1. Natural capital, ecosystem services and no net loss

In environmental policy making natural environments, resources and beings have been increasingly conceptualized as ‘natural capital’. Central to this conceptualization of environmental goods as capital is the concept of ‘ecosystem services’. Natural environments, resources and beings should be understood as assets valued in terms of the services they provide to human beings, either directly for the welfare of those who use them and value their existence, or indirectly as inputs into production and as the background conditions for human life and production. A standard list of ecosystem services accompanies this claim: provisioning services, such as clean water and inputs into food production; regulating services, such as waste assimilation, carbon sequestration and pollination; cultural services, such as recreation or aesthetic enjoyment; supporting services, such as soil formation and photosynthesis (TEEB 2010: 7). On this account, then, environmental goods such as biodiversity are understood as capital that are conditions for the provision of ecosystem services.

This conceptualization of environmental goods in terms of capital and ecosystem services is the starting point in defence of a number of further independent claims that have underpinned environmental policy making. It informs standard approaches to economic valuation: the economic valuation of environmental goods is conceived in terms of ascertaining the monetary value of ecosystem services. It informs standard approaches to sustainability. Sustainability is a matter of maintaining those levels of capital required to provide these services. This understanding of environmental goods is largely shared by both sides of the debate between proponents of weak and strong sustainability in economics (O’Neill Reference O’Neill, Moellendorf and Widdows2015). Both assume that environmental goods should be understood as capital that provides services. The point at issue is whether and how far natural capital can be substituted by forms of human capital. Proponents of weak sustainability are taken to claim that a wide degree of substitutability of goods at the margin is possible and hence that all that matters for sustainability is that the total level of capital, human-made and natural, does not decline. Proponents of strong sustainability are typically taken to claim that there are limits to the substitutability of natural and human-made capital and hence that sustainability requires that there is some level of ‘natural capital’ that does not decline (Beckerman Reference Beckerman1994, Reference Beckerman and Dobson1999; Daly Reference Daly1995; Jacobs Reference Jacobs1995; Dobson Reference Dobson1996; Holland Reference Holland and Foster1997).

Proponents of strong sustainability, while insisting on the limits of substitutability between natural and human capital, can still be open to the possibility of considerable substitutability between different components of natural capital. The aim of policy should be to ensure there is ‘no net loss’ of natural capital such that the total aggregate level of natural capital is maintained. This is possible if the loss in one component of natural capital can be compensated with gains in another.

‘No net loss’ of natural capital has been increasingly appealed to in national and international environmental policy statements. It is associated with biodiversity offsetting markets that have become the standard market-based approach for the realization of the objective of no net loss. Landowners and environmental bodies who create, restore or enhance a habitat or site of biodiversity are assigned credits. Those credits can then be sold to developers to offset environmental damage to natural capital caused by a development. Market transactions can ensure that there is no net loss of habitats, biodiversity and other forms of natural capital. Early versions of biodiversity offsetting were developed in New South Wales, Australia and, in the form of wetland mitigation banking, in the USA. It is now becoming increasingly prevalent elsewhere (Kiesecker et al. Reference Kiesecker, Copeland, Pocewicz, Nibbelink, McKenney, Dahlke, Holloran and Stroud2009; Madsen et al. Reference Madsen, Carroll, Kandy and Bennett2011; UNEP 2011). A policy to ‘ensure there is no net loss of ecosystems and their services (e.g. through compensation or offsetting schemes)’ (European Commission 2011) forms a significant component of the ‘EU Biodiversity Strategy to 2020’. In the UK no net loss forms part of the UK Biodiversity 2020 plan (Defra 2012), where the aim is to be similarly pursued by biodiversity offsetting (Defra 2013). The approach is itself part of the larger international initiative to develop ‘no net loss’ approaches to environmental protection (BBOP 2012). The policy is used increasingly in development projects, for example, in the planning for the High Speed Two train line in the UK (HS2 Limited 2015) and in the justification of mining projects by Rio Tinto Zinc in Madagascar (Seagle Reference Seagle2012; Kill and Franchi Reference Kill and Franchi2016).

In this paper I will argue that major goods that matter to people and their well-being will be lost through a policy of no net loss. I develop that argument through a criticism of the claim that environmental goods should be characterized in terms of natural capital and the ecosystem services they provide which underpins the no net loss approach to environmental policy. My central aim in the paper will be to show that the concepts of natural capital and ecosystem services cannot capture important dimensions of value that matter to human well-being. As such I will not consider all the dimensions of value that cannot be captured by the concepts of natural capital and ecosystem services. The concepts of natural capital and ecosystem services are concerned only with environmental goods insofar as they are good for human beings. The approach is welfarist: only human well-being matters in itself. It does not include those dimensions of value which are claimed to be independent of human well-being. There are two dimensions of value that might be appealed to in this regard. The first is the appeal to the moral status of non-human beings: the well-being of sentient beings, or more generally the good of living things, might be taken to matter in itself independently of the well-being of human beings. The second is the appeal to impersonal values. To claim that some things or states have impersonal value is to claim they have value that is independent of their contribution to the goodness of the life of some being (Brink Reference Brink1989: 219). For example, if biodiversity is intrinsically valuable and not simply a means to the realization of some other good, such as the well-being of particular beings affected by biodiversity loss, then it looks like it has impersonal value.Footnote ¹ Questions about the defensibility of claims about the moral standing of non-human beings or of the existence of impersonal values are important. However, the arguments in this paper will not address those questions. I will rather suggest that, even if we are only concerned with human well-being, there are significant dimensions of environmental value that are not captured by the concepts of natural capital and ecosystem services. Significant goods that matter for human well-being will be lost through a policy of no net loss.

2. Justifying no net loss

The policy of no net loss takes it to be both possible and desirable to maintain or even increase total levels of natural capital through substitution across different dimensions of natural capital. As Helm puts it ‘the aggregate of natural capital should be non-decreasing, but … there can be substitutability between different types of natural assets’ (Helm Reference Helm2014: 111). Helm contrasts this approach with, on the one hand, weaker versions of sustainability which focus ‘on the substitutability between natural and man-made factors of production, with sustainability being achieved if the investment in produced capital equals the reduction in natural capital’ (Helm Reference Helm2014:111) and, on the other hand, versions of deep green environmentalism according to which ‘all and every environmental asset should be preserved from development’ (Helm Reference Helm2014: 111). If the aim of policy is to preserve aggregate natural capital, the approach allows that losses in one natural asset can be compensated for and offset by improvements in another natural asset elsewhere:

Helm suggests that such compensation can take place through monetary payments in properly constituted markets: ‘It remains to determine how the compensation will be carried out. In principle this could be a monetary payment, equal to the value of the damage. A market could be constituted in which bids are made for natural capital improvements …’ (Helm Reference Helm2014: 118).

Two claims need to be distinguished here. The first is that it is possible to substitute one natural asset by another, so that losses in one area of natural capital can be compensated for by gains in another, and aggregate natural capital can be maintained. The second is that compensation can be achieved through monetary payments allowing the development of biodiversity offset markets. The second claim is stronger than the first. Most of this paper will be concerned to criticize the first basic claim. However, I will first make some critical observations about the second stronger claim.

Against the view that monetary compensation is not possible for damages and should rather take a non-monetary physical form, Helm offers the following response:

The claim that ‘unless an asset is of infinite value, there is always a price or a cost that can be paid in compensation for its loss’ assumes that all goods have either a finite or infinite price. There are good grounds for rejecting that disjunction. It is open to the response that some goods are priceless in the sense that either they do not or should not have any price. This is to make a particular incommensurability claim: X cannot be valued in monetary terms or is not appropriately valued in monetary terms. It is important to notice that this claim is logically distinct from the claim that a good X has an infinite value. Indeed, the claim (1) ‘it is not the case that X has or ought to have a price’ is logically inconsistent with the claim (2) ‘X has or ought to have a price of infinity’. The central arguments against the claim that environmental goods can be priced are incommensurability arguments against the possibility or desirability of pricing (O’Neill et al. Reference O’Neill, Holland and Light2008: Ch. 5; O’Neill Reference O’Neill and Spash2017a).Footnote ²

Why should some goods be considered incommensurable with monetary values? One argument for monetary incommensurability to which I will return later appeals to the claim that some social relations and ethical commitments are constituted by a refusal to place monetary values on them (Raz Reference Raz1986: 345ff; O’Neill Reference O’Neill1993: 118–122; O’Neill et al. Reference O’Neill, Holland and Light2008: 77–79). Consider the following passage from a person facing eviction to make way for the development of the Narmada dam:

The point of the rhetorical questions – ‘what price our adivasi life?’ – is not to suggest there is a monetary value that could be placed on the good – be this finite or infinite – but rather to reject the very possibility of properly expressing the value of kinship and a way of life in monetary values. The claim that monetary compensation for a loss is always possible is rejected. I return below to the defensibility and rationality of this refusal to accept monetary compensation for environmental damage.

The claim that specifically monetary substitution is always possible for environmental damage needs to be kept distinct from a second claim that Helm defends and which underpins the policy of no net loss – that it is possible to substitute one natural asset by another so that losses in one area of natural capital can be compensated for by gains in another such that aggregate natural capital can be maintained. The claim here is that environmental goods be understood as natural capital, assets that provide services, and that one asset can be substituted by another if it provides the same services. This claim need not be committed to monetary compensation. It need only assume that physical offsets are available which provide the same services.Footnote ³ My central aim in the rest of the paper is to reject this prior claim: major goods that matter to people and their well-being will be lost through a policy of no net loss.Footnote ⁴

3. Ecosystem services: de re and de dicto valuation

Environmental policy is increasingly framed in terms of environmental goods, such as biodiversity, being understood as natural capital, valued in terms of the ecosystem services they provide. What is it to value goods in terms of the services they provide? A useful starting point here is an influential characterization of the nature of economic goods in terms of the services they provide offered by Ayres and Kneese. The characterization is a general one that pre-dates the specific use of the concept of ecosystem services:

Important here is the claim that objects should be understood as ‘merely the vehicles’ which carry services. What are valued are services. Any object as such is valued simply as a ‘vehicle’ for these services. For what class of objects is this true? What is it to value objects in this way and when is such valuation justified?

A useful starting point for answering these questions is the distinction between de re and de dicto valuation. The distinction can be illustrated by the following mildly funny joke about Zsa Zsa Gabor offered by Hare:

The joke turns on an ambiguity between de re and de dicto valuations. When Zsa Zsa says she does good in finding a way to keep her husband young and healthy we expect her to be concerned about the youth and health of a particular person who is her husband. However, it transpires that Zsa Zsa is only concerned that whoever turns out to fall under the description ‘my husband’ be young and healthy. We expect her to value the person de re, to be valuing a particular object. It turns out that she is valuing de dicto: she values whoever happens to fall under the description of being her husband. Consider the statement ‘Zsa Zsa Gabor values a husband who is young and healthy’. The statement is ambiguous. Under the de re reading what is valued is a particular individual who is young and healthy. Under the de dicto reading what is valued is that she has a husband who falls under a description of being young and healthy.

The distinction allows us to clarify at least part of what is at stake in the use of the language of natural capital and ecosystem services. To value something simply as a vehicle for the provision of services is to assume that it should be valued de dicto and not de re. Consider the statement ‘P values a site of biodiversity which has qualities α₁ … α_n’. The statement displays an ambiguity. One reading is a de re reading. What is valued is a particular place that is a site of biodiversity with qualities α₁ … α_n. Another reading is a de dicto reading. What is valued is that a site falls under the appropriate description as a site of biodiversity with qualities α₁ … α_n. Policy that aims at no net loss and the related practices of biodiversity markets and offsets assume de dicto valuations. A site is valued since it falls under the appropriate description and its loss can be compensated for by its replacement by another site that falls under the same description. Those opposed to the destruction of sites of biodiversity for the purposes of some development to be offset are often moved by an attachment to particular places. They value sites of biodiversity de re. Conflicts emerge where the same specific site is valued de re by people with a particular attachment to a site of biodiversity, but de dicto by a policy maker committed to no net loss. Consider a particular example – Smithy Wood, an 800 year old woodland of 20 acres near Sheffield in the UK. The developers who want to build a new motorway service area on the site of the Smithy Wood plan to offset the loss of woodland with the creation of a new community woodland of up to 600 acres. For locals with a particular attachment to the woodland, what matters is the value of the woodland as a particular place with its particular history. It is valued de re. The policy maker looking at the same woodland in the light of a no net loss policy will understand the value of the woodland differently. It is valued solely in terms of the biodiversity properties α₁ … α_n the wood exhibits. It is valued de dicto.

The competing valuations matter to how far substitutability between different environmental goods is possible. If sites are valued de dicto then one site can be substituted by another with respect to that valuation if and only if it falls under the same description. De dicto valuation does bring with it important restrictions on substitutability. The substitute must fall under the same description as the original. The point is at the centre of many disputes about biodiversity offsetting. Whether or not substitutability is possible will depend upon the description under which substitution takes place. Given a detailed and specific description of the qualities of the site, α₁ … α_n, substitutability will be very difficult or even impossible. On the other hand, given a very general and minimal description of those qualities, β₁ … β_m, substitution will not be difficult at all. One site will not be a substitute for another under one description, say as a woodland that contains a very specific mix of species of plants and trees, but may be under another, say as simply a mixed woodland. The detailed disputes about the adequacy of offsets typically concern the descriptions under which the offsets are made – with the ways in which the equivalence of two sites of biodiversity is characterized (Sullivan Reference Sullivan2013; Carver and Sullivan Reference Carver and Sullivan2017). De dicto valuation is consistent with considerable limits on the possibility of substitution. However, if sites of biodiversity are valued de re, then there exist still stronger limits on the possibility of substitution. The loss of one valued particular site or place cannot be simply substituted by another no matter whether there is a description characterizing their environmental properties which both meet. It is the particular site that is valued. An important background question then concerns what is the scope of objects to which de re and de dicto valuation is appropriate and why?

What objects are owed a de re valuation, what objects are owed de dicto valuation and why are these valuations owed to each? De re value appears to be what is often ethically demanded or appropriate with respect of persons. Zsa Zsa Gabor isn’t finding a ‘way of doing good for people’ by substituting a new person with the required husband-like qualities for an old one. Zsa Zsa husbands are not substitutes for real husbands. Similarly Stepford wives are no substitutes for real wives. Love is a de re attitude to a particular person. As Kraut puts it, ‘it is one thing to love a particular person and quite another to love some property or properties that s/he happens to possess’ (Kraut Reference Kraut1986: 421; cf. Goldie Reference Goldie2010; Grau, Reference Grau2010). To say this is not to deny that the basis for loving a person makes reference to his or her properties – his or her kindness, generosity, sense of humour, and so on. However, it is a mistake to confuse the basis of love and the object of love (Soble Reference Soble1990: 225ff.; cf. Velleman Reference Velleman1999: 368; Kolodny Reference Kolodny2003: 154; Grau Reference Goldie2010: 260). The object of love is the person and not the properties he or she happens to exhibit. If it was simply the properties that were the object of love, then the person could be substituted by another.

On the other hand, there is a large class of objects to which de dicto valuation is appropriate. Some things are valued in virtue of displaying some set of properties. We value whatever falls under a description. We typically value tools de dicto. We are concerned with the job that the tool does, not this or that particular tool. For the most part I am not attached to a particular tool. There are exceptions where the particular history matters. I have, for example, a hammer that was passed on to me from my father. As a tool it could be better. The head sometimes comes loose. However, the hammer is valuable to me de re. It is this particular hammer with its history that I value. However, these are exceptions. A similar story goes for everyday fungible goods like apples and pears. I value any apple that falls under the description of a fruit with particular gustatory properties. Again there are exceptions – for example, the first apple picked off a tree that I planted with my daughter, which I took several hundred miles to give to her. However, these are again exceptions. The range of goods for which de dicto valuation is owed is a wide one. That this is the case indeed matters for the possibility of sustainability. Consider sources of energy. Insofar as sources of energy are valued simply in terms of the needs they are able to satisfy, such as transport, warmth, sustenance and health, they will be typically valued de dicto. What is valued is that there is a source of energy that meets those needs, not any particular source (Brand-Correa and Steinberger Reference Brand-Correa and Steinberger2017). That this is the case is a precondition of sustainability since it allows sources of energy with high environmental impacts, for example, high greenhouse gas emissions such as coal and oil, to be substituted by sources that have lower emissions such as wind and solar power.

The defender of the view of environmental goods as components of natural capital that are to be valued for their ecosystem services assumes this attitude to be appropriate to environmental goods more generally. We are taken to value whatever provides the ecosystem services or functions that are relevant. The particular site or object is a vehicle for the provision of services. The loss of one vehicle can be substituted by another that provides the same services. Is this assumption defensible or are there good reasons to suppose that many environmental goods are goods to which de re valuation is appropriate?

An initial point to note is that the question of substitutability is Janus faced in discussions of sustainability (O’Neill Reference O’Neill, Moellendorf and Widdows2015: 404). On the one hand, sustainability requires the substitutability of environmentally damaging technologies, goods and resources by others that are less environmentally damaging. On the other, sustainability is taken to require constraints on the substitutability of particular places, habitats and beings by others. Correspondingly, for example, where a new source of energy production such as a dam causes the loss of a particular valued place or habitat, it can result in the loss of a good for which there is no substitute (O’Neill Reference O’Neill2017b). Why are there such limits on substitutability? In the following I argue that one source of non-substitutability is the existence of goods for which de re rather than de dicto valuation is appropriate.

4. Constituents and determinants of well-being

Why is de re valuation often appropriate? Consider the case of human persons first. Why are de re valuations required when we are concerned with persons? To answer that question, start with a prior question. What would be wrong with valuing a person simply for the services they provide? One answer that might be offered is a Kantian answer that appeals to the moral attitudes required. We are morally required, from the Kantian perspective, to value persons not merely as means, but as ends in themselves. We should not value persons merely for the services they provide us, but as autonomous rational persons. The Kantian reasons for holding this view are in outline as follows: Every rational agent is necessarily committed to treating himself or herself as a rational agent who governs themselves by their own reason – i.e. treating themselves as ‘ends in themselves’. Since reasons are impersonal, every rational agent recognizes that all other rational agents are also committed to treating themselves as ends in themselves. Hence, all rational agents are committed to treating other rational agents as ends in themselves (Kant Reference Kant1948: 91 [429]). Whether this argument is a good one or not I will not consider here. The argument, if successful, would show that we should not value persons merely as a means. However does it offer grounds for de re valuations of persons?

The difficulty with the suggestion that the argument could provide a justification for de re valuation is that the grounds it offers for treating persons as ends in themselves are impartial. They are reasons that require a particular attitude of moral respect for any rational person whoever he or she is. They do not give reasons for valuing a particular person as that particular person. Such impartial moral concerns look like the wrong place to ground valuations de re. They don’t offer a strong ground for de re valuations as such – although moral respect will always be owed in the end to particular persons. The point is one pressed by Williams against both Kantian and utilitarian ethical positions (Williams Reference Williams1976). A husband who saved his wife over a stranger on the grounds that ‘moral principle can legitimate his preference, yielding the conclusion that in situations of this kind it is at least all right (morally permissible) to save one’s wife’ would have provided ‘one thought too many’ (Williams Reference Williams1976: 18). The motivation to act should be simply that this person was his wife without the additional thought that there was a moral principle that rendered it permissible for a person to save whoever happened to be his wife. De re attitudes are not plausibly grounded in an impartial ethic (Smith Reference Smith1995: 71–76).Footnote ⁵

Similar points apply to considerations about valuing the non-human world. An objection to understanding the value of the natural world through the concept of ecosystem services is that it values environmental goods purely instrumentally for the services they provide to human beings. What is the source of the objection? One possible answer is one that follows a broadly Kantian route. One implication of Kant’s account of the attitudes owed to persons in saying that they are ends in themselves and not to be treated merely as a means is that they have a moral standing as such. They are said to have ‘intrinsic value’ in the sense of being morally considerable.Footnote ⁶ A central move in much environmental ethics has been to extend moral standing beyond the Kantian domain of persons to include non-human beings and states – be this all sentient beings, all living things or even collectives such as ecosystems. Non-humans are said to have intrinsic value in this moral sense.Footnote ⁷ If this is the case, then they cannot be treated merely as a means – so that to treat them as simply providers of ecosystems services is to fail to recognize their value. Whether some non-humans can be said to have intrinsic value in this moral sense, and if they can, how far intrinsic value extends, raises large questions which I will not pursue here. However, again it does not give strong grounds for de re valuations as such being important. The grounds for appropriate ethical attitudes are impartial. For example, if sentient beings have moral standing they have it in virtue of falling under the description of being sentient beings, not in virtue of being this or that individual sentient being. A de re valuation is not plausibly grounded in an impartial ethic of this kind.

What then might be the grounds for de re valuation? A distinct line of argument for de re attitudes being appropriate is an Aristotelian one. Consider again relationships to particular persons. The argument would run something as follows. Relationships to particular persons are a central component of what it is to live a flourishing human life. This claim might appear to take us back into the realm of treating persons as a means to an end, in this case that of leading a flourishing life. However, an important distinction in subsequent interpretations of Aristotle and in the Aristotelian tradition is a distinction between two ways in which something might be said to promote an end. The first is that something is an external causal means to an end. The second is that something is constitutive of that end (Wiggins Reference Wiggins and Rorty1980).Footnote ⁸ Take a well-worked example, the claim that friendship is necessary for a good life. There are two ways in which that might be understood. The first is that friendship is a causally necessary means to a good life, for example that friends will bring you support in hard times and the pleasures of conversation in all times. More recently, social relationships like friendship have been described as forms of ‘social capital’ with the benefits these bring – better employment possibilities, higher incomes, better physical and mental health and so on. All these facts about the causal impact of friendships on life might be true. However, if this was all that there was to the contribution of friendship to a good life it might look as if something was awry. It makes friendship look like an insurance contract with additional benefits. We don’t merely value relationships of friendship as an external causal means to other goods. We value our friends and our relationships with our friends as ends in themselves. Friendships are part of what makes a good life. They are one of its constituents. They are constitutive of the good life. One might say then that friendship is not just a means to a good life in the sense of being a causal determinant, but a ‘means’ in the sense that it is a central constituent of the good life.Footnote ⁹

Similar points can be made with respect to environmental goods (O’Neill Reference O’Neill1993: 23–24; cf. James Reference James2016). Relationships to particular persons and places are central components of human well-being. Relationships to particular places and particular non-human beings and environments are constitutive of the good life. These relationships involve de re attitudes and values. While they are conditions for human well-being, they are not conditions in the sense of being external causal means for independently specified ends. Rather they are constitutive of well-being. They are valued not simply for services they provide. Rather, they are valued for their own sake. As such they do not have substitutes in other objects with the same causal properties. The loss of particular places matters as such.

It is attachments to such particulars that moves opposition to offers of compensation, be this monetary or in the form of a physical replacement. Consider again the passage from a person facing eviction to make way for the development of the Narmada dam quoted earlier. The response does make reference to services provided by the river: the provision of water and fish, and the watering of soil for growing vegetables. However, what matters here is not just the provision of services, but the loss of a particular place. It matters in virtue of the past of a community that is embodied in it and in the way that it is constitutive of particular social relations in the present. What is at stake in the flooding of the place is the disintegration of a community and way of life. The basic dimensions of human well-being – human affiliation and community – are rooted in a particular place.

Similar points apply to the ordinary and everyday places people inhabit. They are relationships to particular environments that embody their histories, personal and collective. Habitats may have species assemblages that might be easy to replace – a pond or copse of woodland – and hence are replaceable with ‘no net loss’ of biodiversity. In that sense they might have replacements, although this requires qualification since there is an historical dimension to biodiversity.Footnote ¹⁰ However, they also have significance as particulars, relationships to which are constitutive of a life. It is this particular pond where I as a child caught fish with my father or this particular copse I walked in with friends and family that matters. Where developments threaten a particular place it is these kinds of attachments to particulars that often motivate resistance.

Many environmental goods that are valued de re matter centrally to the well-being of individuals. Hence the reason why environmental concern with local places is central to so much environmental conflict. Attachment to particulars matters to people. They matter not just as resources that are an external causal condition for the realization of some distinct good, but as a constitutive condition of the lives of persons and communities. As such, as Goodin notes of natural environments, they provide a context in which people can make ‘some sense and patterns to their lives’ (Goodin Reference Goodin1992: 37). Just as it is an error to treat social relationships like friendship merely as forms of ‘social capital’ valued simply for the independent benefits these bring in the form of employment, incomes and health, so it is an error to treat relationships to particular places and environments simply as ‘natural capital’ that deliver benefits in the form of services. The facts about the external benefits of relationships to people and places might be true – but what the metaphors of capital and services disguise are the ways in which these relationships are constitutive of a good life.

5. Rescuing ecosystem services from economics?

The recent conceptual framework developed by the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES) has included concern for what are characterized as ‘relational values’ as opposed to both ‘instrumental’ and ‘intrinsic’ values (Diaz et al. Reference Diaz2015; cf. Chan et al. Reference Chan, Balvanera, Benessaiah, Chapman, Díaz, Gómez-Baggethun, Gould, Hannahs, Jax, Klain, Luck, Martín-López, Muraca, Norton, Ott, Pascual, Satterfield, Tadaki, Taggart and Turner2016). Intrinsic values are defined as ‘those inherent to nature, independent of human judgement’ and are claimed to ‘have no relationship with possible benefits to humans or their quality of life’ (Diaz et al. Reference Diaz2015: 11). Instrumental values are defined in terms of ‘nature’s benefits as far as they allow people to achieve a good quality of life, be it through spiritual enlightenment, aesthetic pleasure or the production or consumption of a commodity’ (Diaz et al. Reference Diaz2015: 11). In contrast to both, relational values are characterized as those ‘imbedded in desirable (sought after) relationships, including those between people and nature’ (Diaz et al. Reference Diaz2015: 11). In another recent paper that defends this approach the contrast is introduced thus:

These relational values are taken in the IPBES conceptual framework to fall outside the scope of economic valuation. Space will not permit a critical examination of these conceptual contrasts and the way they are drawn here. However, the introduction of the concept of ‘relational values’ is consistent with the arguments developed in this paper. Indeed, one way of understanding the arguments here might be as a way of characterizing at least one class of environmental relational values and contrasting them both with instrumental values and with values that reflect the independent moral status of non-human beings. Relationships to particular other humans and non-humans are constitutive of a good life for human beings – rather than being external instrumental means. The de re valuations that underpin them are not reducible to impartial claims about the moral status of human and non-human beings. The arguments here are also clearly consistent with the recognition in the IPBES framework that these relational values cannot be captured by ‘an economic valuation framework’ (Diaz et al. Reference Diaz2015: 11).

There is another question however. Can they still be integrated into a framework of ecosystem services? After all, the framework offered is that of Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services. What response might be made to the arguments in this paper? One response that has been put to me by proponents of the IPBES approach is this – that my arguments assume a particular narrow economic understanding of the concept of services.Footnote ¹¹ They assume the framework outlined by Ayres and Kneese:

The response runs that we can extend the concept of services beyond this narrow economic reading – and that once expanded it can include what have been characterized as ‘relational values’. The point might be pressed that there are after all other uses of the concept of services that are wider than a narrowly market-based concept. Consider for example, ‘public services’ or ‘social services’ or the concept of ‘being of service to someone’. Care for other people and things might be included within this wider concept of services.

This line of argument concerns the scope of the concept of service. There are features of the Ayres and Kneese reading that are clearly narrowly economic in scope – in particular the characterization of services in terms of consumer preferences, either directly through their use or indirectly through their contribution to manufacturing processes. However, it is not the scope of the concept that has been the object of criticism in this paper. Rather, the paper has been concerned with two features of the very concept of service as it is employed in its standard uses.

1. 1. Services are instrumental in the sense of being external means to an end.

2. 2. An object, insofar as it is valued for its services, is valued de dicto. It is valued only as a vehicle for those services.

If services are understood in this way, then relations to particular others that are constitutive of well-being cannot be captured in the language of services. Since this is the standard use of the concept, there are good grounds to be sceptical of the view that the relational values that IPBES properly want to include can be included under the heading of services – that the term ‘service’ is appropriate for describing relationships to particular persons, places, activities and goods that matter for the lives of individuals. Interestingly, in more recent documents IPBES has shifted its language from ‘services’ to ‘contributions’, specifically to talk of ‘nature’s contributions to people’ (Pascual et al. Reference Pascual, Balvanera and Díaz2017). Part of the motivation for that move is a recognition that many values, in particular relational values, cannot be captured by the concept of ecosystem services. The notion of ‘contribution’ is undoubtedly an improvement with respect to the arguments of this paper in that it allows for both causal and constitutive contributions to well-being. Given that the concept of ‘ecosystem services’ in its current use is closely tied to the extension of the economic language of ‘capital’ and has the features just outlined, clarity of discussion may be best served by dropping the concept.

One final response to this argument is to question whether the concept of services needs to be understood as having the two features noted in its economic use outlined above: services are external means to a good; an object valued for its services is valued de dicto. Could there be uses of the concept that lack these two features? Consider the following passage from Robert Louis Stevenson in a discussion of services:

The concept of the ‘true services of life’ that Stevenson employs here is not one that ties services to the external causal means to some good, or makes the service provider a mere vehicle. The services that Stevenson outlines here, such as kind words and tender behaviour, are not just the external means to friendship. Rather they are the very stuff of friendship. They are themselves constitutive parts of relationships of friendship, not means to some distinct good. Friendship is a relationship constituted by reciprocal attitudes, behaviours and beliefs. Insofar as friendship itself is one of life’s central goods, offering and receiving those services form part of what makes a good life. Conversations with friends and caring attitudes and behaviour with respect to them are all part of what makes for a flourishing life. If Stevenson’s usage of the concept of services is taken seriously here, then it might be that one can extend the concept so that it includes relationships to particular others that are constitutive of a good life.

How far is this account of services applicable to the environmental sphere? There are clear problems in simply taking a relationship like friendship as a paradigm case. The attitudes, behaviours and beliefs that make up friendship are reciprocal in nature. This in part distinguishes friendship from other relationships, such as romantic love, in which reciprocation may be absent. With relationships to particular places central to a good life, like the case of the Narmada river to the community threatened by its loss, the concept of reciprocated beliefs, attitudes and behaviours of the kind found in friendship does not apply. The river does not reciprocate in this sense. However, the idea of services of life that are constitutive of the relationship might have some hold. The gifts of the river – ‘for her fish, her water, for vegetables that grow along her banks’ – are central to the relationship not simply as an external means to other goods, but also as central contributors to what the river means for that community and a way of life in which it plays such a central part. Fishing, growing vegetables, the other livelihood activities themselves make up the way of life that is valued and is threatened by loss. They are not external means to something else that is a way of life. At the same time, the relationship is constituted on the other side by the practices, attitudes and beliefs that are brought to the river by the members of the community. What makes the river central to the way of life of a community is not just the history of the community that is embodied in it, but its continued presence through these practices, stories and attitudes. The loss of a relationship can be the result of the direct destruction of the place – as is the case with the Narmada dam. However, its loss is also made possible through the loss of the beliefs and practices of a community that are constitutive of the relationship. The livelihood practices central to the meaning of place may disappear with technological change or the emigration of young people from the community. The beliefs and attitudes may no longer make sense to a person – she becomes alienated from them. Correspondingly, the evolving relationship to a place is sustained through engagement in practices, through stories told and retold, through everyday activities in which a place, like a river, are central. In that sense the services of life of the kind that Stevenson hints at are not just gifts from nature, but also services that members of a community bring to the place. A place is constituted by attitudes, beliefs and practices that members of a community bring to it.

Whether or not the concept of services can be rescued in this way is a question that I will leave open. However, if thus used it involves a departure from the standard economic sense of a benefit stream received. The concept of services, if it is to be used at all, would require a radical inversion of the traditional use of services in the economic sense. Services are not simply something received, but something that has to be brought to the relationship that they constitute.

Whatever the resolutions to this difference of whether the term services should be employed or not, neither view will give support for a policy of no net loss. For, once the concept of services is extended to include relations to particular others that are constitutive of a good life, then it does not allow for the substitution of one good for another that provides the same services. Where a relation to a particular is central to a life, the loss of that particular person or place cannot be substituted by another. The loss of that good, like the loss of a real husband or wife, cannot be substituted by another that provides the same services. The policy of no net loss relies on the concept of ecosystem services that I have criticized in the earlier sections of this paper. It assumes goods are valued de dicto such that any good can be substituted by others that provide the same services so that the aggregate remains the same.

One final objection that has been made to the arguments for the significance of relations to particulars is that it would lead to a form of radical policy conservatism where the value of relations to particular persons and places always trumps other goods no matter how important they are. This objection misses the point of the arguments in this paper. No such claim about comparative value of relations to others and other central components of human well-being has been made in this paper. Not all relations to particulars valued de re as particulars are constitutive of a good life. Many are not. Sometimes, where other vital needs are at stake, the relationships to particulars that are constitutive of a good life will be lost. The seriousness of conflicting vital needs might justify their loss.Footnote ¹² However, where they are lost there is a need to recognize the values at stake. Where a relationship to a particular that is central to the good life of a community or person is lost, there is no substitute. When a close relation dies the loss is a loss that cannot be replaced. If a husband or wife dies and the relationship was a good one, there is no substitute. The loss is a loss and nothing can replace or compensate that loss. Some goods are irreplaceable. A central problem with the policy aim of no net loss is that it fails to recognize the importance of irreplaceable goods grounded in relationships to particulars. It assumes goods are valued de dicto. In doing so it justifies the destruction of places that matter to persons and communities on the grounds that substitutes exist. However, environmental goods valued de re that are central to human well-being do not have substitutes. There is a necessary loss. Consideration of policy needs to recognize this and not proceed on the illusion that it is otherwise.