Studies in History and Philosophy of Biological and Biomedical Sciences 45 (2014) 101–104
Contents lists available at ScienceDirect
Studies in History and Philosophy of Biological and Biomedical Sciences journal homepage: www.elsevier.com/locate/shpsc
Biodiversity, conservation biology, and rational choice David Frank Environmental Studies, Center for Bioethics, New York University, 285 Mercer St. #908, New York, NY 10003, United States
a r t i c l e
i n f o
Article history: Available online 9 November 2013 Keywords: Biodiversity Conservation biology Decision theory Environmental philosophy
a b s t r a c t This paper critically discusses two areas of Sahotra Sarkar’s recent work in environmental philosophy: biodiversity and conservation biology and roles for decision theory in incorporating values explicitly in the environmental policy process. I argue that Sarkar’s emphasis on the practices of conservation biologists, and especially the role of social and cultural values in the choice of biodiversity constituents, restricts his conception of biodiversity to particular practical conservation contexts. I argue that life scientists have many reasons to measure many types of diversity, and that biodiversity metrics could be value-free. I argue that Sarkar’s emphasis on the limitations of normative decision theory is in tension with his statement that decision theory can ‘‘put science and ethics together.’’ I also challenge his claim that multi-criteria decision tools lacking axiomatic foundations in preference and utility theory are ‘‘without a rational basis,’’ by presenting a case of a simple ‘‘outranking’’ multi-criteria decision rule that can violate a basic normative requirement of preferences (transitivity) and ask whether there may nevertheless be contexts in which such a procedure might assist decision makers. Ó 2013 Elsevier Ltd. All rights reserved.
When citing this paper, please use the full journal title Studies in History and Philosophy of Biological and Biomedical Sciences
1. Introduction Environmental Philosophy From Theory to Practice is a welcome, compact statement of Sahotra Sarkar’s unique approach to environmental philosophy. This approach is informed as much by philosophy of biology and ecology, economics, and conservation biology as it is by environmental ethics and politics. Sarkar covers much ground for such a slim volume, discussing, among other topics, debates in environmental value theory (ch. 3), biodiversity and conservation biology (ch. 5), the integration of values in policy-making using decision protocols (ch. 4), environmental restoration (ch. 6), sustainability (ch. 7), climate change (ch. 4, 8), and environmental justice (ch. 8). My constructive comments below will focus primarily on two areas that the book discusses in some detail: biodiversity and conservation biology and roles for decision theory in incorporating values explicitly in the environmental policy process. However it is worth stating one general criticism of the book at the outset: Sarkar attempts to cover too many fascinating topics in too few words. A paragon of the virtue of concision, the book’s brevity can also be frustrating. (For example, the papers in Gardiner, Caney, Jamieson, and Shue
E-mail address: [email protected]
1369-8486/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.shpsc.2013.10.005
(2010) could usefully supplement Sarkar’s limited discussion of climate change ethics.) My criticisms proceed as follows. In Section 2, I characterize the value-laden constituents-based approach to biodiversity taken in Chapter 5. Sarkar’s emphasis on the practices of conservation biologists, and especially the role of social and cultural values in the choice of biodiversity constituents, restricts his conception of biodiversity to particular practical conservation contexts. Whether or not Sarkar’s skepticism about the explanatory power of species diversity in ecology is justiﬁed, I argue that life scientists more generally have many reasons to measure many types of diversity. There are also other practical contexts where metrics of biodiversity might be important, and these are ruled out by Sarkar’s adequacy conditions. I also suggest that biodiversity metrics could be value-free, in a limited sense analogous to Boorse’s (1977) biostatistical theory of health: once biodiversity is speciﬁed, operationalized, and estimated in a geographical area, it remains an open question whether and how that biodiversity should be conserved or managed. In Section 3, I argue that Sarkar’s emphasis on the limitations of normative decision theory, especially their paradoxes and the
D. Frank / Studies in History and Philosophy of Biological and Biomedical Sciences 45 (2014) 101–104
idealizations built into the standard, static notion of individual preference, is in tension with his statement that decision theory can ‘‘put science and ethics together.’’ (80) I also question his claim that multi-criteria decision tools lacking axiomatic foundations in preference and utility theory are ‘‘without a rational basis.’’ (91) I present the case of a simple ‘‘outranking’’ multi-criteria decision rule (the so-called Regime method; Hinloopen, Nijkamp, & Rietveld, 1983) that can violate a basic normative requirement of preferences (namely transitivity) and ask whether there may nevertheless be contexts in which such a procedure might assist decision makers. 2. What is biodiversity? Life sciences and human values Sarkar’s work on biodiversity (2002, 2005, 2007, 2008, 2012) has defended the view that ‘biodiversity’ should be deﬁned by the practices of conservation biologists. As he puts it most recently (2012, p. 99): ‘‘biodiversity is simply the goal pursued by the practice of conservation biology.’’ Sarkar has also emphasized the normative problems inherent to deﬁning the concept, noting that since it is impossible to conserve all of life’s diversity, any practical attempt at biodiversity conservation must operationalize ‘biodiversity’ in terms of particular populations, species, higher taxa, biological ‘‘events’’ (e.g. migrations), etc. that society deems worth conserving. These are ‘‘constituents’’ of biodiversity. Furthermore, considerations of economy are always on the table, for conservation biologists and planners as much as for society more broadly. This choice of biodiversity constituents and the associated tradeoffs clearly involve value judgments. A philosopher might initially wonder whether this approach to deﬁning ‘biodiversity’ makes success inevitable.1 That is, if biodiversity just is whatever goal is pursued by conservation biology, then conservation biologists cannot be wrong about its deﬁnition. Analogously, if we take ‘health’ to be deﬁned as the goal pursued by doctors, then doctors, as long as they consistently pursue this goal, cannot fail to pursue health. Sarkar’s likely response would be to note that though there are conventional elements in the deﬁnition of ‘biodiversity,’ our practices and the biological world impose general adequacy conditions on any acceptable deﬁnition.2 Sarkar offers the following adequacy conditions for sets of biodiversity constituents (116): (1) ‘‘They must be biotic features; (2) Variability must be captured . . . we are dealing with biodiversity; (3) Taxonomic spread is important; (4) Concern should not be limited to material resource use.’’ Beyond the many values that may be involved in an initial choice of biotic features for conservation, these constraints explicitly incorporate values, particularly constraint (4), which, as Sarkar notes, is consistent with both anthropocentric and nonanthropocentric theories of the value of biodiversity. However this constraint might rule out a metric of, say, genetic or phenotypic diversity in an agroecosystem from being a metric of biodiversity per se, as long as the goal is to maximize resilience or yield of economically valuable food crops, to the exclusion of goals related to non-resource value. Sarkar explicitly restricts biodiversity to practical contexts where at least one goal is conservation for non-resource value. It is unclear whether constraint (4) rules out conservation for option value, where option value might in some 1
cases be limited to option value for future material resource use. Presumably this would depend on the intentions of the conservation biologists and ecosystem managers. An obvious consequence of this view is that the same set of constituents might count as biodiversity constituents in one context and not in another. Not only does Sarkar rule out practical contexts where biodiversity is being measured for reasons that exclude conservation for non-resource value, his value-laden approach does not explicitly embrace the possibility of value-free theoretical concepts and operationalizations of biodiversity. Consider a value-free speciﬁcation and operationalization of biodiversity as vertebrate species richness. Given some taxonomic theory of vertebrate species, we could estimate vertebrate species richness in different areas. Here, one might think, we are measuring a kind of biodiversity. However, it certainly remains an open question whether vertebrate species richness ought to be conserved, or maximized, etc. For example, we might be measuring vertebrate species richness with an eye toward questions in biogeography or ecology, or even evolutionary biology. Constraint (3) also places a value on taxonomic spread. In a footnote Sarkar admits that constraint (3) might be entailed by constraint (2), but variability can be cashed out in a number of ways, including the genetic or phenotypic variability within, say, a single higher taxon, for example as limited to vertebrates, or limited to a single species. Biologists have also devised many metrics of phylogenetic diversity. Indeed, life scientists have a general interest in diversity, as Sarkar’s discussion of diversity concepts and metrics in ecology as well as any perusal of life science oriented (as opposed to ethical) discussions of biodiversity, will attest (see, e.g., Gaston, 1996; Magurran & McGill, 2011). One might go so far as to argue that a central problem of the life sciences is the explanation of the diversity of living systems. Whether or not Sarkar’s skepticism of the diversity-stability relationship in ecology is justiﬁed, a philosophical account of biodiversity should not exclude the practices of many life scientists. This is an important difference between Sarkar’s approach and Maclaurin and Sterelny’s (2008) ‘‘units-and-differences’’ approach to biodiversity. Maclaurin and Sterelny argue that biodiversity concepts (especially species richness supplemented by various accounts of disparity—morphological, phylogenetic, ecological) play important roles in specifying causes and consequences in biological explanations. Based on their reading of previous work (especially his 2002 and 2005), Maclaurin and Sterelny accuse Sarkar of holding the view that biodiversity is ‘‘whatever we think is valuable about a biological system.’’ (8) While the adequacy conditions enumerated above go some way toward answering this charge by restricting sets of constituents, I have argued here that they are actually too restrictive. Dropping constraints (3) and (4) would accommodate the practices of both conservation biologists and other pure and applied life scientists interested in various kinds of biological heterogeneity. There is a deeper philosophical problem looming, which I will motivate without resolving. This problem would apply particularly to my suggested broad conception of biodiversity contained only in Sarkar’s constraints (1) and (2), as well as Maclaurin and Sterelny’s pluralistic units-and-differences approach. The problem is whether there is any such thing as biodiversity in general such that bringing together these concepts of biological heterogeneity is theoretically useful or interesting. DeLong’s (1996) ‘‘consensus’’ deﬁnition of biodiversity, based on consulting more than 80 published deﬁnitions, deﬁned biodiversity extremely broadly:
I owe this point to discussion with Laura Franklin-Hall. Odenbaugh mentions in his comments that Sarkar’s conventionalist approach to biodiversity cannot rely on Lewis’s (1969) account of convention. He is right, since Lewis only considers games where individuals are indifferent between coordinative equilibria. But a generalized version of Lewis’s game-theoretic approach where individual utilities for different equilibria differ, perhaps contextually, could work. 2
D. Frank / Studies in History and Philosophy of Biological and Biomedical Sciences 45 (2014) 101–104
Biodiversity is a state or attribute of a site or area and speciﬁcally refers to the variety within and among living organisms, assemblages of living organisms, biotic communities, and biotic processes, whether naturally occurring or modiﬁed by humans. Biodiversity can be measured in terms of genetic diversity and the identity and number of different types of species, assemblages of species, biotic communities, and biotic processes, and the amount (e.g., abundance, biomass, cover, rate) and structure of each. It can be observed and measured at any spatial scale ranging from microsites and habitat patches to the entire biosphere. (1996, p. 745) As Sarkar points out, such deﬁnitions are useless in conservation contexts. But they are also useless in theoretical contexts. What do measures of genetic heterogeneity in a population of humans have to do with measures of species richness across habitat patches? Very little of theoretical interest, I imagine. 3. Decision theory and environmental philosophy: putting science and ethics together? Sarkar is fond of optimization problems: he devotes much space to decision problems in conservation planning, and believes that decision theory can ‘‘put science and ethics together.’’ (80) (His is also probably the only text in environmental philosophy with a compact and clear description of Arrow’s theorem.) What does decision theory have to do with environmental philosophy? Firstly, decision theory is the foundation of contemporary economics, and economic approaches to environmental valuation have become ubiquitous despite critiques from philosophers, biologists, and economists themselves. Secondly, despite the fact that many economic approaches attempt to estimate the monetary value of environmental goods and services, decision theory is more general in that it can accommodate the incorporation of multiple values, including natural values, where the use of money as a numeraire is strictly optional. A good example of this kind of use of decision theory is found in Sarkar’s description of multi-criteria decision analysis for conservation prioritization in Namibia (box 4.4, 88– 90). Thirdly, the use of decision theory makes values explicit, which can be helpful for environmental decisions in democratic societies. Sarkar stresses that the use of decision protocols incorporating tools from decision theory can make environmental decisions transparent to decision-makers and stakeholders alike. A foundational notion of decision theory is the notion of preference: agents are presumed to have at least a weak ordering of alternatives or outcomes, perhaps one for each criterion of evaluation in the multi-criteria case, perhaps extending to a cardinal or interval ranking of ‘‘overall desirability’’ (the utility) of the alternatives or outcomes. These judgments of value are combined with probabilities that various states of the world will obtain to compute expected utilities of courses of action.3 Sarkar’s focus is on normative as opposed to descriptive uses of decision theory. For such normative uses we are supposed to take the axioms of, say, von Neumann and Morgenstern’s utility theory, as constraints of rationality. This is the approach taken by the ﬁeld of applied decision theory known as decision analysis (Edwards & von Winterfeldt, 1986). What is the philosophical basis for this approach, other than the intuitive plausibility of the axioms? So-called ‘‘Dutch Book’’-style arguments are supposed to show that, if an agent’s preferences or judgments of probability do not accord with the axioms of utility and probability theory, the agent can be offered a bet or series of bets such that he or she is guaranteed to lose value, whatever the agent cares about (Vineberg, 2011). 3
However, Sarkar also stresses the limitations and paradoxes of normative decision theory. He asks, ‘‘is it even clear that we should. . .try to maximize our [expected] utility?’’ (85) Here Sarkar asks the reader to consider the choice between an alternative with a low probability of an extremely high reward and an alternative with a high probability of a much lower reward, where the former has a higher expected utility but is intuitively less attractive. (It is worth noting that this simple example assumes a risk-neutral utility function.) Elsewhere Sarkar enumerates problems arising in applications of decision theory to group decisions, including issues with majority rule voting, the interpersonal comparability of preferences, as well as Arrow’s theorem. Sarkar also points out that time constraints unique to some environmental decisions introduce further dilemmas about the value of gathering decision-relevant information. The problems go deeper, as Sarkar argues earlier in the book and elsewhere (e.g. in his 2005) that human preferences (as revealed by demand) do not necessarily take into account the transformative value of the environment. That is, some aspects of the environment are valuable just because they are able to alter or even reverse our preferences. Decision theory often abstracts away from processes of preference formation and preference change: preferences are static rankings that represent an agent’s values at a time. While a dynamic view of preferences is of course possible (Grüne-Yanoff & Hansson, 2009), Sarkar does not provide a normative account of rational preference change that takes into account transformative value. These critiques of normative decision theory, and especially Sarkar’s endorsement of transformative value, seem to be in tension with Sarkar’s claim that decision theory can ‘‘put science and ethics together.’’ How could the transformative value of an area’s biodiversity be taken into account by decision-makers who are committed to basing everything on agents’ preferences? Eliciting people’s current preferences, one may be led to the conclusion that the area is worth more dead than alive. If transformative value counts ethically, then standard decision theory (i.e. decision theory that does not model preferences as dynamic) cannot put science and ethics together. Furthermore, Sarkar claims, provocatively, that methods for multi-criteria analysis that have ‘‘no foundation in accounts of agents’ preferences . . . therefore . . . have no rational basis.’’ (91) By foundation, I take it that Sarkar means an axiomatic foundation in preference or utility theory. Some multi-criteria decision methods, such as multi-attribute utility theory, have explicit foundations in utility theory, while others do not. Putting the issue of transformative value to the side, consider the multi-criteria decision problem in the table below, where Areas A, B, and C are potential conservation areas, and the rankings represent ordinal rankings of species richness, community uniqueness, and taxonomic spread. For each criterion, 1 is the best, 2 the second best, 3 the worst. Let’s assume that we can only come up with ordinal rankings for these criteria.
Area A Area B Area C
1 2 3
2 3 1
3 1 2
In this decision problem, each area is non-dominated, so simple dominance analysis cannot be used to rule out any of the
This presentation idealizes a bit. Utility functions are often elicited via pairwise comparisons (ordinal rankings) of probabilistic gambles.
D. Frank / Studies in History and Philosophy of Biological and Biomedical Sciences 45 (2014) 101–104
alternatives.4 One way to further enrich the basis for making decisions in these cases is to ordinally rank the criteria themselves. Thus a complete, transitive, weak ordering is induced on the criteria. This allows for use of the ‘‘Regime method’’ (Hinloopen et al., 1983), one of many proposed rules for multi-criteria decision-making that is not based on utility theory (Moffett & Sarkar, 2006). The Regime method works as follows. For any two alternatives a1 and a2, let K+ be the set of criteria where a1 > a2 (a1 is strictly preferred to a2 on those criteria), and K the set of criteria where a2 > a1. Say that a1 outranks a2 if and only if K+ is non-empty and there exists an injective (one-to-one) function where each criterion in K is mapped to a more important (higher ranked) criterion in K+. This method yields as a solution the set of alternatives that are not outranked. It was proposed as an ‘‘extension’’ of the notion of dominance. Assume that for the decision problem in Table 1, Species richness is more important to the decision-maker than Taxonomic spread, which is more important than Community uniqueness. Comparing Area A and Area B with the Regime method results in Area A outranking Area B, since Taxonomic spread is the only member of the set K , which can be mapped to Species richness in the set K+. Comparing Area B and Area C results in Area B outranking Area C, since Community uniqueness is the only member of the set K , which can be mapped to Taxonomic spread or Species richness in the set K+. Comparing Area A and Area C results in neither outranking the other. Thus the only remaining alternative is Area A, which is outranked by neither Area B nor Area C. This example shows that Regime’s outranking relation is not transitive in all cases, since in this case Area A outranks Area B, and Area B outranks Area C, but Area A does not outrank Area C. The intransitivity of Regime’s outranking relation is problematic if one takes transitivity to be a constraint on any rational ranking procedure for a single decision-maker in multi-criteria decision problems, as classical decision theory assumes for the preference relation. The usual argument for this is that intransitive preferences can lead one to take a series of trades such that one is guaranteed to lose value. Say I prefer A to B, and B to C, but I am indifferent between A and C. If I start with B, someone could sell me A, then trade me for C, since I am indifferent between A and C. But then someone could sell me B, since I prefer B to C. If I start with A, someone could trade me for C, then sell me B and then A again. If I start with C, someone could sell me B, and then A, but then trade me for C. In all of these scenarios I lose value and end up with the alternative I started with. But it is unclear what we should make of this result. Some philosophers have even rejected the idea that transitivity of preference is a constraint of rationality (Anand, 1993). We could reject the rule out of hand, since it can sometimes produce intransitive outranking relations. Sarkar seems to commit himself to the view that the Regime method and all other multi-criteria decision making methods that are inconsistent with preference or utility theory should be rejected. But we could also take a more nuanced view. While I will not develop such a view here, it would reject the idea that preference and utility theory are global constraints on practical rationality—rather they are useful tools in some contexts and not others. That is, in some contexts, the Regime method might be helpful. Given Sarkar’s generally pragmatic outlook, I assume he would ﬁnd this weak claim at least somewhat attractive. 4. Conclusion
measures of biodiversity that might prove useful in theoretical contexts as well as many important practical contexts. In Section 3 I argued that Sarkar’s enthusiasm for decision protocols is undermined by the critiques of decision theory he himself discusses, especially his views on transformative value. I also suggested that he commits himself to a strong view about the normative status of preference and utility theory. Overall, Sarkar’s Environmental Philosophy presents a compelling case for what he calls ‘‘integrative biocultural environmentalism,’’ the view that ‘‘natural values cannot be successfully pursued without simultaneously pursuing human cultural values and vice versa.’’ (176) If natural values like biodiversity are pursued without human values, conservationists become imperialists, forcing their idiosyncratic vision of the world on the rest of us. That is, successful conservation biologists must negotiate the value of biodiversity with other human values such as public health, the economic use of landscapes, and cultural values placed on particular species. If human values are pursued without natural values, then humans are impoverished, culturally and scientiﬁcally as well as economically and ecologically. Acknowledgements This paper was originally presented at the American Philosophical Association-Paciﬁc Division 2013. For improvements to this paper, acknowledgements are due to the session attendees, the other presenters, and an anonymous referee. References Anand, P. (1993). The philosophy of intransitive preferences. The Economic Journal, 103, 337–346. Boorse, C. (1977). Health as a theoretical concept. Philosophy of Science, 44(4), 542–573. DeLong, D. C. (1996). Deﬁning biodiversity. Wildlife Society Bulletin, 24, 738–749. Edwards, W., & von Winterfeldt, D. (1986). Decision analysis and behavioral research. New York, NY: Cambridge University Press. Gardiner, S. M., Caney, S., Jamieson, D., & Shue, H. (Eds.). (2010). Climate ethics: Essential readings. Oxford: New York. Gaston, K. (Ed.). (1996). Biodiversity: A biology of numbers and difference. Oxford: Blackwell. Grüne-Yanoff, T., & Hansson, S. O. (Eds.). (2009). Preference change: Approaches from philosophy, economics and psychology. New York: Springer. Hinloopen, E., Nijkamp, P., & Rietveld, P. (1983). The regime method: A new multicriteria method. In P. Hansen (Ed.), Essays and surveys on multiple criteria decision making (pp. 146–155). Berlin: Springer. Lewis, D. (1969). Convention. Cambridge, MA: Harvard University Press. Maclaurin, J., & Sterelny, K. (2008). What is biodiversity? Chicago, IL: University of Chicago Press. Magurran, A. E., & McGill, B. J. (Eds.). (2011). Biological diversity: Frontiers in measurement and assessment. Oxford: New York, NY. Moffett, A., & Sarkar, S. (2006). Incorporating multiple criteria into the design of conservation area networks: A minireview with recommendations. Diversity and Distributions, 12, 125–137. Sarkar, S. (2002). Deﬁning ‘biodiversity’; assessing biodiversity. The Monist, 85(1), 131–155. Sarkar, S. (2005). Biodiversity and environmental philosophy: An introduction. New York, NY: Cambridge University Press. Sarkar, S. (2007). From ecological diversity to biodiversity. In D. Hull & M. Ruse (Eds.), The Cambridge companion to the philosophy of biology (pp. 388–409). Cambridge: Cambridge University Press. Sarkar, S. (2008). Norms and the conservation of biodiversity. Resonance(July), 627–637. Sarkar, S. (2012). Environmental philosophy: From theory to practice. Malden, MA: Wiley-Blackwell. Vineberg, S. (2011). Dutch book arguments. In E. Zalta (Ed.), The Stanford encyclopedia of philosophy. URL = Accessed March 2013.
I argued in Section 2 that Sarkar’s account of biodiversity is too parochial to conservation biology: it rules out concepts and
An alternative dominates another just in case it is at least as good on all criteria and better on at least one criterion.