists in rejecting the classic view. Contrary to the classic view, ecological systems do not, by their very nature, exist above and beyond their parts. Philosophers generally agree with the species+ claim that an ecological system’s properties should be studied in terms of the causal interactions of local species and abiotic factors, as opposed to pan-system generalizations (Lean Reference Lean2018, 216; Odenbaugh Reference Odenbaugh and Hazlett2010, 246-7). That being said, some ecological systems are composed so that they meet the criteria for existing above and beyond their parts (Lean Reference Lean2018; Odenbaugh Reference Odenbaugh2007; Sterelny Reference Sterelny2006). So, philosophers might resist the species+ claim that ecological systems do not belong in ecology’s taxonomy. In sum, the conflict between the classic and species+ positions on ecology’s taxonomy has structured debates about the nature of ecological systems. Ecologists and philosophers have rejected the classic view and have some, perhaps imperfect, sympathies with the species+ view.

2.2 The internal organization of ecological systems as causally robust or historical accidents

Ecology’s taxonomy is partially determined by the degree to which ecological systems possess a causally robust internal structure. For classic theorists, ecological systems possess a causally robust, universal structure. According to species+ theorists, an ecological system’s internal organization is determined by historically contingent local interactions. The clash between the classic and species+ views has structured scientific controversies about niche theory and philosophical disputes regarding the concept of “function” in ecology.

The classic proponent claims that there are robust patterns in internal organization that are shared across all ecological systems. This view motivated mid-twentieth-century niche theory. Let “niche” refer to the biotic and abiotic conditions that a species requires in order to survive.Footnote ⁷ Some niche theorists analogize an ecological system to a box packed with balls, where the box represents the available resources in the ecological system and each ball is a species’ niche (Ricklefs Reference Ricklefs1987, 167). Niche theorists study how balls are packed; they study pan-system generalizations about niche size and spacing and their relationship to processes like competition and specialization, e.g., the broken stick model (Kingsland Reference Kingsland1985, chapter 8). The classic positions on ecological system organization and ecology’s taxonomy are linked through a shared commitment to pan-system generalizations. If there are pan-system generalizations, then ecological systems are worthwhile units of scientific investigation and likely possess robust internal organization.

For the species+ theorist, the internal organization of an ecological system is a matter of happenstance interactions between individual species and local abiotic factors. As a result, ecological systems are unlikely to possess shared, robust patterns in internal organization. If a particular ecological system possesses a robust internal organization, then ecologists would explain this as the product of interactions between the ecological system’s components. These claims about the internal organization of ecological systems fit naturally with the species+ position on ecology’s taxonomy, according to which an ecological system is nothing more than a collection of individual species and abiotic factors.

The rise and decline of niche theory during the second half of the twentieth century illustrates a general shift from the classic towards the species+ view of ecological system organization in ecology. As an articulation of the classic view, niche theory was the source of major theoretical innovations during the mid-twentieth century, e.g., niche breadth and limiting similarity. However, niche theory soon came under fire as researchers struggled with conceptual and empirical difficulties in measuring niches (Kingsland Reference Kingsland1985). Moreover, ecologists increasingly recognized the importance of place-specific processes, like migration history, in shaping ecological system organization (Ricklefs Reference Ricklefs1987). As a result, during the second half of the twentieth century ecologists shifted away from general theories of ecological structure towards a greater interest in the organization of particular ecological systems.Footnote ⁸ This represents a general shift away from the classic view towards the species+ view among ecologists.

Let us turn to philosophical discussions regarding the internal organization of ecological systems. I will suggest that philosophers have approached this topic indirectly via accounts of the concept of “function” in ecology (Dussault and Bouchard Reference Dussault and Bouchard2017; Lean Reference Lean2021; Millstein Reference Millstein2020; Morrow Reference Morrow2023; Nunes-Neto et al. Reference Nunes-Neto, Moreno and El-Hani2014; Odenbaugh Reference Odenbaugh and Hazlett2010). Although unconventional, I will show that framing the functions literature in terms of the classic and species+ views is useful for interpreting this literature and understanding its stakes. There are two main theories of scientific functions: the causal role theory and the selected effects theory. A causal role (CR) function refers to an entity’s causal contribution to a capacity of the system that it is part of (Cummins Reference Cummins1975). For example, ecosystem ecologists study the capacity of ecological systems to cycle energy. In this context, deer have the CR function of consuming vegetation because this is how they contribute to energy flow. In contrast, an entity has a selected effects (SE) function to F if and only if it was selected to F, where the relevant selection process is natural selection (Millikan Reference Millikan1984; Neander Reference Neander1991). For example, the heart has the SE function to pump blood because it was selected to do so.

A number of philosophers have argued that ecological functions are CR functions, not SE functions (Lean Reference Lean2021; Maclaurin & Sterelny Reference Maclaurin and Sterelny2008, chapter 6; Morrow Reference Morrow2023; Odenbaugh Reference Odenbaugh and Hazlett2010; for dissent, see Millstein Reference Millstein2020). Briefly, ecological functions are not selected effects functions because ecological systems are typically not a unit of selection. On the other hand, CR functions do a reasonably good job of tracking how ecologists use the term “function” (Morrow Reference Morrow2023; Odenbaugh Reference Odenbaugh and Hazlett2010). Ecologists often study how current causal interactions between organisms and abiotic factors contribute to system-level processes. So, a number of philosophers have concluded that ecological functions are CR functions.

The two theories of function map naturally onto the classic and species+ views. Historically, many ecologists assumed that ecological systems possess robust causal organization because they are the product of system-level selection, e.g., Forbes, Elton, Lindeman, and the Odum brothers (Hagen Reference Hagen1992, 151). So, the classic view of ecological system organization implies that ecological functions are SE functions. The species+ view pairs naturally with the CR theory. The species+ theorist views the organization of an ecological system as nothing more than the interactions between the local species and abiotic factors that happen to be part of that system. This suggests that ecological functions are nothing more than the current causal interactions within a system. Similarly, for an entity to have a CR function, it needs to causally contribute to a system-level capacity, without meeting any further requirements for why it has that ability. So, the CR theory of ecological function and a species+ view of ecological organization are natural bedfellows; they both impose minimal requirements on the internal structure of ecological systems. By analyzing ecological functions as CR rather than SE functions, philosophers, like ecologists, have rejected the classic view and adopted a position sympathetic to the species+ view of ecosystem organization.

2.3 (Dis)unity of ecology

Ecology’s taxonomy and the structure of ecological systems shape the structure of science as a whole. Some sciences are unified by a set of core generalizations. For example, the crux of particle physics is the Standard Model, and the crux of evolutionary biology is natural selection theory. Similarly, according to the classic view, ecology is unified by its study of pan-system generalizations, e.g., generalizations about ecological systems’ developmental trajectory. The classic position on ecology’s structure fits naturally with its positions on ecology’s taxonomy and ecological system organization; ecology is unified by a shared commitment to pan-system generalizations because it is the study of ecological systems as well-defined entities with robust internal organization.

According to the species+ view, ecology is a collection of case studies about the interactions between individual species and abiotic factors in particular locations, like North American boreal forests or Midwestern prairies (Shrader-Frechette & McCoy Reference Shrader-Frechette and McCoy1993). The species+ theorist views ecological systems as conventionally designated aggregates of species and local abiotic factors which lack shared, robust internal organization. As a result, ecological systems are studied on a case-by-case basis; these case studies are not unified by cross-system generalizations.

Despite early ambitions to develop pan-system generalizations, ecological practice falls far short of the classic view. Ecology has developed as “a bush with multiple stems” in the form of diverse subdisciplines and study sites, rather than a “a tree with single, well-defined trunk” of core generalizations (McIntosh Reference McIntosh1985, 7). For example, we can turn to the subdiscipline of fire ecology. Ecologists have found few high-level generalizations about the relationship between fire, vegetation, and weather. Rather, fire ecology is largely composed of extended case studies of particular ecological systems, e.g., the Ponderosa pine forests of the Sierra Nevada Mountains and the chaparral of Southern California (van Wagtendonk et al. Reference Wagtendonk and Jan2018). Each region has generated its own literature, with relatively little crosstalk between regions.Footnote ⁹

A number of philosophers have noted ecology’s disunified structure and rejected the classic position on ecology’s structure (e.g., Cooper Reference Cooper2003, chapter 4; Justus Reference Justus2021, chapter 2). Beyond this rejection, relatively few philosophers have offered positive characterizations of ecology’s structure. Shrader-Frechette and McCoy side firmly with the species+ view that ecology is a collection of case studies (Reference Shrader-Frechette and McCoy1993). In contrast, Elliott-Graves argues that ecologists have found “moderate” generalizations that hold across some, but not all, ecological systems (Reference Elliott-Graves2018). I am sympathetic to this view and will build upon it via my account of functional roles (see footnote 4). In short, ecology is a rather disparate discipline. Philosophers have rejected the classic view of ecology’s structure and either endorse the species+ view or an intermediate position.

Let us take stock. According to the classic view, ecology is the study of ecological systems as real, well-defined entities with robust causal features that obey pan-system generalizations. According to the species+ view, ecology is a collection of case studies on how species and abiotic factors interact in particular places. These two views are a useful framework for understanding major debates in ecology and philosophy. Across the board, philosophers have rejected the classic view. In particular domains, they either endorse the species+ view or a moderate position that lies in between both views.

4.1 The account of functional roles

I will first give a high-level overview of functional roles and then iron out various wrinkles in section 4.2. For presentation purposes, I offer a definition of functional roles in terms of necessary and sufficient conditions. These conditions are best understood as a bundle of features that are typically found in ecological kinds. Let F be a functional kind, or a part of ecological systems that is picked out by its causal disposition to interact with other parts of ecological systems in a particular way, e.g. predator. F is a functional role if and only if:Footnote ¹¹

(1) there is a generalization, G, about F, which captures a regularity in interactions among organisms and between organisms and their environment;

(2) G is a cross-system generalization;Footnote ¹² and

(3) multiple species or general abiotic factors are members of F.

All of the examples in section 3 are functional roles. For example, let F be “consumer.” “Consumer” picks out members of an ecological system based on their relationship with other members of the ecological system, namely consumption. Let G be the generalization that three-level food chains tend to exhibit top-down control. Although G does not explicitly mention “consumer,” food chains are built out of consumption relations. In this way, G is about F. G holds across a range of ecological systems, from Yellowstone National Park to the Aleutian Islands. Third, a wide range of species are “consumers,” from caterpillars to cows. So, “consumer” is a functional role. Analogously, other trophic categories, like “producer” and “decomposer,” are also functional roles.

“Predator” and “prey” are also functional roles. They pick out creatures based on their interactions with others, and they are both multi-species kinds. “Predator” and “prey” are featured in cross-system generalizations, like the Lotka–Volterra equations and the life–dinner principle. Symbionts are functional roles as well. For example, “competitor” is a multi-species functional kind that is part of cross-system generalizations, like the evolutionary arms race. “Parasite” and “host” are also multi-species functional kinds that obey a cross-system generalization that high partner fidelity leads to decreased virulence.

Ecosystem ecologists also study functional roles. Generally, they characterize entities functionally in terms of their contribution to a cycle of energy or matter. In the water cycle, “wetland” is a functional role. “Wetland” is defined in terms of its role in mediating between bodies of water and the surrounding land. Wetlands are featured in cross-system generalizations; wetlands are a buffer for flooding. Disturbance ecologists study functional roles too. “Fire embracer” is a multi-species kind that is featured in cross-system generalizations, e.g., fire embracers have high ladder fuel loading. “Foehn wind” is a functional role as well; it picks out a general interaction between air flow and topography. It is part of a cross-system generalization: Foehn winds cause extreme fire behavior.

In short, the examples from the previous section are all functional roles, and so functional roles capture a wide swathe of ecology. These examples show that functional roles are highly heterogeneous; some functional roles may be subsets of or crosscut other functional roles. For example, “herbivore” is a functional role which is itself a type of consumer. As a kind, functional roles are unified by their role in cross-system generalizations.

4.2 Spelling out the three criteria of functional roles

Now that we have a general sense of how functional roles work, let us spell out some details regarding the three criteria of functional roles. We will go out of order. First, functional roles are part of generalizations that hold across multiple ecological systems. Two comments. First, this criterion strikes a balance between the classic and species+ views. Contrary to the species+ view, ecologists have found generalizations that hold across ecological systems. Contrary to the classic view, these are cross-system but not pan-system generalizations; they hold across some but not all ecological systems. Second, a defining feature of science is that it produces general knowledge, or knowledge that holds across a diverse range of circumstances. In ecology, a relevant range of circumstances is different ecological systems. So, it should be unsurprising that central ecological generalizations hold across ecological systems.

Second, the members of a functional role include multiple species or general abiotic factors. This criterion is closely related to the previous one. The scope of a generalization corresponds to a particular level of generality of kinds. Given the structure of the empirical world, cross-system generalizations typically involve kinds whose members include multiple species or general abiotic factors. Two comments on abiotic factors in the context of functional roles. First, I am using “general” abiotic factors in contrast to “local” abiotic factors, or the abiotic factors found in a particular place. For example, Foehn winds are general abiotic factors, which are instantiated by local abiotic factors, like the Santa Ana and Diablo winds. I will not precisify the “general” versus “local” distinction; for our purposes, our intuitive grasp of the distinction will suffice (Carroll Reference Carroll, Edward and Stanford2020). The motivation for focusing on general, rather than local, abiotic factors is to pick out kinds that recur across ecological systems.

Second, we have largely focused on functional roles that include either multiple species or general abiotic factors. Many functional roles include both living and non-living beings. For example, Bond and Keeley proposed that fire acts as a universal herbivore; it indiscriminately consumes the vegetation in its path (Reference Bond and Keeley2005). So, “herbivore” is a functional role that includes different species, like deer and locusts, and general abiotic factors, like fire. In the nitrogen cycle, nitrogen fixation refers to the process of converting atmospheric nitrogen to ammonia. Nitrogen fixation is a functional role that includes living beings, such as rhizobia, and general abiotic factors, like lightning (Odenbaugh Reference Odenbaugh and Hazlett2010). As a branch of biology, ecology is distinctive in its equal attention to both biotic and abiotic factors.

Let us turn to the final criterion: functional roles are featured in ecological generalizations, or generalizations that capture a regularity in the interactions amongst living beings and their abiotic environment. This account of functional roles does not presume any particular account of scientific generalizations, so you can plug in your favorite.Footnote ¹³ Ecologists study a small subset of possible generalizations about functional roles. For example, ecologists do not study generalizations about the walking of insects. This is despite the fact that “walking insect” is a multi-species functional kind that is probably part of ecological regularities. On the other hand, ecologists study generalizations about the walking of ruminants, like cows and sheep. By walking, ruminants compact soil, leading to decreased plant growth (Hamza & Anderson Reference Hamza and Anderson2005). Ecologists study some functional roles but not others due to their scientific and pragmatic interests. Ruminant, unlike insect, walking has predictive and explanatory value for understanding rangeland productivity. More broadly, my claim is not that all functional roles are studied by ecologists. Rather, I am arguing that of the kinds studied by ecologists, many are functional roles. This is compatible with ecologists ignoring many other functional roles because they are uninteresting.

To recap, functional roles split the difference between the classic and species+ views; they pick out causally robust, functional kinds that are more general than individual species and local abiotic factors but more specific than entire ecological systems. Functional roles are featured in cross-system generalizations. A wide range of ecological kinds, from consumers and prey to parasites and fire embracers, are functional roles.

5.1 Functional roles as a central object of study

Recall that the “taxonomy of ecology” refers to the collection of kinds studied by ecologists. Since functional roles are central to ecology, they are part of its taxonomy. My account of functional roles allows us to reorient existing philosophical discussions of ecology’s taxonomy. By approaching the topic via generalizations, rather than metaphysics, we can shift our focus from assessing whether ecological systems are metaphysically real to characterizing the distinctive phenomena studied by ecologists, namely functional roles.

In addition, functional roles serve as a friendly supplement to accounts of the metaphysics of ecological systems. For example, Odenbaugh analyzes ecological systems as composed of entities connected by “ecological relation[s],” like “predator–prey, interspecific competition, mutualism,” and symbiosis (Reference Odenbaugh and Hazlett2010, 241). These relations are none other than functional roles. Since Odenbaugh does not provide an account of “ecological relations,” we can plug in my account of functional roles. More broadly, ecologists study ecological systems in terms of their component functional roles. For example, they study ecological systems as food webs, where creatures interact via consumption, production, and decomposition. In short, functional roles are the scientifically relevant building blocks of ecological systems and have pride of place in ecology’s taxonomy.

5.2 Functional roles as the basis of causally robust ecological system organization

Next, we can use functional roles to characterize the organization of ecological systems. Contrary to the classic view, ecological systems do not possess a universal, tightly integrated internal structure composed of selected effects functions. Contrary to the species+ view, there is more to ecological system organization than the interactions between a happenstance collection of local species and abiotic factors, as captured by the causal role theory. Rather, ecological systems are structured by functional roles: general, functional kinds with robust causal features that recur across ecological systems. I will develop the functional role view of ecological system organization by spelling out its relationship with CR functions.

First, we must attend to some details. When discussing “function” in ecology, philosophers are interested in what it means to “have” a function. Let a member of a functional role, F, have the functional role of performing whatever interaction is definitive of F. For example, kestrels are predators, so they have the functional role of predating, or killing and eating other creatures. Two comments. First, an entity’s functional role is generalization relative. In the context of generalizations about predation, kestrels’ functional role is “predator.” In the context of generalizations about food webs, kestrels’ functional role is “consumer.” Second, the same functional role may be featured in multiple generalizations. Both the Lotka–Volterra equations and life–dinner principle are generalizations about predation. The two generalizations characterize “predator” differently. In the Lotka–Volterra equations, predators are defined in terms of their effect on prey population density, while in the life–dinner principle, predators are characterized in terms of the selection pressure that they exert on prey. That being said, both generalizations pick out high-level patterns regarding the same phenomena, namely creatures that kill and consume others. So, both generalizations are about the functional role of predation.

Functional roles pick out a kind of CR function that is of interest to ecologists. Recall that CR functions are defined in relation to an analytical strategy (Cummins Reference Cummins1975). For example, an investigator may study the heart as part of the body’s circulatory system. In this context, the CR function of the heart is to pump blood. For an investigator of the acoustics of the human body, the CR function of the heart is to make a “whoosh” sound. The CR theory is agnostic regarding the appropriate analytic strategy for picking out functions. As a result, the CR theory struggles to explain why ecologists study some causal interactions, but not others (Morrow Reference Morrow2023).

Functional roles specify a major analytic strategy of ecology: decomposing ecological systems into general functional kinds that support cross-system generalizations. For example, fire ecologists pick out fire embracers as a unit of study, as opposed to the collection of a lodgepole pine, ground squirrel, and two huckleberry bushes. Under the appropriate analytic strategies, both the gerrymandered collection and fire embracers have CR functions. However, ecologists only attribute function to fire embracers because they have robust causal features that recur across ecological systems. More broadly, functional roles take part in ecologically significant causal interactions because they support cross-system generalizations.

Morrow has recently advanced an alternative account of the analytic strategy of ecology: ecologists study CR functions that are robust across the relevant alternate regimes of an ecological system (Reference Morrow2023, 440–41). Let us call these functions “KM functions.” The main difference between KM functions and functional roles is in how they use robustness to pick out ecological kinds.Footnote ¹⁴ KM functions are causal interactions that occur across alternate regimes of the same ecological system, while functional roles are causal interactions that are robust across different ecological systems. Before comparing these two theories, recall that I am a pluralist about ecological kinds. Ecologists study different types of kinds, which may require different analyses. I am happy to acknowledge that KM functions capture an important type of ecological kind. My claim is that given the centrality of cross-system generalizations to ecology, functional roles capture a prevalent and scientifically powerful ecological kind that is not adequately handled by existing theories.

To see why functional roles better capture the predictive and explanatory power of ecology than KM functions, we can return to our example of lodgepole pines as fire embracers. There is a cross-system generalization about fire embracers: they possess adaptations, like high ladder fuel loading, in order to facilitate their own reproduction through high-severity fire. In the United States, the government has systematically suppressed fire for over a century (Pyne Reference Pyne1982). Even in fire-suppressed forests, lodgepole pines still possess adaptations that dispose them to burn, and so continue to fall under generalizations about high-severity fire. In fire-suppressed forests, lodgepole pines continue to have the functional role of facilitating reproduction by spreading high-severity fire. In contrast, lodgepole pines lack the KM function to burn with high severity. For lodgepole pines to have this KM function, they must be able to burn with high severity across all relevant ecological regimes. Since the mid-twentieth century the US has suppressed forest fires, causing profound changes in forest structure and composition. As a result, a relevant ecological regime of lodgepole forests is fire suppression. In fire-suppressed forests, lodgepole pines are largely prevented from burning. Since lodgepole pines lack the ability to burn in a relevant ecological regime, fire suppression, lodgepoles lack the KM function to burn.

The fact that lodgepoles maintain their functional role during fire suppression better captures the scientific use of “fire embracer.” Land managers use generalizations about the anticipated fire behavior of fire embracers when estimating fire risk (e.g., Ager et al. Reference Ager2021). In contrast, lodgepoles lack the relevant KM function. So, KM functions fail to capture how researchers rely on the dispositions of lodgepole pines in prediction and explanation. More broadly, given the centrality of cross-system generalizations in ecology, functional roles better capture how ecologists pick out kinds for the purposes of explanation and prediction.

To summarize, I have argued that functional roles are a form of ecological system organization: recurring, causally robust building blocks that are featured in cross-system generalizations. In comparison to CR and KM functions, functional roles better capture how ecologists pick out kinds for the purpose of formulating explanatorily useful generalizations.

5.3 Ecology as unified by a shared level of analysis

We can turn to the third and final topic addressed by the classic and species+ views, the structure of ecology as a science. The classic view suggests that ecology is unified by a shared commitment to central generalizations. In practice, ecology lacks this structure. This leaves us with the species+ view, according to which ecology is a collection of case studies. The species+ view leaves an important question unanswered: what binds together different case studies into the science of ecology? If ecology is a multi-stem shrub, then what is their shared rootstock? To answer this question, let us return to functional roles. The branches of ecology have a shared level of analysis: robust functional kinds which are featured in cross-system generalizations.

Although generalizations about functional roles are central to ecology, they may be too general for conservation and land management. For example, when modeling fire behavior, fire ecologists don’t just want to know that lodgepoles promote high-severity fire. They want to know the flame height, rate of spread, and combustion temperature. To make predictions at the level of granularity needed for applying ecology to the real world, researchers often need to supplement generalizations about functional roles with place- and species-specific details.Footnote ¹⁵ For example, scientifically informed land management decisions often require a case study on local conditions, such as the Environmental Impact Statements required by the National Environmental Protection Act. This gives applied ecology the case study structure described by Shrader-Frechette and McCoy (Reference Shrader-Frechette and McCoy1993).

To be clear, my disagreement with the species+ view in this context originates from a difference in emphasis, rather than substance. If we want to understand the structure of ecology from the perspective of producing predictions relevant for application, then ecology looks like a collection of case studies. If we want to understand how the science as a whole hangs together, then ecology is unified by a shared level of analysis, namely the study of general, functional kinds.