'Sustainability' concerns the relationship between humans and their natural environment over time. But there are various ethical understandings of that relationship with quite different implications. Two popular definitions actually repudiate human interdependence with nature by either making human interests completely subservient to a sacred nature, or by making nature completely subservient to human interests. Gro Brundtland's famous definition points in the right direction by focussing on the goal of sustainably meeting humanitarian needs, but her picture of human interests seems too narrow and technocratic. What we need is a definition that is humanistic without necessarily being human-centred.
1. Perhaps most popular understanding of sustainability in the context of environmental ethics is the maxim 'do no harm'. i.e. human activities should be structured in such a way that they have the least possible ecological footprint. The strong point of this definition is in pointing out that human activity does do a lot of harm. E.g. the entire sub-discipline of 'ecological economics' has been inspired by this perspective to analyse the ways in which human activities impact on the environment, and on how much human activities depend on 'ecological services' (such as wetlands providing sea defences, forests as carbon sinks, etc). They point out, quite rightly, that the so-called New Economy that is supposed to be all about weightless services and information, actually retains an enormous physical industrial sector. They also point out that doing harm to the environment can come back to bite us, with estimates that China's GDP growth, if adjusted for the damage caused by air and water pollution to human health and ecological services, might actually be negative.
This makes possible a distinction between 2 forms of anthropocentricism. Man is the measure of all things - we can never leave our human perspective behind - but that doesn't mean that human evaluation of environments is limited to asking "what's in it for us?". We can bring a range of other perspectives to bear, such as asking "what do polar bears need to flourish" or admiring the intricate co-ordination of a bee colony independently of their use-value for human interests.
Nevertheless sustainability as 'do no harm' fails to be a morally convincing or practically plausible normative restriction on human activity. It is liable to fetishising aspects of the environment by dogmatically asserting that their intrinsic value has priority over human interests. But the intrinsic, human independent value of the environment is difficult to see, let alone the justification for its priority over human interests.
Ecosystems for example are just contingent arrangements of various species in a competitive equilibrium, and the analogy to an economy is an interesting one. Like economies they change form and make-up depending on the different forces and elements they are subject to (climate, different species, soils, etc), and those are always changing, with or without human interference. If any particular ecosystem is intrinsically valuable, then they all are, whether deserts or spruce forests, and no one particular ecosystem can be any more valuable than others, except for non-intrinsic reasons, like being rare, beautiful, or useful to us (or to other species we care about, like Pandas and their special bamboo).
When one eco-system changes into another, whether as a result of human action or other causes, it's unclear what net harm has really been done, but the form of the 'do no harm' principle considers all change as loss. A strict interpretation is so strongly conservative that it is incompatible with human existence, since it would require an ephemeral humanity that would float ghostlike above the earth without materially affecting it. The logic leads straight to the Voluntary Human Extinction Movement.
2. Another popular version considers the environment as a stock of resources and focuses on human inter-generational justice. i.e. making sure that we maintain the stock of resources that our grandchildren, or great-great-great-grandchildren might want to use. Here we have a straightforward constrained maximisation problem that depends on our time preferences: more consumption now at the expense of future consumption, or vice versa (i.e. 'Sustainability'). Therefore present consumption should be limited to renewable, preferably self-renewing, resources (i.e. not fossil fuels) that also do not destroy other resources (e.g. through climate change). This avoids fetishising the environment since it concerns itself directly with human concerns, and it raises the important issue of the interests of future human-beings (who are normally ignored).
But it still seems unduly conservative and restrictive in at least 2 senses: i) why only consider resources, and ii) why only consider future generations?
i) Why should we only think about the environment in terms of resources? This seems to unnecessarily restrict how we can value its other aspects, such as beauty. Further, why is conserving resources so important? Because they are supposedly scarce while human wants are unlimited? There is a kind of Malthusian logic here in which constraints are taken to be absolute (almost theological) before which humans can and should only submit. But Malthus's 'iron law' is empirically flawed, since we know from history that technological and social innovation can discover and produce new resources, from methods of producing power from fossil fuels or uranium to the 1970s green revolution in agriculture. There isn't much natural about 'natural resources': they are all produced by the combination of human knowledge, innovation, and labour, and their particular natural inputs are only contingent. But even if resources are constrained, why is saving them the only ethical concern, rather than using them efficiently (not wastefully) for valuable projects? Surely the point of resources is their instrumental value - what they can do for us. So what are we saving those resources for, if every future generation will also follow the same ethical argument and save them for their future generations?
ii) As well as the rights of future generations, shouldn't we be concerned with the rights of present generations? This sustainability principle criticises the status quo because too many resources are being used up (in the aggregate), but says nothing about the current distribution of that resource consumption, implicitly going along with its present great global inequality.
3. In contrast, Gro Brundtland's influential 1987 report On Our Common Future links environmental, intergenerational, and distributive concerns. Her famous definition of sustainable development is "development that meets the needs of the present without compromising the ability of future generations to meet their own needs" where 'needs' means overriding priority for the essential needs of the world's poor, under limitations that are not absolute but are imposed by the state of technology and social organization on the environment's resources and ability to absorb the effects of human activities. Brundtland was optimistic that economic growth that met those requirements could achieve a more prosperous, just and secure future.
This is a much better, and more ethical, definition of sustainability. It doesn't fetishise the environment, but incorporates it as instrumentally relevant into a just, feasible and human well-being centred view of progress. The only criticism would be that Brundtland's concept of the person is not ambitious enough: essential needs are not the only things people have reason to care about and use resources for, so restricting our concerns for sustainable development merely to sustaining needs is not enough (Amartya Sen in The Idea of Justice, and also this shorter LRB article).
4. Amartya Sen argues that we should be much more ambitious about expanding our idea of basic needs as we get richer (development means raising our standards as well as meeting them), but he also emphasises the importance of the agency of individuals - of freedoms - over and above our basic physiological needs. Sen's 'agency perspective' includes for example that:
i) people can have reasons for conservation that are non-instrumental to meeting basic needs, whether aesthetic or otherwise (this re-incorporates the possibility of 'do no harm' as one goal for human activity among others);
ii) there is a value to participatory public discussion about what sustainable development entails, even if some technocratic policy analysis is more efficient (if a society is going to decide to value the non-instrumental value of rare and beautiful species and ecosystems for example, perhaps at the cost of economic growth, this should be a public decision);
iii) people may want to distinguish and particularly protect certain opportunities apart from their aggregate well-being impact e.g. even if our grandchildren are much richer than us and can live for 200 years, we might still like them to be able to breathe fresh air outdoors (i.e. justice for future generations is more than securing their general purpose wealth - we may want to secure them particular things that we particularly cherish);
iv) it is important to consider not only aggregate well-being and specific opportunities, but also how they are accomplished (e.g. procedural constraints on policies like population control that fail to sustain our freedoms right now).
Sen therefore proposes what seems to me the best definition of sustainability so far:
the preservation, and when possible expansion, of the substantive freedoms and capabilities of people today without compromising the capability of future generations to have similar - or more - freedom.