. índice . Prefacio . Preface . . aguas . 1 . 2 . 3 . 4 . 5 . 6 . . contamina 1 . 2 . 3 . 4 . 5 . 6 . . holocausto 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 10 . 11 . 12 . 13 . . lineas 1 . 2 . 3 . 4 . . hidrotermias 1 . 2 . 3 . 4 . 5 . 6 . . nuevas 1 . 2 . 3 . . Reconquista 1 . 2 . . hidrogeo 1 . 2 . 3 . 4 . 5 . 6 . . esbozos 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . . corredorcentral 1 . 2 . 3 . 4 . 5 . . cordones 1 . 2 . 3 . 4 . 5 . . epiola 1 . 2 . 3 . 4 . 5 . 6 . . deriva 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 10 . 11 . 12 . 13 . 14 . . archivo 1 . 2 . 3 . 4 . . Halcrow 1 . 2 . 3 . 4 . 5 . 6 . . frentehalino 1 . 2 . 3 . 4 . 5 . 6 . 7 . . emicampanaoculto 1 . 2 . 3 . 4 . 5 . 6 . 7 . . Costa del Plata 0 . 1 . 2 . 3 . 4 . 5 . 6 . . Costa del oro 1 . 2 . . IRSA 1 . 2 . 3 . 4 . . flujos . . segmentos . . pendientes 1 . 2 . 3 . 4 . 5 . 6 . . delta 1 . 2 . 3 . 4 . 5 . . propuesta . 1 . 2 . . correconvectivo 1 . 2 . 3 . 4 . 5 . 6 . . plataforma 1 . 2 . . termodinamica 1 . 2 . 3 . . ABL 1 . 2 . . congreso . . girh . . Acumar 1 . 2 . 3 . 4 . . evaluacion 1 . 2 . . BocaRiachuelo 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 10 . 11 . 12 . 13 . 14 . 15 . 16 . 17 . 18 . 19 . 20 . . StoDomingo . . urgenciasatadas 1 . 2 . . inundabaires 1 . 2 . 3 . 4 . . sinsustento 1 . 2 . . emisarios 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 10 . . UAG 1 . 2 . 3 . . áreas nuevas 1 . 2 . 3 . . acreencias 1 . 2 . 3 . 4 . 5 . . audiencia 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . . Valls 1 . 2 . . contrastes 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 10 . . convexterna . . playas 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . . Plan Maestro 1 . 2 . 3 . . Parque Norte . 1 . 2 . . ribera . 1 . 2 . 3 . 4 . 5 . . jurisdiccion 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 10 . 11 . 12 . . CSJNpisamr 1 . 2 . 3 . 4 . . zonas muertas . . Bermejo 1 . 2 . . Pilcomayo . . Samborombon . . Salado . . Uruguay 1 . 2 . . Parana . . Mar del Plata 1 . 2 . 3 . 4 . 5 . . PuntaRasa 1 . 2 . . PuntaMedanos . . Mar Chiquita . . Necochea . . Areco 1 . 2 . . Colonia . . MartinGarcia 1 . 2 . 3 . . Puertos 1 . 2 . . formula1 . . disocio . . senderos . . bajante . . . . oceano 1 . 2 . . hidrolinea 1 . 2 . 3 . . sustentable. 1 . 2 . . agua 1 . 2 . 3 . . antarticflows . . derrame . . luna 1 . 2 . 3 . 4 . 5 . 6 . . index .



Department of Philosophy and Religious Studies
Iowa State University

Department of Philosophy
Environmental Studies
Lewis and Clark College



Philosophers of the life sciences have devoted considerably more attention to evolutionary theory and genetics than to the various subdisciplines of ecology, but recent work in the philosophy of ecology reflects a growing interest in this area (Keller and Golley 2000; Cooper 2003; Ginzburg and Colyvan 2004). However, philosophers of biology and ecology have focused almost entirely on conceptual and methodological issues in population and community ecology; conspicuously absent are foundational investigations in ecosystem ecology. This situation is regrettable.

Ecosystem concepts play a central role in many branches of theoretical and applied ecology, and in environmental literature generally. Indeed, for some historians, the division of ecological theory into populationcommunity and ecosystem research traditions, and the methodological and conceptual debates that have arisen between workers in these respective camps, is the distinguishing feature of 20 th century ecological science (Hagen 1992; Golley 1993).

These include debates over, among others: reductionistic vs. holistic research methodologies; the existence and metaphysics of ecological “kinds”; the relationship between evolutionary mechanisms and ecosystem phenomena; and the nature and scope of ecological science and its relationship to other branches of natural and social science. Philosophers of ecology have written on all these topics, but almost exclusively from the theoretical perspective of population, community or evolutionary ecology.

Philosophical attention to these issues from the perspective of ecosystem ecology is long overdue. It would be misleading to assert that philosophers in general have ignored ecosystem ecology. Environmental philosophers, including environmental ethicists, socalled “radical” environmental philosophers (deep ecologists, social ecologists, ecofeminists, etc.), and policy theorists, have had a longstanding interest in ecosystem ecology (e.g. Callicott 1986; Cahen 1988; Warren and Cheney 1993; Westra 1994; Sagoff 1997; Fitzsimmons 1999).

However, these investigations have typically focused on the ethical and policy implications of accepting certain holistic interpretations of ecosystem science.

When environmental philosophers do address conceptual or methodological issues in ecosystem science, their analysis tends to be narrowly focused on one or two theoretical traditions within ecosystem ecology, and fails to reflect the intellectual background and analytic techniques of contemporary philosophy of science. Despite these weaknesses, the environmental philosophy and policy literature has produced some provocative commentaries on ecosystem science.

Among the most provocative have been the writings of Mark Sagoff. Sagoff’s expertise is in the ethical and economic dimensions of environmental policy, but he has written on the relationship between environmental ethics, policy and environmental science for over twenty years (Sagoff 1985, 1997, 2003). Sagoff believes that (1) due to the complexity, historicity and contingency of ecological phenomena, general ecological theory is not capable of functioning as a sound guide to environmental policy, but that nevertheless, (2) many workers in the area of environmental ethics and policy have been, and continue to be, influenced by holistic conceptions of nature that they have gleaned from the theoretical ecology literature (e.g. that there is a “balance of nature” and that ecosystems exhibit selforganizing behaviors that are directed toward increasing complexity and stability, that ecosystem theory legitimizes such notions as “ecosystem health” and “ecosystem integrity”, etc.).

Sagoff argues that this holistic tradition of ecological science is not scientifically wellfounded, and consequently that environmental policy based on this tradition is misguided and should be challenged.

In his efforts to redirect environmental policy away from an uncritical acceptance of the concepts and theories of theoretical ecology, Sagoff has written several papers that challenge the methodological and conceptual foundations of ecosystem science. In his “The Plaza and the Pendulum: Two Concepts of Ecological Science” (2003), Sagoff argues that ecosystem ecology needs to overcome four “conceptual obstacles” before it may be counted as a successful part of ecological science:

· it should provide a definition of the concept of an ‘ecosystem’ that makes it possible to unambiguously demarcate ecosystems, classify different types of ecosystems, and track changes in the state of an ecosystem;

· it should identify criteria for testing ecosystem theories and models, and implement these criteria in ecosystem research;

· it should provide a plausible account of the causes of ecosystem structure and organization;

· it should be effective in helping to solve pressing environmental problems.

Sagoff believes that, at present at least, ecosystem ecology fails on all four counts, and so should not be regarded as successful ecological science. Indeed, in failing to meet the first two conditions, Sagoff suspects that theoretical ecosystem ecology may fail to qualify as empirical science at all, that it may be better characterized as “a formal science that studies the mathematical consequences of assumptions without regard to the relation of these assumptions to the world” (2003: 529). This claim may come as a surprise to the many professional ecologists who work in this area, but Sagoff is not alone in expressing skepticism about ecosystem ecology, and theoretical ecology generally (Peters 1991; ShraderFrechette and McCoy 1993; Fitzsimmons 1999; Marshall 2002).

While we agree that ecosystem ecology faces a number of conceptual and methodological challenges, we take issue with Sagoff’s characterization of these challenges, and with much of his argumentation. Our primary aim, however, is not to refute Sagoff specifically, but rather to correct some of the common misconceptions and errors that are found in the critical literature on ecosystem ecology, and to use Sagoff’s critique as a vehicle for exploring the philosophical issues raised by this understudied branch of ecology.


The paper is organized as follows:

Section one distinguishes three importantly different senses of the term ‘ecosystem’ and gives a brief overview of the different areas of ecosystem ecology. Ecosystem ecology is a more methodologically and theoretically heterogeneous discipline than is commonly appreciated, a fact that by itself is sufficient to blunt many criticisms that make excessively broad generalizations about the discipline as a whole.

Section two presents and evaluates the background assumptions about the nature of ecological science and ecological systems that Sagoff (among others) brings to his criticism of ecosystem ecology. We characterize these assumptions in terms of three related “dualisms”, and argue that all three present false dichotomies; consequently, arguments that presuppose them are guilty of this fallacy.

In the remaining sections we focus more narrowly on Sagoff’s four objections to ecosystem ecology.

Section three considers Sagoff’s argument that ecosystem theory cannot progress unless ecosystem ecologists can agree on identity conditions for ecosystems. We argue that the field can and has progressed despite a lack of agreement on the status of ecosystems, and that debate over key theoretical terms is not an uncommon situation in science.

Section four addresses the charge that ecosystem ecologists are more concerned with model building than with testing models against empirical data. We grant that different branches of ecology suffer to varying degrees from a lack of productive interaction between theory and data, but argue that the charge is overstated, and that there are important roles for theoretical modelbuilding in ecosystem ecology apart from generation of empirically testable hypotheses.

Section five considers the nature of theoretical explanation in ecosystem ecology and the relationship between evolutionary theory and ecosystem theories. We note that the explanatory goals of ecosystem theories vary across the different branches of ecosystem ecology, but acknowledge that theoretical and explanatory connections between evolutionary mechanisms and ecosystem processes remain an important area for future research.

Section six addresses the charge that ecosystem theory has failed to provide a sound theoretical basis for environmental management and policy. We defend the applicability of ecosystem ecology to environmental problemsolving but question its relevance for assessing the epistemic status of ecosystem theory.



Los filósofos de las ciencias de la vida han dedicado mucha más atención a la teoría evolutiva y a la genética que a las diversas subdisciplinas de la ecología, pero estudios recientes en la filosofía de la ecología refleja un creciente interés en este campo ( Keller y Golley 2000; Cooper 2003; Ginzburg y Colyvan 2004). Sin embargo, los filósofos de la biología y la ecología se han centrado casi exclusivamente en los aspectos conceptuales y metodológicos de la ecología de la población y la comunidad; brillando por su ausencia las investigaciones fundamentales en ecología del ecosistema. Esta situación es lamentable.

Los conceptos en ecosistema desempeñan un papel central en muchas ramas de la ecología teórica y aplicada, y en la literatura del medio ambiente en general. De hecho, para algunos historiadores , la división de la teoría ecológica en población-comunidad y tradiciones de investigación en ecosistemas, y los debates metodológicos y conceptuales que han surgido entre los trabajadores de estos campos respectivos , es la característica distintivo de la ciencia ecológica del siglo 20 (Hagen 1992; Golley 1993 ) .

Estos incluyen debates sobre , entre otros : metodologías de investigación reduccionista vs holística; la existencia y metafísica de "tipos" ecológicos, la relación entre los mecanismos evolutivos y los fenómenos del ecosistema, y la naturaleza y el alcance de la ciencia ecológica y su relación con otras ramas de la física y las ciencias sociales. Los filósofos de la ecología han escrito sobre todos estos temas , pero casi exclusivamente desde la perspectiva teórica de la población , comunidad o la ecología evolutiva .

La atención filosófica a estos problemas desde la perspectiva de la ecología del ecosistema es de larga data. Sería engañoso afirmar que los filósofos en general han ignorado la ecología de los ecosistemas. Filósofos del medio ambiente , incluidos los especialistas en ética del medio ambiente, tambiñen llamados filósofos del medio ambiente "radical" ( ecologistas profundos, ecologistas sociales, ecofeministas , etc ) , y los teóricos de la política , han tenido un interés de larga data en la ecología del ecosistema ( por ejemplo, Callicott 1986; Cahen 1988 , Warren y Cheney 1993 ; Westra 1994; Sagoff 1997; Fitzsimmons 1999).

Sin embargo , estas investigaciones se han centrado habitualmente en las implicaciones éticas y políticas de aceptar determinadas interpretaciones holísticas de la ciencia de los ecosistemas.

Cuando los filósofos medioambientales se ocupan de temas conceptuales o metodológicos de la ciencia de los ecosistemas, sus análisis tienden a ser muy centrados en una o dos tradiciones teóricas en ecología de ecosistemas , y no tiene en cuenta los antecedentes intelectuales y técnicas analíticas de la filosofía de la ciencia contemporánea. A pesar de estas debilidades, la filosofía y la literatura política de medio ambiente ha producido algunos comentarios provocativos sobre ciencia de los ecosistemas .

Entre los más provocativos, los escritos de Marcos Sagoff . La experiencia de Sagoff se descubre en las dimensiones éticas y económicas de la política ambiental , pero ha escrito sobre la relación entre la ética del medio ambiente , la política y la ciencia del medio ambiente por más de veinte años ( Sagoff 1985 , 1997 , 2003). Sagoff cree que (1) debido a la complejidad , la historicidad y la contingencia de los fenómenos ecológicos, la teoría ecológica general no es capaz de funcionar como guía sonora para la política medioambiental, pero que, sin embargo , (2) a muchos trabajadores en el ámbito de la ética ambiental y políticas han sido y siguen siendo , influenciados por las concepciones holísticas de la naturaleza que han recogido de la literatura ecología teórica (por ejemplo, que existe un "equilibrio de la naturaleza "y que los ecosistemas exhiben comportamientos auto organizados que apuntan a complejidad y estabilidad cada vez mayor, que la teoría de los ecosistemas legitima nociones tales como la salud del ecosistema "y " la integridad del ecosistema , etc.)

Sagoff sostiene que esta tradición holística de la ciencia ecológica no está bien fundamentada científicamente, y por lo tanto, la política medioambiental basada en esta tradición es errónea y debe ser cuestionados.

En sus esfuerzos por reorientar la política ambiental lejos de una aceptación acrítica de los conceptos y teorías de la ecología teórica, Sagoff ha escrito varios artículos que cuestionan los fundamentos metodológicos y conceptuales de la ciencia de los ecosistemas .

En su " La Plaza y el péndulo : Dos conceptos de ciencias ecológicas "(2003 ) , Sagoff sostiene que la ecología de los ecosistemas debe superar cuatro " obstáculos conceptuales " antes de que pueda considerarse como parte exitosa de la ciencia ecológica :

· Se debe proporcionar una definición del concepto de un "ecosistema" que permita delimitar de forma inequívoca los ecosistemas , clasificar los diferentes tipos de ecosistemas, y seguir los cambios en el estado de un ecosistema;

· Se deben identificar los criterios para la determinación de teorías y modelos de ecosistemas , y aplicar estos criterios en la investigación de los ecosistemas;

· Se debe proporcionar una explicación plausible de las causas de la estructura del ecosistema y la organización ;

· Se debe ser eficaz para ayudar a resolver los urgentes problemas ambientales.

Sagoff considera que, en la actualidad por lo menos, la ecología de ecosistemas falla en todos los cuatro cargos , por lo que no debe considerarse exitosa a la ciencia ecológica.

De hecho, al no cumplir con las dos primeras condiciones , Sagoff sospecha que la teoría de la ecología de ecosistemas puede fallar por completo para calificar como ciencia empírica; que puede ser mejor caracterizada como "una ciencia formal que estudia las consecuencias matemáticas de supuestos sin tener en cuenta la relación de estos supuestos para el mundo" (2003: 529).

Esta demanda puede resultar una sorpresa para muchos ecologistas profesionales que trabajan en esta área, pero Sagoff no es el único en manifestarse escéptico alededor de la ecología de los ecosistemas y la teóría ecológica en general (Peters 1991; ShraderFrechette y McCoy 1993; Fitzsimmons 1999 , Marshall , 2002).

Al tiempo de estar de acuerdo con que la ecología de los ecosistemas enfrenta varios retos conceptuales y metodológicos, tomamos partido con la caracterización Sagoff de estos desafíos y con gran parte de su argumentación. Nuestro principal objetivo no apunta a refutar específicamente a Sagoff, sino corregir algunas de las ideas falsas y errores comunes que se encuentran en la literatura crítica sobre la ecología de los ecosistemas y utilizamos la crítica Sagoff como vehículo para explorar las cuestiones filosóficas planteadas por esta rama de la ecología poco estudiada.

El documento está organizado de la siguiente manera :

La sección primera distingue tres sentidos diferentes importantes del término "ecosistema" y acerca una breve descripción de las diferentes áreas de la ecología de los ecosistemas. La ecología de los ecosistemas es una disciplina metodológica y teorica más heterogénea de lo que comúnmente se aprecia, un hecho que por sí sólo es suficiente para paliar muchas críticas generalizadoras excesivamente amplias acerca de la disciplina en su conjunto

La segunda sección presenta y evalúa los supuestos básicos sobre la naturaleza de la ciencia ecológica y los sistemas ecológicos que Sagoff (entre otros) acerca a su crítica de la ecología de los ecosistemas. Nosotros caracterizamos a estos supuestos en términos de tres " dualismos " relacionados y argumentamos que los tres presentan falsas dicotomías y en consecuencia, argumentos que suponen que ellos son culpables de esta falacia

En las secciones restantes nos centramos más estrechamente en las cuatro objeciones de Sagoff a la ecología de ecosistemas .

La tercera sección considera el argumento de Sagoff de que la teoría de ecosistemas no puede avanzar a menos que los ecologistas logren acordar en las condiciones de identidad de los ecosistemas. Nosotros sostenemos que el campo puede y ha progresado a pesar de la falta de acuerdo sobre el status de los ecosistemas y el debate sobre términos teóricos clave que no es una situación infrecuente en la ciencia.

La cuarta sección refiere a la acusación de que los ecologistas de los ecosistemas están más preocupados por la construcción de modelos que comprobar los modelos con datos empíricos. Concedemos que las diferentes ramas de la ecología sufren diferentes grados de falta de interacción productiva entre teoría y datos, pero argumentan que la carga es exagerada, y que hay importantes roles en la construcción de modelos teóricos en ecología de ecosistemas, aparte de la generación de hipótesis empíricamente comprobables.

La quinta sección considera la naturaleza de la explicación teórica en la ecología del ecosistema y la relación entre teoría de la evolución y las teorías de los ecosistemas. Advertimos que los objetivos de las teorías de los ecosistemas varían a través de las diferentes ramas de la ecología de los ecosistemas, pero reconocen que las conexiones teóricas y explicativas entre los mecanismos evolutivos y procesos de los ecosistemas siguen siendo un área importante para futuras investigaciones.

El sexto apartado se refiere a la acusación de que la teoría de los ecosistemas no ha proporcionado una base teórica sólida para la gestión ambiental y política . Defendemos la aplicabilidad de la ecología de ecosistemas para la solución de problemas del medio ambiente, pero cuestionamos su relevancia para evaluar el estado epistémico de la teoría de ecosistemas.



Sagoff’s critique is ostensibly directed at all branches of ecology that seek to “identify general rules or principles that govern the assembly, structure, and emergent properties of ecological systems” (2003: 530), but the brunt of his critique is borne by research traditions that posit the existence of ecosystems possessing emergent properties, as represented, for example, by the work of Simon Levin (Levin 1998) and Sven Jorgensen and his colleagues (Jorgensen and Muller 2000). But there are important differences in theoretical orientation between Levin and Jorgensen, and there are branches of theoretical ecosystem ecology that focus on different types of ecosystem phenomena and that use the ecosystem concept in different ways. These different research traditions and different senses of the term ‘ecosystem’ need to be distinguished.

The concept of an ecosystem can be used in the context of an object of scientific study, theories of the nature of such objects, or a general methodology for doing science.

Ecosystem as Object

The ecosystem as object or entity is commonly described as a community of organisms together with its physical environment, or the total set of biotic elements and physico chemical processes present within a particular spatial region. This concept appears in its modern form with Tansley (1935). Ecologists of all theoretical orientations have found the ecosystem concept useful for one purpose or another (Cherrett 1989). However, it is most commonly used in theoretical contexts when there is some intention to study the interrelations of the biotic and abiotic elements in the ecosystem, or to treat the ecosystem as a dynamical unit. We address the metaphysics and semantics of ecosystems in section three.

Ecosystem Theory

Ecosystem theory didn’t emerge until almost ten years after Tansley. The seminal paper is Raymond Lindeman’s (1942) “The TrophicDynamic Aspect of Ecology”, in which Lindeman reinterpreted Charles Elton’s theory of trophic dynamics and community structure as a theory of biogeochemical cycling driven by the capture and conversion of solar energy. The primary focus of ecosystem theory is the description and explanation of the flow of matter and energy in ecosystems.

Ecosystem theory comes in a variety of forms.

If the focus of study is primarily the flow of matter (i.e. carbon, nitrogen, phosphorus, water, etc.) then it is common to use the term “biogeochemistry” to describe this branch of ecosystem theory (e.g. Likens et al. 1977).

If the focus is more narrowly on balances and ratios of chemical elements in ecosystems, then it is called “ecological stoichiometry” (e.g. Sterner and Elser 2002). If the focus is on the flow of energy and the constraints imposed by the laws of thermodynamics on this flow, then one is doing “ecological energetics” (e.g. Weigert 1988).

Finally, if one is primarily interested in studying the organizational and developmental properties of ecosystems taken as wholes, then it is common to describe theories of such properties as belonging to the field of “systems ecology”. Systems ecologists use a variety of formal techniques to describe these structural and dynamical properties (network analysis, control theory, information theory, catastrophe theory, nonlinear dynamics, etc.) (e.g. Halfon 1979; Odum 1994; Patten and Jorgensen 1995; Ulanowicz 1997; Jorgensen 2002).

Of course the differences between these different approaches to the study of ecosystems is a matter of degree, and there is often a good deal of overlap: the study of nutrient cycling and its relationship to (say) total productivity and respiration necessarily involves considerations of energy storage and flow; network descriptions of ecosystems are given in terms of flows of some material or energetic currency (e.g. grams of carbon or kilocalories). Nevertheless, it is not difficult to distinguish the canonical works of those who focus on biogeochemistry from, say, those who focus on abstract network theoretic descriptions of the feedback structure of ecosystems.

To further complicate the picture, there also exist research traditions that are explicitly aimed at investigating relationships between ecosystem processes and population and community dynamics (e.g. recent work on community diversity and ecosystem functioning; see Kinzig et al 2002). These research programs may have a strong ecosystem component but they cross over traditional subdisciplinary boundaries in ecology.

Ecosystem Methodology

There is another important use of the term ‘ecosystem’ that must be considered. When we type “an ecosystem approach to” in our Google search engine we get over six hundred thousand hits. Browsing the first few pages of hits we see the phrase modifying terms like “conservation”, “human health”, “fisheries”, “urban settlements”, “understanding cities”, “public education”, “management”, “human wellbeing”, “the integrity of the Great Lakes”, and “family therapy”. In this context the term “ecosystem” clearly connotes a certain style or philosophy of research rather than a specific object of investigation or a theory of the behavior of such objects.

What is this style of research? It will vary from application to application, but it typically involves distinguishing a focal system of investigation (a fishery, a business, a family unit, a forest, whatever) and situating it explicitly within some broader environmental context. To take an “ecosystem approach” is to assume that at least some of the properties and behaviors of the focal system will depend on interactions or relations between the focal system and its surrounding environment, and that a proper understanding these properties and behaviors will require bringing these system environment relations explicitly within the field of investigation.

The analytic tools that are brought to bear on such investigations will vary from field to field, and will range from informal word or picture models to formal network descriptions and quantitative mathematical models. What such investigations have in common is a belief that anchoring one’s investigations on phenomena operating at only a single spatiotemporal scale risks missing or obscuring the influence of processes operating at smaller and larger spatiotemporal scales on the focal level dynamics.

Within the literature on biological and ecological complexity the term “hierarchy theory”is commonly used to refer to a set of theoretical principles for conducting research on complex systems from a multiscalar perspective (Allen and Starr 1982; O’Neill et al. 1986; Allen and Hoekstra 1992). Though discussions of the compositional hierarchy of atural systems are often taken to have ontological import, hierarchy theory is possibly est understood in epistemological or methodological terms, as a framework for nvestigating complex natural and social systems; one might view hierarchy theory as a et of prescriptions for how to implement an “ecosystem approach to X”.

Thus, we see that the term ‘ecosystem’ has a range of meanings that need to be kept istinct. It is a common strategy among critics of ecosystem ecology to argue, for example, that ambiguities in the ecosystem concept as an object of scientific inquiry somehow, by itself, damages the credibility of ecosystem theories. But the logical relations between these uses are not straightforward. There are different kinds of ecosystem theory, not all of which require or presuppose any particular conception of ecosystems as natural kinds. And some areas of ecosystem theory are better understood as attempts to formalize or situate an ecosystem methodology (e.g. hierarchy theory) rather than as descriptions or models of ecosystem processes (i.e. flows of matter and energy). We elaborate on these points in sections three and four.



It is tempting to now begin our evaluation of Sagoff’s four objections to ecosystem ecology, but this is still premature. Sagoff brings a suite of background assumptions to his evaluation of ecosystem ecology. These assumptions pertain to the whole field of ecology and concern the methodology of ecological science and the ontology of ecological systems. Sagoff’s analysis of ecosystem ecology is framed against the backdrop of these assumptions, and so merit close scrutiny. We argue in this section that these assumptions are relatively common among ecological commentators, and that they should be resisted.

The “Story” of Ecology: Holism versus Reductionism

There is a story that one commonly encounters in writings on the history and philosophy of ecology. It is a story of two competing research traditions, one holistic, the other reductionistic.

In this story, the holists believe that ecological systems exhibit order, structure and regularity at population, community and ecosystem levels of organization, with higher level properties and regularities both emerging out of and constraining lowerlevel properties and regularities. Hence, holists believe the search for lawlike generalizations governing the behavior of populations, communities and ecosystems is a reasonable and desirable goal of ecological research, and formal investigations of community and ecosystem structure are a worthwhile – indeed, indispensable – activity. The ecosystem concept has its home within this broad holistic picture of ecological systems.

The reductionists, on the other hand, believe that ecological systems are nothing more than assemblages of individual species populations whose behavior is primarily determined by response to local environmental conditions. There are no such things as “communities” or “ecosystems” with emergent, causal properties of their own; any properties they have are, at best, epiphenomenal statistical properties of the collection of species populations that compose them. The ecological properties of species populations are best understood in evolutionary terms, as products of natural selection and other evolutionary mechanisms. Consequently, reductionists eschew the search for general laws governing large classes of ecological systems, for it is assumed there are none to discover; rather, their focus is on local, historically contingent, site specific investigations of population behaviors and environmental conditions.

In the opening sections of his 2003 paper, Sagoff makes it clear that he embraces this story. He claims that ecological science pursues two different kinds of inquiry, bottom up and topdown (2003: 531). Bottom up inquiry is experimental and observational research concerning local populations (and maybe communities) that seeks, by induction on data, to uncover the causal processes at work in specific ecological systems. Top down inquiry uses mathematical theory to account for the general structure and function of largescale systems. It employs simplified, abstract representations of ecological systems and uses formal methods to deduce the properties and expected behaviors of whole classes of ecological systems. Thus, we have a set of contrasts: inductive vs. deductive; observational vs. theoretical; smallscale vs. largescale ; case-specific vs. general. Sagoff treats these contrasting methodological strategies as mutually exclusive.

He writes, How can one tell whether the ecological goings on in a lake, forest, or estuary exhibit the kinds of patterns or processes that warrant a theoretical topdown mathematical approach as contrasted with a casebased bottom up inductive inquiry? (531)


No one knows where the future of ecology lies. If the sites that ecologists study are like the swirl of activity at the Luxembourg Gardens, the properties of which are historically contingent, ecology may mature into an inductive and experimental science like medicine… On the other hand, if ecosystems, once def ned and delimited, turn out to be regulated systems like Foucault's pendulum, ecology may become a mathematical and deductive science like physics. (548)

In the above quote we see how Sagoff uses the metaphors of the "plaza" and the "pendulum" to characterize reductionistic and holistic conceptions of ecological reality, respectively. He assumes that ecological systems are either stochastic assemblages of individual species each doing their own thing (like a crowd of people in a plaza), or integrated, wellbehaved dynamical systems with emergent properties amenable to law like generalizations and mathematical description (like a pendulum, or an ideal gas). In his paper, Sagoff identifies theoretical ecosystem ecology with the "pendulum" view.

Though his paper is titled "The Plaza and the Pendulum", what is implied is a choice: the Plaza or the Pendulum.

Sagoff is not alone in endorsing a classification of ecological research as either bottom up or topdown (e.g. Shrader Frechette and McCoy 1993) 1 . Nor is he alone in associating these respective research methodologies with two competing conceptions of the nature of ecological systems. We can summarize this portrait in terms of three dualisms:

· a methodological dualism that assumes that research in ecology is either bottomup or topdown;

· an ontological dualism that assumes that ecological systems are either like the plaza or like the pendulum; and

· a combinatorial dualism that assumes that bottomup research is always associated with the plaza conception of ecological reality, while topdown down research is always associated with the pendulum conception (i.e. that these are the only viable combinations of methodological approach and ontological commitment).

There are elements of truth to the story of ecology as a battle between holists and reductionists, but not enough to support this dichotomous portrait: as generalizations, all three dualisms are false.


Two Examples of Ecosystem Research

The dualisms described above are false for ecology generally, but we will use two examples from the ecosystem ecology literature to help make the point.

The Hubbard Brook Ecosystem Study

Consider the work of ecologists like Gene Likens and Herbert Bormann at the Hubbard Brook Ecosystem Study in New Hampshire. Hubbard Brook is a secondgrowth hardwood forest ecosystem which has a relatively impermeable rock beneath soil. The area is divided into several watersheds where all the runoff goes to several streams. This has allowed ecologists to experiment on different watersheds and provide detailed observational and experimental studies on forest succession and the flow of nutrients in this ecosystem (see Bocking 1997: 11650; see Likens et al 1970 for an example study).

These ecosystem experiments developed in conjunction with a number of theoretical modeling projects. One influential Hubbard Brook model was JABOWA, developed by theoreticians Jim Janak, Daniel Botkin, and Jim Wallis (Janak, Botkin, and Wallis) to determine how the forest would change with time as the result of many different factors.
In effect, the model simulated the growth of individual trees in different species using information from tree competition, reproduction and morality, and physical variables like soil moisture, light, and nutrients.


Regime Shifts in Lake Ecosystems

Consider next the work of Stephen Carpenter and his colleagues on "regime shifts" in lake ecosystems of Wisconsin (Carpenter 2003). A regime shift is a sudden change in the global structure and functioning of an ecosystem. A muchstudied example of a regime shift in lake ecosystems is the shift from oligotrophy to eutrophy (or vice versa).

Eutrophic lakes are rich in mineral and organic nutrients that promote a proliferation of plant life (typically algae) which reduces the dissolved oxygen content and often causes the extinction of other marine organisms. Such algae-dominated lakes are turbid and smelly, low in biodiversity and undesirable for human use.

Oligotrophic lakes are lacking in plant nutrients and have large amounts of dissolved oxygen; they are clear and can support a greater diversity of fish and other marine life, and consequently support a wider range of human uses. Gradual changes in nutrient storages and flows, water quality and species composition occur all the time, but typically within oligotrophic or eutrophic "regimes". Sometimes, however, a lake ecosystem will unexpectedly shift from one regime to the other.

What factors control such shifts? In the case of eutrophic lakes the key variables are relatively well understood. Growth of algae is stimulated by increased levels of nitrogen and phosphorus, but algae are also the base of a food chain that can include micro zooplankton, larger zooplankton, planktivorous (planktoneating) fish, and piscivorous (fisheating) fish.

Phosphorus levels are a key controlling variable, but so is a lake's food web structure, and each are affected by a host of factors, including external phosphorus inputs from soil runoff, habitat alteration, species invasions and fishing practices.

Carpenter and his colleagues have applied both bottomup experimental and topdown theoretical methods in their investigations of the dynamics of regime shifts. They use longterm timeseries data on single lakes and comparative analysis of timeseriesdata on multiple lakes, and conduct wholelake ecosystem experiments that induce regime shifts to test competing hypotheses on the controlling factors and effects of regime shifts (Carpenter et al 2001).

These empirical investigations are accompanied by modeling activities aimed at identifying the key variables responsible for the dynamics of regime shifts. The models can account for many of the features observed in regime shifts, and are commonly used to evaluate and anticipate the effects of management interventions on lake water quality.

With these examples in mind, let us examine the three dualistic claims outlined above.


Against methodological dualism

It is simply false to categorize all ecological research as either topdown work conducted by theoreticians or bottomup work conducted by empiricists. Indeed, one can argue that the best work in ecology is conducted by individuals or teams that involve a close association between theory and empirical studies.

Such appears to be the case in our two examples. In the Hubbard Brook study and the lake ecosystem study we have experiment, observation, and theory are all being used together to study the same ecosystem. Granted, it is often different individuals who carry out the different tasks, but in all sciences there is a division of labor, and ecology is no exception. There are ecologists whose bread and butter is field experiment and natural history, others who are trained for work in the lab, and others who are mathematically knowledgeable and make their home in front of a computer screen. To respond to Sagoff's implicit query, "what is the future of ecology?"; it most certainly lies at the intersection of theory, experiment and observation (Karieva 1989).


Against ontological dualism

Sagoff describes ecology as presenting us with a choice between two mutually exclusive conceptions of the nature of ecological systems. The "plaza" is a direct descendent of Henry Gleason's "individualistic" conception of the plant community that treats each species as a "law unto itself" (Gleason 1926). The "pendulum" has associations with Clements' organismic conception of the plant community (Clements 1916), but also the dynamical, law governed ecosystem perspective of Tansley (1935) and Eugene Odum (1969), and the various traditions of community ecology that posit the existence of emergent communitylevel properties and processes that are amenable to mathematical generalization (MacArthur 1972). No doubt there is some truth in the description of ecology as a field divided between these two competing conceptions of ecological reality.

However, we believe there are no compelling theoretical or empirical reasons to accept this strict ontological dichotomy. The most common theoretical arguments for the Gleasonian ontology are based on two assumptions, (1) that natural selection acts primarily at the level of individuals, and (2) that evolutionary theory is sufficient to explain all of the organization, diversity and complexity observed in the natural world.

The return of multilevel selection theory to respectability (in some quarters, at least) has taken the edge off the first assumption (Wilson 1992; Sterelny 1996), but the major problem is with the second assumption: with respect to the issue at hand it a) begs the question against the existence of organizing principles that are not grounded in evolutionary theory, and b) doesn't entail the Gleasonian individualistic ontology even if it is true; there are plenty of models of selforganization in complex systems that showhow higher level regularities can emerge out of selection acting on locally interacting individuals (e.g. Holland 1995; Watts 1999).

As an empirical matter, while there is considerable evidence that many communities are not the wellordered, clearly delineated, equilibrium structures that tradition associates with Clements, there has never been any dispute that phenomenological regularities exist at community and ecosystem levels; ecology journals and textbooks are full of graphs and charts depicting such regularities. The issues that are debated concern the scope and generality of such regularities, and the mechanisms that explain them, not their existence.

The plaza and the pendulum are more plausibly viewed as ends of a continuum of possible ecological states that exhibit varying degrees of historical particularity, stochasticity, regularity, cohesion and hierarchical organization. Consequently, the law like regularities that are the object of theoreticallyoriented ecological inquiry may vary in domain, scope and "nomic force" (Cooper 2003). Where a given ecological system fits along this continuum is a contingent empirical matter, not a question to be answered a priori . 3


Against combinatorial dualism

Is topdown theoretical research always associated with a conception of communities and ecosystems as holistic entities with emergent properties? Is bottomup research always associated with a Gleasonian individualistic conception of nature? Given our objections to methodological and ontological dualism, these questions are no longer wellposed.

If the general rule is that ecological research involves both topdown and bottomup methods, then simply noting the ontological commitments of ecological researchers will not establish any kind of informative correlation.

Indeed, the fact that individual-and agent-based models are commonly used techniques for investigating higherlevel regularities in complex systems serves to undermine any straightforward contrast between reductionistic and holistic research traditions in ecology. Consider again the Hubbard Brook study. Bormann and Likens sought to investigate the biogeochemical, thermodynamic and community properties of developing ecosystems, and were certainly committed to a holistic methodology for studying watershed ecosystems, but were they also committed to the existence of irreducible organizing principles operating at the ecosystem level? 4

Bormann and Likens believed that ecosystems had homeostatic capacities to respond to disturbance, but their willingness to use gapforest individual based models like JABOWA indicates a willingness to search for explanations of ecosystem level regularities in terms of the properties of the component parts, i.e. what would normally be characterized as a reductionist approach consistent with (though not requiring) Gleasonian individualism.
Philosophers will analyze these cases by pointing out the various different senses in which ecosystems and ecosystem methods may be described as holistic or reductionistic, but such analyses will simply highlight the point being made here, that ecology doesn't support any straightforward mapping between research methodologies and ontological commitments (Odenbaugh 2005). 5

Implications for Sagoff's Argument

The remarks made in this section bear on the general character of debate in ecology, which tends to reflect the false dichotomies described here. But what is the relevance of all this to Sagoff's critique of ecosystem ecology, and to his broader concern with the use of ecological theory in environmental management and policy contexts?

In "The Plaza and the Pendulum" Sagoff is assuming the stance of the toughminded but constructive critic of ecosystem ecology. He refrains from explicitly endorsing the bottom up mode of inquiry with its attendant Gleasonian conception of ecological reality, but in his other writings he is not so equivocal. As noted in the introduction, Sagoff is strongly critical of approaches to environmental ethics and policy that appeal to hypothesized properties of whole ecosystems like "health" or "integrity". He is attracted to a Gleasonian model of ecological reality that, in his view, undercuts all these approaches. Though these motivations are not explicit in this paper, it is clear that Sagoff intends his critique of ecosystem ecology to be persuasive - he doesn't expect ecosystem ecology to succeed in overcoming the conceptual obstacles that he outlines. Sagoff wants to present bottomup inquiry as the only viable methodology for ecological science, and the Gleasonian plaza as the only viable conception of ecological reality.

The considerations raised in this section suggest that Sagoff's overall strategy for undermining ecosystem ecology, and theoretical ecology generally, is ill conceived from the outset. Even if we were to grant that theoretical ecosystem ecology is a failed branch of ecological science and 'ecosystem' too vague or arbitrary a concept to function as a proper object of scientific study, it does not follow that bottomup methods are the only viable mode of inquiry for ecology, nor does it follow that the Gleasonian plaza is the only viable ontology for ecology, nor does it follow that theoretical ecology cannot make valuable contributions to environmental management and policy contexts.

That said, let us now turn to Sagoff's arguments against ecosystem ecology.



Sagoff's first challenge to ecosystem ecology concerns the status of its central concept as an object of scientific investigation "ecosystem". He notes (correctly) that there is no consensus among ecologists on how to define or classify ecosystems. This state of affairs raises two methodological concerns for Sagoff.

First, "No theory can be tested unless it defines the class of objects the behavior of which it seeks to understand" (2003: 535). If we cannot define the kind ecosystem in a satisfactory way, or agree on how to classify ecosystems of different types, then we cannot properly evaluate claims made about ecosystem structure and function. Sagoff believes that a satisfactory definition and classification system must allow ecologists to demarcate ecosystems, distinguish different types of ecosystems, and distinguish changes within an ecosystem type from changes between ecosystem types (53739). Lacking such a classification scheme, ecosystem theories are simply untestable.

Second, the lack of such a classification scheme makes it impossible to properly distinguish substantive empirical claims about ecosystems from mere analytic tautologies, claims that are true by definition. Consider for example the following definition by Eugene Odum:

[An ecosystem] is a unit of biological organization made up of all of the organisms in a given area (that is, 'community') interacting with the physical environment so that a flow of energy leads to characteristic trophic structure and material cycles within the system (1969: 262)

Now consider the following question: "Do ecosystems have a characteristic trophic structure that is determined by energy flow through the system?" Given this definition, the answer is trivially yes, by definition. But surely this is a substantive empirical claim about ecosystems, requiring evidence and argument for its justification, not a matter to be settled by postulation.

On the other hand, if we truncate this definition to resemble something more commonly encountered in ecology textbooks -"An ecosystem is a community of organisms interacting with the physical environment" - then we end up with a definition that excludes almost nothing; "a kitchen sink, a head full of lice, a yeast infection, an orchard … a dung pile or whale carcass …" would all qualify as ecosystems (Sagoff 2003: 5378). Such a definition includes entities that most ecologists would hesitate to call ecosystems, but more importantly, it threatens to render the concept meaningless by making it universal in scope.

This second set of objections amounts to an additional constraint on any proposed classification scheme for ecosystem ecology; it must define the kind ecosystem in a way that avoids the problems of overinclusive and underinclusive definitions. Sagoff claims that ecologists have yet to offer any classification scheme that avoids both these problems.


Definitions, Classification and Testability

Does a lack of consensus on the definition of a fundamental scientific kind render theories of such kinds untestable? There seem to be plenty of counterexamples. The concept species is fundamental to evolutionary biology, but one can find over a dozen distinct, well motivated definitions of the species concept in the biological literature that carve up the biological world in different ways (see Claridge et al. 1997; Ereshefsky 2001; Stamos 2003). 6 At least three of these concepts are fairly well known to philosophers of biology:

· the biological species concept, which defines a species as a group of organisms that can interbreed and produce fertile offspring;

· the phylogenetic species concept, which defines a species in terms of evolutionary lineage or ancestry; and

· the ecological species concept, which defines a species as a lineage occupying a distinct ecological niche.

There is also persisting debate among biologists and philosophers of biology over whether we should keep searching for a single, correct species concept (socalled "species monism") or whether we should abandon such a search and accept a plurality of species concepts as equally correct and useful for different purposes ("species pluralism"). Should we conclude, then, that evolutionary theory is untestable?

What about the concept of "gene"? The classical gene concept is a unit of function, something that controls or affects the phenotype. All the early geneticists picked out individual alleles by beginning with a phenotypic trait and defining the associated gene as "whatever segment of nucleic acid that causes the trait to take the form it does". But the complexities of gene expression and gene regulation are such that there is yet no consensus among geneticists about how to identify such segments, and none expected (Kitcher 1982; Rosenberg 1985, chapter 4; Moore 2001). Should we conclude that classical genetics is untestable?

Consider mass, surely one of the most fundamental concepts of physics. Do physicists share a single, unambiguous concept of mass? Undergraduate textbooks may convey this impression, but physicists recognize a plurality of mass concepts - "inertial mass", "active gravitational mass", "passive gravitational mass", "relativistic mass", etc. - about which there remains considerable debate about their natures and their empirical and logical relations to one another (Jammer 2000). Should we conclude that physical theories that employ this concept are untestable?

These are rhetorical questions, of course. All of these areas of science have enjoyed considerable theoretical and empirical success. It is simply a mistake to think, as Sagoff appears to, that consensus opinion on the definitions of scientific kinds and is necessary for a successful science that refers to those kinds. To see how such success is possible in the presence of ambiguity and debate over the fundamental natures of scientific kinds, we need to understand - contra Sagoff, who asserts that "[t]he first step in a theoretical science is taxonomy" (2003, 538), (emphasis added) - that scientific classification is an on going and theoryladen enterprise, not something that scientists do prior to the development and testing of theories. Definitions of scientific kind terms are a product of science, and they often have multiple senses that reflect the multiple theoretical contexts in which they arise and the multiple uses to which they are put.

To illustrate further, consider the concept of temperature (Sklar 1999). The first temperature concepts arose as modes of description of immediate experience (e.g. temperature as hot or cold). In the early days of thermodynamics this sense was replaced with a meaning fixed by the measuring instruments used to determine quantitative values (e.g temperature as that which is measured by a gas thermometer). As equilibrium thermodynamics became formalized, temperature acquired a meaning in terms of the role it played in basic thermodynamic laws (e.g. via its roles in the zeroth and second laws of thermodynamics, physicists introduced the concept of "empirical" and "absolute" temperature, respectively). 7

With the rise of statistical physics the temperature concept came to be applied more broadly to characterize equilibrium states of systems, sometimes of wildly divergent kinds (e.g. a blob of matter and a region of electromagnetic radiation can share the same equilibrium temperature). Temperature has an even more important role in statistical mechanics, however, "as a way of characterizing ensembles, that is as a parameter appearing in the appropriate probability distribution over the microscopic states of individual systems" (Sklar 1999: 195). We are long way from "hot" and "cold"!

This kind of conceptual elaboration and pluralism is the rule in theoretical science, not the exception, and it is necessary for theoretical and empirical advancement. Thus, we should be neither surprised nor dismayed to find a plurality of ecosystem concepts and classification schemes at work in the ecological literature, given the multiple uses and contexts in which the term can and has been used.

We can use this point to address some of the concerns expressed by Sagoff over the use of overinclusive definitions of ecosystems. Earlier we distinguished the ecosystem as an object of scientific study from ecosystem theories on the one hand, and ecosystem methodologies on the other. It is important to see that classifying spatial units into ecosystem types is only one possible role that the ecosystem concept might play, and one that may or may not be a relevant or useful activity for an ecosystem researcher.

When so called "hierarchy" theorists write about the ecosystem concept, for example, they are often employing the concept in a methodological sense, as denoting a set of conceptual and empirical strategies for investigating complex natural systems ("a way to explain a part of the universe selected for study" (O'Neill et al 1986: 38)), not as a basis for classifying landscape units, nor as a theoretical concept within a particular tradition of ecosystem theory describing energetic and material flows, i.e. biogeochemistry, ecological energetics, systems ecology, etc.

The "ecosystem strategy" involves situating a given phenomenon of interest within a hierarchical spatial and temporal context, decomposing the system into component parts that reflect "natural" boundaries (via comparison of characteristic process rates), studying and modeling the relationships between the system components across varying scales, etc. (see O'Neill et al 1986: 75 100). This method is sometimes used to define ecosystem as objects, but it is important not to conflate the process and the product. It is common, though, for ecosystem theorists to shift back and forth between the various different senses of 'ecosystem' without being careful to distinguish them, with sometimes confusing results.

Sagoff is also concerned with under inclusive ecosystem definitions. His main concern is that these render certain substantive empirical hypotheses about ecosystems untestable because they follow immediately from the definition. But how does the fact that they are part of a definition make them untestable? One of the important lessons that Hilary Putnam and Saul Kripke have taught philosophers is that definitions in science are often empirical hypotheses.

The claim that "water is H2O" is a claim about the essence, constitution or nature of a substance or kind water. The question "Is water made of H2O?" has a trivial answer just in case water is constituted by H2O and one knows this. However, the claim that a given liquid sample is water is not a trivial claim, nor is the claim that it consists in H2O. These are both testable claims.

The same is true in the case of ecosystems. The definition given above by Odum, if correct, renders the question "Do ecosystems have a characteristic trophic structure?" trivially true. However, the claim that a group of biotic and abiotic components is an ecosystem is not trivial (there can be reasons for treating or not treating some collection of biotic and abiotic components as an ecosystem), nor is the claim that they exhibit a particular trophic structure or a particular material flow.

To sum up, we reject the claim that, because there is no consensus on a definition of the ecosystem concept or on a classification scheme for ecosystems, hypotheses about ecosystem structure, function or behavior are untestable.



Sagoff offers the following criticism of mathematical theory as it is developed in ecosystem ecology:

Theories of ecosystem structure and function confront a second problem simply because there are so many of them. The abundance of untested mathematical theory threatens to turn ecology into a formal rather than an empirical science … (2003: 536).

He also writes, Other sciences tend to replace one theory with another, for example, the phlogiston with the oxidative theory of combustion. Ecology, in contrast, seeks to add one theory to another. (2000: 541)

Here, Sagoff isn't arguing that ecosystem theories and models are untestable (though of course he suspects they are), but that ecosystem ecologists seem unwilling or uninterested in testing them. This is a claim about the methodological practices and attitudes of ecologists, not a claim about ecosystem theories.

Sagoff supports this charge by noting the diversity of theoretical approaches represented in a collection of essays titled Handbook of Ecosystem Theories and Management (Jorgensen and Muller, 2000). There he finds chapters that describe ecosystems as "Functional Entities", as "Self Organizing Holarchic Open Systems", as "Subjects of SelfOrganizing Processes", as "Dynamic Networks", as "Hierarchical Systems", as "Chaotic Systems", as "States of Ecological Successions", and so on ... (541)

But Sagoff sees no evidence among the authors of any desire to prune such lists; rather, the sole aim appears to be to rationalize how they can all be true and equally useful ways of understanding ecosystems. Sagoff's problem isn't with pluralism per se, but rather with what he views as "unconstrained production of theory that lacks relevance to empirical puzzles or problems" (542).

There are several things to say about this charge.

First, it doesn't apply equally to all branches of ecosystem ecology. As noted earlier, ecologists who study mineral, nutrient and element flows are in general more likely to have their modeling activities be closely allied to empirical data and research programs than, say, theoreticallyoriented systems ecologists . The collection that Sagoff cites to illustrate the plurality of untested ecosystem models is firmly within the systems ecology tradition and thus is not necessarily representative of the theoretical modeling activities of ecosystem science taken as a whole (cf. the Hubbard Brook and "regime shift" examples).

Second, even restricting ourselves to systems ecologists, it is not obvious that there is a general lack of concern among systems ecologists with issues of testability. In the introductory chapter of the above volume Jorgensen and Muller write, Ecosystem research provides the data needed for our understanding of ecosystems because without data our models cannot be tested and could easily give a flawed image of the ecosystem. Models developed in a data vacuum are (almost) useless - with perhaps a few exceptions. There should therefore be strong interactions between ecological modelling and ecosystem research. (2000: 9).

Jorgensen himself has tried to make a point of drawing out empirical predictions from his own thermodynamic model of ecosystem function and comparing them with empirical data sets when possible (e.g. Jorgensen et al 2002; Marques and Jorgensen 2002).

Granted, close associations between systems modellers and empirical ecosystem researchers are not as common today as they were in the 1960s during the heyday of largescale, federally funded ecosystem modelling projects, but there are many factors other than methodological complacency that might account for this situation (see Hagen 1992; Golley 1993).

Third, systems ecologists (indeed, scientists generally) employ formal methods for avariety of purposes, not all of which are aimed at generating falsifiable empirical predictions (Canham et al 2003). These may include, for example, analysis and synthesis of existing empirical data, exploration of formal properties of real and hypothetical systems, forecasting and scenario planning for management purposes, etc. To get a real sense of how empirical data bears on the theoretical activities of systems ecologists one must survey a broader sample of the literature than Sagoff cites.

Fourth, even among systems ecologists there remains confusion about the nature of the evidence that would bear on ecosystem theory in this tradition. Systems ecologists are often charged with creating formal models that ignore biologically relevant detail, but it is important to understand the specific explanatory aims of these models. Systems ecologists are perhaps better described as "complex systems" ecologists; they seek to describe and explain macrolevel patterns in the structure and behavior of ecosystems that may be characteristic of certain generic classes of complex systems, like self organization. A striking feature of such complex systems patterns is that they can often be realized in systems of different kinds (physical, chemical, biological, ecological, etc.).

To give just two examples: 1) the same critical point phenomena observed in phase transitions in gases and fluids can be observed in the transition from ferromagnetic to paramagnetic state in magnetic materials; and 2) the same "perioddoubling route" to chaotic dynamics has been observed in systems as diverse as fluids, chemical clocks, electrical circuits, lasers and acoustic systems (Lesne 1998). These and other complex systems behaviors have the following generic features (see Batterman 2002: 13):

· The details of the system (those details that would feature in a complete causal mechanical explanation of the system's behavior) are largely irrelevant for describing the behavior of interest.

· Many different systems with completely different "micro" details will exhibit the identical behavior.

What do fluids, chemical systems, electrical circuits, lasers and acoustic systems have in common that would explain their common period doubling route to chaotic dynamics?

Whatever it is, it cannot have much to do with the specific material properties of the components that make up these systems. Any explanation must refer to the relational or structural features that the systems have in common - in short, it must abstract away from the "matter" to identify the underlying "form" that is common to all the systems in question. In this sense, one can justifiably call a science that attempts to describe and explain such features a "formal science" (see Franklin 1994; de Laplante 1999).

Sagoff uses the term "formal science" to disparage ecosystem theory, but systems ecologists might do better to embrace it, for it helps to clarify much that is distinctive of systems ecology, and it offers a principled response to critics who charge ecosystem theory with using formalisms and constructing models that ignore biological detail.

If ecosystems do have emergent "formal" properties characteristic of complex systems, they will not be explained by appeal to the details of the composition and behavior of constituent components. Rather, they will be explained by the organizational properties common to a generic class of such systems, and it is through modelbuilding that these organizational properties are identified. This line of response certainly doesn't resolve all the challenges facing systems ecology but it does help to rebut the charge that such models lack empirical relevance simply in virtue of being abstract.



Sagoff's third challenge to ecosystem ecology imposes a condition on what a satisfactory theory of ecosystem organization and structure should be able to do.
If ecologists succeed in establishing a general theory to describe the design, function, structure, organization, etc. of ecosystems, they must identify the force, power, or agency that constitutes the efficient cause of that design or order. (543)

Sagoff argues that ecosystem ecologists have yet to offer any plausible mechanism that would generate the order and organization that ecosystem theorists claim is to be found at the ecosystem level. He considers two candidates for such a mechanism, self organization and evolution. Let us look at his objections to these proposed mechanisms in turn.


Self Organization

Sagoff writes:
One might reply that ecosystems organize themselves, but this is problematic. There are objects, such as crystals, that may be said to be selforganizing, but they have no obvious similarity with ecosystems. The mathematician Per Bak (1996) identified what he called self organized criticality in the way piles form when sand is dripped on a plate. One cannot think of two items more unlike, however, than, say, a forest and a dribbled pile of sand. ... That dribbled sand piles display selforganized criticality suggests nothing whatsoever about ... ecosystems. (543)

Sagoff appears to be arguing that, because the model systems that are commonly used to illustrate and analyze self organizing dynamics (like Bak's sandpile) are very different from real world ecosystems, then nothing about the actual or potential behavior of ecosystems can be inferred from such model systems.

This objection reveals a profound ignorance of the explanatory goals and methods of complexity theories. The key points were introduced in the previous section but we can say a bit more here. A (not necessarily exhaustive) list of "complex systems" or "complexity" sciences would include the following:

· the theory of universality of critical state phenomena in equilibrium statistical mechanics and solid state physics (Domb 1996);

· the theory of universality in non linear dynamical systems and deterministic chaos (Cvitanovic 1996);

· the theory of universality classes in cellular automata theory (Langton 1995);

· theories of self organized criticality in non equilibrium dynamical systems, cellular automata, and "small worlds" networks (Kauffman 1993; Bak 1996; Newman et al 2003)

· theories of the selforganization of dissipative structures in far from equilibrium thermodynamic systems (Nicolis and Prigogine 1977; Schneider and Kay 1994);

· network and other "goal function" approaches in systems ecology (Muller and Leupelt 1998; Jorgensen 2002)

What all these theories have in common is the following: they seek to describe and explain the formation of macro level patterns in systems composed of many interacting micro level constituents or components. The patterns may be spatial, temporal, or more abstractly dynamical. It is important to distinguish complex systems phenomenology - the identification and description of observable patterns - from theoretical frameworks that attempt to explain this phenomenology. Both are part of complex systems theory, though the theoretical component is more developed in some disciplines than in others.

As can be seen from the list, theories of selforganization are a part of complexity theory but they do not exhaust the category, and there are various distinct types of self organization theory - selforganized criticality theory (or SOC theory, as it is often called) is just one of them. Mathematical ecologists have borrowed from all areas of complexity theory in their efforts to describe and explain ecological patterns and processes.

In the case of SOC, the relevant phenomenology involves power law distributions. Power laws appear to show up everywhere, from the magnitudes of earthquakes, to the flow of rivers, to light from quasars, to highway traffic (Bak 1996: 21). They have the same general characteristics, with the value of a variable inversely proportional to the frequency of occurrence of that value. Such distributions are not random, and their presence in systems of radically different types, and at radically different spatial and temporal scales, begs for an explanation.

Power law distributions can be generated by superimposing periodic signals of all frequencies and time scales, indicating a selfsimilar "fractal" structure in the spatial and temporal correlations of the system generating the signal. Such correlations are the distinctive signature of critical state phenomena observed in phase transitions in equilibrium thermodynamic and statistical systems.

Away from the critical state, the system response to fluctuations is linear and localized; a small change has only a localized effect that decays exponentially in space and/or time. As a critical temperature is approached the correlation length increases and goes to infinity at the critical temperature, and the system's response to fluctuations becomes nonlinear and potentially global, with the result that a small change can cascade throughout the system, causing dramatic fluctuations at all length scales. 8

However, equilibrium systems do not spontaneously evolve to critical states, they require a control parameter to be "tuned" to the critical value, "by hand" (e.g. turning up the heat under a beaker of water). Bak was concerned with power law phenomena in nonequilibrium systems where there is no "external" tuning of control parameters. He argued that the ubiquity of power law phenomena could be explained if, under very general conditions, nonequilibrium dynamical systems spontaneously evolve to a critical state. This is what the claim of "self organization" amounts to in SOC - the system naturally gravitates toward a critical state without the need for external tuning.

This is a bold hypothesis. Bak and his colleagues (1987) introduced a toy mathematical model (the famous "sandpile" model) that, they believed, captured the relevant general conditions, and showed that it exhibited spontaneous evolution toward a critical state with no characteristic length scale, and argued that the general conditions should be widely instantiated in natural systems.

There has been much experimental and theoretical work done on SOC since 1987, and many claims for the theory have been challenged and qualified, but there is a consensus that under certain conditions, SOC behavior can be expected. In particular, the system must be a "slowly driven, interactiondominated threshold system" (Jensen 1998: 126). Basically, if a system is slowly forced away from equilibrium (slow relative to the natural relaxation time of the system), and the interactions between the components dominates the dynamics of the system, and the system is locally stabilized by the existence of potential thresholds that must be overcome before the system changes configuration, then it should exhibit SOC behavior. SOC models specify the form that the physical constraints must realize for the complex systems phenomenology to be observed.

This is one sense in which ecosystems may be said to "self organize". Contrary to Sagoff's casual dismissal, this is a perfectly legitimate sense in which an ecosystem can be similar to a sandpile. There are other models of self organization that may apply to ecosystems as well. The relevant scientific question is whether we have any reason to think that ecosystems do in fact self organize in any of these ways. In the case of SOC, it involves looking for evidence of power law distributions and selfsimilar scaling relations among ecological variables, and ruling out alternative explanations of such distributions that don't entail selforganization to the critical state. This is a challenging research program that has yet to yield conclusive results, but it is a perfectly coherent research program (e.g. Keitt et al 1996; Milne 1998; Jorgensen at al 1998).


Simon Levin has argued that an evolutionary perspective is essential to understanding the structure and functioning of ecosystems (Levin 1999). He treats ecosystems as "complex adaptive systems", the essential elements of which are

· Diversity and individuality of components: This feature also implies that there are mechanisms, such as mutation or genetic recombination, that continually refresh diversity

· Localized interactions among the components: In natural systems, these interactions include processes such as competition for food, predation, and sexual reproduction.

· An autonomous process (such as natural selection) that uses the outcomes of those local interactions to select a subset of those components for replication or enhancement. (1999: 12)

Levin believes that large scale patterns in ecosystem organization arise not from processes operating at the level of whole ecosystems, but rather from the localized interactions of components subject to selection processes. In this respect he can be viewed as a "bottomup" complexity theorist, in contrast with "topdown" complexity theorists who posit organizing principles operating at the system level.

Sagoff presents an interesting objection to the hypothesis that natural selection on the individual species components of ecosystems is responsible for the organizational properties of ecosystems. He asks us to consider the difference between a) recently created ecosystems (such as a lake created by a hydroelectric dam) and ecosystems comprised of mostly exotic species (due to biological invasion or conscious design), and b) ancient ecosystems that contain only indigenous organisms.

If evolutionary forces, which work slowly, build up ecosystems, this process would take time to accomplish. Evolutionary forces, then, would not appear to be responsible for organizing recently created ecosystems and those that comprise mostly exotic species. (2003: 544)

Thus, we should expect to see a difference in the organizational structure of "heirloom" ecosystems from either recent ecosystems or ecosystems with a high proportion of exotic species.

However, Sagoff claims, we don't see such differences. He gives two examples to illustrate. Deep Creek Lake in Maryland was created in 1920 by a hydroelectric dam.

The people who settled around the newly flooded valley introduced whatever fish species they thought desirable, and fish populations persist to the present day. Sagoff reports that in a biological study conducted on the lake in the 1970s (Howell et al. 1978), ecologists concluded that no structural differences distinguish this lake, created recently by a corporation, from one that "evolved" or selfassembled over millennia. The youthfulness of the lake did not detract from its dynamic properties. (2003: 544)

In the second example, Sagoff considers the San Francisco Bay. The waters in this area are dominated by invasive or exotic species. Sagoff surmises that if largescale patterns result from evolution at lower levels, the presence or absence of these patterns would indicate whether a site is free of or full of introduced species. Yet, no one can tell whether or which species are exotic except by consulting the historical record. (544)

Sagoff generalizes from these examples to conclude that the effects of natural selection on ecosystem structure and function are either unobservable or non existent.

We agree that vulnerability of ecological communities to biological invasion and establishment of exotic species is prima facie evidence against a conception of communities as stable, tightly connected species assemblages. And we strongly agree with Sagoff that ecosystem ecology must direct more resources toward establishing the empirical consequences of ecosystem hypotheses (in this case hypotheses concerning the effects of natural selection on ecosystem structure and function) and checking these against the empirical evidence. The line of reasoning that Sagoff is presenting here is an interesting one and should be developed.

However, we view the specific examples used here as inconclusive. Ecological systems exhibit varying degrees of resistance and vulnerability to invasion (Naeem et al 2000), a situation more consistent with the "graded" conception of ecological systems we are advocating here (i.e. of ecosystems exhibiting differing degrees of order and randomness) than with an extreme Gleasonian individualism. Of the Deep Creek Lake example one is tempted to ask how ecologistscould have known what structure to expect of a lake that had "self assembled over millennia”, to compare with the structure of the recently created lake.

More importantly, the "structural differences" in question must be elaborated in order to assess the plausibility of the claim that such differences should be expected in heirloom and recent ecosystems. Hypotheses of this type will also depend on particular views concerning the relationship between evolutionary and ecological dynamics, and there is wide variance on this issue among theorists (cf. Levin 1999, chapter 8; Allen et al 2003, chapter 6; OdlingSmee et al 2003, chapter 5).


Another False Dichotomy?

Sagoff sets up his discussion of the causes of ecosystem organization by offering two choices, self organization or evolution by natural selection. But many theorists would view this as a false dichotomy. Bruce Weber and David Depew offer a helpful survey of different views on the relationship between selforganization and evolution among theorists (Weber and Depew, 1996). They distinguish seven positions:

1. Natural selection, and not self organization, drives evolution. (traditional Darwinism, e.g. Dawkins 1976))

2. Self organization constrains natural selection. (e.g. Lewontin 1970, Gould 1982)

3. Self organization is the null hypothesis against which evolutionary change is to be measured. (e.g. Burian and Richardson 1991, Kauffman 1993)

4. Selforganization is an auxiliary to natural selection in causing evolutionary change.(e.g. Wimsatt 1986)

5. Self organization drives evolution, but is constrained by natural selection. (e.g. Brooks and Wiley 1988, Goodwin 1989, Salthe 1993)

6. Natural selection is itself a form of selforganization. (e.g. Swenson 1989, Schneider and Kay 1994)

7. Natural selection and self organization are two aspects of a single evolutionary process. (Weber and Depew 1996, via Kauffman 1993 and Wicken 1987)

Sagoff may view such lists as further evidence of unchained theoretical promiscuity, but the point being made here is simply this: the question of the origins of order in ecosystems is bound up with the questions that drive the theoretical projects listed above.

We are looking at one direction that the future of biological and ecological thought might take, and it would be hasty to bet on the outcome at this stage.



Sagoff states that "the most convincing way to demonstrate the legitimacy of mathematical ecosystem theory is to show how it can be applied to help solve problems of environmental policy" (2003: 545), but argues that ecosystem theory has failed to make any substantive contributions in this area. By "environmental policy" Sagoff is including a) policies for conserving threatened species and maintaining or restoring biodiversity, and b) policies for protecting, maintaining or restoring the "health" or "integrity" of ecosystem.

Why is ecosystem ecology an ineffective guide for environmental policy? Sagoff cites the skeptical judgments of ecologists who argue that the generalizing tendencies of ecological theory are illsuited to the task of predicting "which species will be found at which site and which species will not coexist. The necessary research ... generally falls under the rubric of natural history." (Simberloff et al 1999). "Ecosystem management" is also concerned with preserving and maintaining biodiversity, but via the broader management goal of maintaining the "integrity" of whole ecosystems (Grumbine 1994).

Sagoff argues that terms like ecosystem "integrity" and "health" presume that we can establish a baseline of "normal" or "desirable" ecosystem functioning against which to judge deviations, but that in the absence of a proper definition of the class ecosystem, or an accepted classification scheme for ecosystems, or agreement about the meaning and measurement of emergent ecosystem properties, there is no way of operationalizing such concepts. These are not new arguments, of course (Calow 1992; Shrader Frechette and McCoy 1993; Lackey 2001).

We view the applicability of ecology to environmental problemsolving as highly dependent on the context in which an environmental problem or issue is addressed. If we are asking about the prospects for survival of a particular endangered species, site specific information will be vital to answering the question with any reliability, and general ecological theory will likely function only in the background, perhaps in the form of management heuristics for situations similar to the one under consideration.

On the other hand, if we are asking about, say, the importance of phosphorus inputs and stocks for water quality issues, or the effects of acid rain on forest regeneration, or the relationship of biodiversity to ecosystem productivity or resistance to invasion, or the impact of habitat destruction or fossil fuel consumption on global nutrient cycles, then the specific predictions of higherlevel theories and models have a more obvious role to play. And indeed, such theories and models have been influential in directing attention to environmental problems that might otherwise have gone unnoticed (e.g. Vitousek et al 1997; Driscoll et al 2001).

Here we wish to make two additional points concerning the application of ecological theory to environmental problem solving.

First, we agree with Sagoff and other critics, like Schrader Frechette and McCoy (1993), that environmental policy issues are fundamentally decision problems, not problems of scientific inference; they require rational consideration of competing individual and social values within conditions of (sometimes radical) factual uncertainty. But then the right answer to the question “Should ecological theory determine environmental policy”? is going to be “No”, regardless of the empirical credentials of the theories under consideration, since the kind of factual information that science provides is only one component of a more complex ethical and sociopolitical problem context. Thus, we agree with Sagoff that there are perils in formulating environmental policy that is driven by a particular conception of ecological reality (ecosystem or otherwise), to the exclusion of ethical and sociopolitical
issues, but one can agree with this point without denigrating ecological theory.

Second, while one might argue that if ecosystem theories were to prove effective in environmental problem solving then this would help to support the empirical adequacy and explanatory power of ecosystem theory, the converse doesn’t seem to follow: if ecosystem theories prove to be empirically adequate and explanatorily successful, does it follow that they should also prove effective in resolving environmental problems? Why should they?

This view seems to presuppose that environmental problems are primarily the result of scientific ignorance, that if we better understood the ecological impacts of our activities we would see the error of our ways and change our behavior. Such a view is at least contentious. If environmental problems are ultimately tied to the effects of neoclassical economics, global capitalism or gender inequality, as many environmentalists and environmental philosophers suggest, then the changes that will be most effective in resolving environmental problems will be ethical and sociopolitical in nature, not scientific. Ecological science may well prove ineffective in resolving environmental problems, but it doesn’t follow from this that ecological theory is lacking in predictive or explanatory power.



This paper had three broad aims. The first was to bring to the attention of philosophers some of the interesting methodological and conceptual features of ecosystem ecology that distinguish it from other branches of ecological science, with the hope of stimulating further work in this area.

The second was to highlight some common misconceptions about the nature of ecological science and ecological reality that frequently appear in the philosophical and scientific literature on ecology.

The third was to respond to specific criticisms of the ecosystem concept and ecosystem theory that have been raised by environmental writers concerned with the influence of holistic ecosystem science on environmental management and policy.

We used Mark Sagoff’s (2003) critique of ecosystem ecology as a vehicle for exploring these interrelated issues. Sagoff presents a dualistic conception of ecological science and ecological reality that associates “topdown” theoretical ecology with a conception of ecological systems as highly ordered (“the pendulum”), and “bottomup” empirical ecology with a conception of ecological systems as random, disordered assemblages (“the plaza”). Similar dualistic schemes are widely represented in the ecological literature.

We used examples from ecosystem ecology to show that this dichotomous portrait is not supported by the actual practices of ecologists, and that arguments that presuppose such a portrait are unsound.

Concerning Sagoff’s specific objections to ecosystem ecology, we defended the scientific credentials of ecosystem ecology against objections based on 1) the plurality of ecosystem concepts and classification schemes, 2) the plurality of ecosystem models and formalisms, 3) difficulties in identifying a plausible mechanism for ecosystem organization, and 4) the perceived ineffectiveness of ecosystem ecology in solving environmental problems. In our responses to 1 and 2 we argued that conceptual and methodological pluralism is well represented in many successful sciences and consequently is not, by itself, a mark of theoretical or methodological impoverishment.

In our response to 3 we suggested that the problem facing ecosystem ecology is not that there is a lack of plausible mechanisms that might explain ecosystem organization, but that there are too many – theories of ecosystem organization are undetermined by the available data, not excluded by them.

Finally, in our response to 4 we defended the general applicability of ecosystem theory to environmental problem solving, but resisted the assumption that the inability of ecological theory to resolve environmental problems is a sign of theoretical inadequacy.



Allen, T.F.H., and Hoekstra, T.W. (1992) Toward a Unified Ecology. New York: Columbia University Press.
Allen, T.F.H., and Starr, T. (1982) Hierarchy: Perspectives for Ecological Complexity. Chicago: University of Chicago Press.
Allen, T.F.H., Tainter, J.A., and Hoekstra, T.W. (2003) SupplySide
Sustainability. New York: Columbia University Press.
Bak, P. (1996) How Nature Works. New York: SpringerVerlag.
Bak, P., Tang, C., and Wiesenfeld, K. (1987) “SelfOrganized Criticality: An Explanation of 1/f Noise”. Physical Review Letters 59: 381384.
Batterman, R. (2002) The Devil In The Details: Asymptotic Reasoning In Explanation, Reduction, And Emergence. New York: Oxford University Press.
Bocking, S. (1997) Ecologists and Environmental Politics: A History of Contemporary Ecology. New Haven, Conn.: Yale University Press.
Brooks, D.R. and Wiley, E.O. (1988) Evolution as Entropy: Toward a Unified Theory of Biology. Chicago: University of Chicago Press.
Burian, R., and Richardson, R.C. (1991) “Form and Order in Evolutionary Biology: Stuart Kauffman’s Transformation of Theoretical Biology”. PSA 1990 2: 26787.
Cahen, H. (1988) “Against the Moral Considerability of Ecosystems”. Environmental Ethics 10: 195216.
Callicott, J.B.(1986) “The Metaphysical Implications of Ecology”. Environmental Ethics 8: 30116.
Calow, P. (1992) “Can Ecosystems be Healthy? Critical Consideration of Concepts”. Journal of Aquatic Ecosystem Health 1: 15.
Canham, D.C, Cole, J.J. and Laurenroth, W.K. (eds.) (2003) Models in Ecosystem Science. New York: Princeton University Press.
Carpenter, S.R. (2003) Regime Shifts in Lake Ecosystems: Pattern and Variation, in O. Kinne (eds.) Excellence in Ecology 15, Oldendorf/Luhe: Ecology Institute.
Cherrett, J. M. (1989). “Key Concepts: The Results of a Survey of Our Members’ Opinions”, in J.M. Cherrett, (ed.) Ecological Concepts: The Contribution of Ecology to an Understanding of the Natural World, Oxford: Blackwell, pp. 116.
Claridge, M.F., Dawah, H.A, and Wilson, M.R. (eds) (1997) Species: The Units of Biodiversity. London: Chapman & Hall.
Clements, F.E. (1916) Plant Succession: An Analysis of the Development of Vegetation. Washington D.C.: Carnegie Institution.
Clements, F.E. (1928) Plant Succession and Indicators: A Definitive Edition of Plant Succession and Plant Indicators. New York: H.W. Wilson Company.
Cooper, G. (2003) The Science for the Struggle for Existence. Cambridge: Cambridge University Press.
Cvitanovic, P. (ed) (1996) Universality in Chaos, 2 nd Ed. Philadelphia: Institute of Physics Publishing.
Dawkins, R. (1976) The Selfish Gene. Oxford: Oxford University Press.
de Laplante, K. (1999) “Certainty and Domain Independence in The Sciences of Complexity: A Critique of James Franklin’s Account of Formal Science”. Studies in History and Philosophy of Science 30: 699720.
Domb, C. (1996) The Critical Point: A Historical Introduction to the Modern Theory of Critical Phenomena. London: Taylor & Francis Ltd.
Driscoll, C.T., Lawrence, G.B., Bulger, A.J., Butler, T.J., Cronan, C.S., Eager, C., Lambert, K.F., Likens, G.E., Stoddard, J.L. and Weathers, K.C. (2001) “Acidic Deposition in the Northeastern U.S.: Sources and Inputs, Ecosystem effects, and Management Strategies”. BioScience 51: 180198.
Erechefsky, M. (2001) The Poverty of the Linnaean Hierarchy: A Philosophical Study of Biological Taxonomy. Cambridge: Cambridge University Press.
Fitzsimmons, A. (1999) Defending Illusions. Lanham, MD: Rowman and Littlefield Press.
Franklin, J. (1994) “The Formal Sciences Discover the Philosophers’ Stone”. Studies in History and Philosophy of Science 25: 513533.
Ginzburg, L, and Colyvan, M. (2004) Ecological Orbits: How Planets Move and Populations Grow. New York: Oxford Unviersity Press.
Gleason, H.A. (1926) “The Individualistic Concept of the Plant Association”. Bulletin of the Torrey Botanical Club 53: 120.
Golley, F. (1993) A History of the Ecosystem Concept in Ecology: More Than the Sum of the Parts. New Haven: Yale University Press.
Goodwin, B.C. (1989) “Evolution and the Generative Order”, in Goodwin B. and Saunders, P. (eds) Theoretical Biology: Epigenetic and Evolutionary Order from Complex Systems. Edinburgh: Edinburgh University Press.
Gould, S.J. (1982) “Darwinism and the Expansion of Evolutionary Theory”. Science 216: 38087.
Grumbine, R.E. (1994) “What is Ecosystem Management?” Conservation Biology 8: 27 38.
Hagen, J.B. (1992) An Entangled Bank: The Origins of Ecosystem Ecology. New Brunswick, NJ: Rutgers University Press.
Halfon, E. (1979). “Theoretical Systems Ecology: Advances and Case Studies.” Academic Press, New York.
Holland, J. (1995) Hidden Order: How Adaptation Builds Complexity. Reading, MASS: Addison Wesley.
Holman, J.P. (1988) Thermodynamics, 4 th Edition. New York: McGrawHill.
Howell, J. et al. (1978) Deep Creek Lake: A Biological Study. Alleghany Community College, Mimeograph.
Jammer, M. (2000) Concepts of Mass in Contemporary Physics and Philosophy. Princeton: Princeton University Press.
Jensen, H. J. (1998) SelfOrganized Criticality: Emergent Complex Behavior in Physical and Biological Systems. Cambridge: Cambridge University Press.
Jorgensen, S. E. (2002). Integration of Ecosystem Theories: A Pattern. Third Edition, Dordecht: Kluwer.
Jorgensen, S.E., Marques, J.C., and Nielsen, S.N. (2002) “Structural Changes in an Estuary, Described by Models and Using Exergy as Orientor”. Ecological Modelling 158: 23340.
Jorgensen, S.E., Mejer, H., and Nielsen, S.N. (1998) “Ecosystems as Self Organizing Critical Systems”. Ecological Modelling 111: 26168.
Jorgensen, S.E. and Muller, F. (eds) (2000) Handbook of Ecosystem Theories and Management. Boca Raton, FL: Lewis Publishers.
Karieva, P. (1989) “Renewing the Dialogue between Theory and Experiment in Population Ecology”, in J. Roughgarden, R. May, and S. Levin. (eds.) Perspectives in Ecological Theory. Princeton: Princeton University Press.
Kauffman, S.A. (1993) The Origins of Order. Oxford: Oxford University Press.
Keitt, T.H., and Marquet, P.A. (1996) “The Introduced Hawaiian Avifauna Reconsidered: Evidence for Self Organized Criticality?” Journal of Theoretical Biology 182: 16167.
Keller, D.R. , and Golley, F.B. (eds) (2000) The Philosophy of Ecology: From Science to Synthesis. Athens: University of Georgia Press.
Kinzig, A., Pacala, S. and Tilman, D. (eds) (2002) The Functional Consequences of Biodiversity. Princeton: Princeton University Press.
Kitcher, P. (1982) “Genes”. British Journal for the Philosophy of Science 33: 33759.
Kitcher, P. (1985) “Two Approaches to Explanation”. Journal of Philosophy 82: 63239.
Kitcher, P. (1989) “Explanatory Unification and the Causal Structure of the World”, in P. Kitcher and W.C. Salmon (eds.) Scientific Explanation, Vol. 13, Minnesota Studies in the Philosophy of Science, Minneapolis: University of Minnesota Press, pp. 410506.
Lackey, R.T. (2001) “Values, Policy, and Ecosystem Health”. BioScience 51: 437443
Langton, C. (1986) “Studying Artificial Life with Cellular Automata”. Physica D 22.
Langton, C.G. (ed) (1995) Artificial Life: An Overview. Cambridge, MA: MIT Press.
Lesne, A. (1998) Renormalization Methods: Critical Phenomena, Chaos, Fractal Structures. New York: John Wyley & Sons.
Levin, S.A. (1998) “Ecosystems and the Biosphere as Complex Adaptive Systems”. Ecosystems 1: 43136.
Levin, S.A. (1999) Fragile Dominion: Complexity and the Commons. Reading, MA: Perseus Books
Lewontin, R.C. (1970) “The Units of Selection”. Annual Review of Ecology and Systematics 1: 118.
Likens, G., Bormann F.H., Johnson, N. M., Fisher, D. W., and Pierce, R. S. (1970) “Effects of Forest Cutting and Herbicide Treatment on Nutrient Budgets in the Hubbard Brook Watershed Ecosystem”, Ecological Monographs 40:2347.
Likens, G.E., Bormann, F.H., Pierce, R.S., Eaton, J. S., and Johnson, N.M. (1977). Biogeochemistry of a Forested Ecosystem. New York:SpringerVerlag.
Lindeman, R.L. (1942) “The TrophicDynamic Aspect of Ecology”. Ecology 23: 399 418.
MacArthur, R.H. (1972) Geographical Ecology. New York: Harper & Row.
Marques, J.C., and Jorgensen, S.E. (2002) “Three Selected Observations Interpreted in Terms of a Thermodynamic Hypothesis”. Ecological Modelling 158: 21322.
Marshall, A. (2002) The Unity of Nature: Wholeness and Disintegration in Ecology and Science. London: Imperial College Press.
Milne, B.T. (1998) “Motivation and Benefits of Complex Systems Approaches in Ecology”. Ecosystems 1: 44956.
Moore, D.S. (2001) The Dependent Gene: The Fallacy of “Nature vs. Nurture”. New York: Times Books.
Muller, F., and Leupelt, M. (eds) (1998) Eco Targets, Goal Functions and Orientors. Berlin: SpringerVerlag.
Naeem, S., Knops, J.M.H., Tilman, D., Howe, K.M., Kennedy, T., and Gale, S. (2000) “Plant Diversity Increases Resistance to Invasion in the Absence of Covarying Extrinsic Factors”. Oikos 91: 97108.
Newman, M. E. J., Barabasi, A. L. and Watts, D. J. (Eds). (2003) The Structure and Dynamics of Complex Networks. Princeton: Princeton University Press.
Nicolis, G. and Prigogine, I. (1977) SelfOrganization in NonEquilibrium Systems: From Dissipative Structures to Order Through Fluctuations. New York: J. Wiley & Sons.
Odenbaugh, J. (2005) The 'Structure' of Population Ecology: Philosophical Reflections on Unstructured and Structured Models, in B. Beisner and K. Cuddington (eds.) Paradigm's Lost: Theory Change in Ecology, Academic Press)
OdlingSmee, F. J., Laland, K.N., and Feldman, M.W. (2003) Niche Construction: The Neglected Process in Evolution. Princeton: Princeton University Press.
Odum, E.P. (1969) “The Strategy of Ecosystem Development,” Science 164: 262 – 270.
Odum, H.T. (1994) Ecological and General Systems: An Introduction to Systems Ecology. University Press of Colorado.
O’Neill, R. V., DeAngelis, D. L., Waide, J. B., and Allen, T. F. H. (1986). A Hierarchical Concept of Ecosystems. Princeton University Press, Princeton.
Patten, B.C., and Jorgensen, S.E. (eds) (1995) Complex Ecology: The Part Whole Relation in Ecosystems. Englewood Cliffs, NJ: Prentice Hall.
Rosenberg, A. (1985) The Structure of Biological Science. Cambridge: Cambridge University Press.
Sagoff, M. (1985) “Fact and Value in Ecological Science". Environmental Ethics 7: 99 116.
Sagoff, M. (1997) “Muddle or Muddle Through? Takings Jurisprudence Meets the Endangered Species Act”. William and Mary Law Review 38: 825993.
Sagoff, M. (2003) “The Plaza and the Pendulum: Two Concepts of Ecological Science”. Biology and Philosophy 18: 52952.
Salmon, W.C. (1989) “Four Decades of Scientific Explanation”, in P. Kitcher and W.C. Salmon (eds.) Scientific Explanation, Vol. 13, Minnesota Studies in the Philosophy of Science, Minneapolis: University of Minnesota Press, pp. 3219.
Salthe, S.N. (1993) Development and Evolution: Complexity and Change in Biology. Cambridge, MA: Bradford/MIT Press.
Schneider, E.D. and Kay, J.J. (1994) “Life as a Manifestation of the Second Law of Thermodynamics”, Mathematical and Computer Modelling 19: 2548.
ShraderFrechette, K. and McCoy, E. D. (1993) Method in Ecology. Cambridge: Cambridge University Press.
Simberloff, D., Stone, L, and Dayan, T. (1999) “Ruling Out a Community Assembly Rule: The Method of Favored States”, in E. Weher and P. Keddy (eds) Ecological Assembly Rules: Perspectives, Advances, Retreats. Cambridge: Cambridge University Press, pp. 5874.
Sklar, L. (1999) “The Reduction(?) of Thermodynamics to Statistical Mechanics”. Philosophical Studies 95: 187202.
Stamos, D.N. (2003) The Species Problem: Biological Species, Ontology, and the Metaphysics of Biology. Lanham, MD: Lexington Books.
Sterelny, K. (1996) “Return of the Group”. Philosophy of Science 63: 56284.
Sterner, R. W., and Elser, J. J. (2002). Ecological Stoichiometry: The Biology of
Elements from Molecules to the Biosphere. Princeton University Press, Princeton.
Swenson, R. (1989) “Emergent Attractors and the Law of Maximum Entropy Production: Foundations to a Theory of General Evolution”. Systems Research 6: 18797.
Tansley, A.G. (1935) “The Use and Abuse of Vegetational Concepts and Terms”. Ecology 16: 284307.
Ulanowicz, R.E. (1986). Ecology, The Ascendent Perspective. New York: Columbia University Press.
Vitousek, P.M., Mooney, H.A., Lubchenco, J., and Melillo, J.M. (1997) “Human Domination of Earth’s Ecosystems”. Science 277: 49499.
Warren, K. and Cheney, J. (1993) “Ecosystem Ecology and Metaphysical Ecology: A Case Study”. Environmental Ethics 15: 99116.
Watts, D.J. (1999) Small Worlds: The Dynamics of Networks Between Order and Randomness. Princeton: Princeton University Press.
Weber, B.H, and Depew, D.J. (1996) “Natural Selection and SelfOrganization”. Biology and Philosophy 11: 3365.
Weigert, R.G. (1988). “The Past, Present, and Future of Ecological Energetics”, in L.R. Pomeroy and J.J. Alberts, (eds) Concepts of Ecosystem Ecology: A Comparative View. New York: SpringerVerlag, pp. 2955.
Westra, L. (1994) An Environmental Proposal for Ethics: The Principle of Integrity. Lanham, MD: Rowman & Littlefield.
Wicken, J.S. (1987) Evolution, Thermodynamics and Information: Extending the Darwinian Program. Oxford: Oxford University Press.
Wilson, D.S. (1992) “Complex Interactions in Metacommunities, with Implications for Biodiversity and Higher Levels of Selection”. Ecology 73: 19842000.
Wimsatt, W.C. (1986) “Developmental Constraints, Generative Entrenchment and the InnateAcquired Distinction”, in Bechtel, W. (ed.) Integrating Scientific Disciplines. Dordrecht, the Netherlands: Martinus Nijhoff, , pp. 185 208.


1. Note that Sagoff inherits the “topdown”, “bottomup” distinction from Shrader Frechette and McCoy (1993), who inherit it from Kitcher (1985, 1989) and Salmon (1989), who used it to characterize the distinction between “epistemic” or “unificatory” and “ontic” or “causal” approaches to scientific explanation, respectively. But in this original context the terms are used to distinguish positions on a specific philosophical issue concerning the nature of scientific explanation, not the broad packages of methodological concepts and modeling strategies that Sagoff associates with the distinction.

2. Whether Clements actually held this extreme conception of community organization is itself open to debate. Quotes such as the following suggest that ecologists may be guilty of exaggerating the differences between Clements and Gleason: “Even where the final community seems most homogeneous and its factors uniform, quantitative study by quadrat and instrument reveals a swing of population and a variation in the controlling
factors” (Clements 1928: 3).

3. See Leibold and Mikkelson (2002) for a discussion of the spectrum of distribution patterns one might find in ecological systems, and the challenges of discerning which patterns are best supported by data.

4. Historian Joel Hagen writes: “Bormann and Likens were imbued with Eugene Odum’s idea of a “new ecology” in which the ecosystem was the central focus of ecological study. They adopted Odum’s holistic perspective: the parts of the system could only be understood within the context of the entire system” (Hagen 1992: 182).

5. The key word here is “straightforward”. Of course there will be individual ecologists and research groups for whom ontological commitment and methodology are strongly interrelated. Our claim is that across such groups one will find sufficient diversity to frustrate any simple mapping between these categories.

6. The species debate has been prominent in the philosophy of biology literature. Thus it is startling to read that, according to Sagoff, “problems of [species] classification ... were resolved in zoology by the time of Aristotle” (2003: 542).

7. We can elaborate on this point a bit. The zeroth law of thermodynamics states that two systems in thermal equilibrium with a third are in thermal equilibrium with each other, where “thermal equilibrium” involves lack of change in material and chemical properties when brought into contact. Thermally equivalent states are said to have the same “empirical temperature”, but temperatures sodefined will differ for different substances.
The second law of thermodynamics allows us to define an absolute temperature scale applicable to all substances. One version of the second law is the KelvinPlanck formulation: it is impossible to construct a heat engine that produces no other effect than the extraction of heat (Q) from a single source and the production of an equivalent amount of work. From this statement it can be shown that any two reversible engines operating between the same thermal reservoirs have the same efficiency (equal to 1 (Q1/Q2)), and that this efficiency is a function of the empirical temperatures of the
reservoirs, yielding the relation Q1/Q2 = f(T1)/f(T2). Kelvin proposed that the temperature function simply be T1/T2 and the relation Q1/Q2 = T1/T2 be used to define an absolute temperature scale applicable to all substances (see Holman 1988: 187190).

8. Correlation functions describe how fluctuations of given variable at a certain place and time influence the values of that variable at different places and times. Shorter correlation lengths mean that such influences fall off quickly, longer correlation lengths mean that fluctuations propagate over greater distances and persist over greater lengths of time. In statistical and solid state systems the variables in question include density (as in the water gas transition) and net magnetization (as in the ferromagnet paramagnet transition). As temperature is increased toward the critical point the correlation functions for these variables increase and shoot off to infinity at the critical temperature, leading to a globally “connected” system in which a small fluctuation precipitates wholesale macroscopic reorganization of the system.