Garcia-Portela, L. (2020) Moral Responsibility for Climate Change Loss and Damage: A response to the Excusable Ignorance Objection. Teorema. International Journal of Philosophy , 39 (1) :7-24

I don’t normally read this kind of paper, but I should because I really enjoyed it. It’s clearly written and nicely argued.

I’d come across the “excusable ignorance objection” before – the idea that “an agent should not be considered morally responsible for the harmful consequences of his actions if they could not have known or foreseen them at the time his action took place.” – in conjunction with event attribution, but not in those exact terms-1.

Normally, event attribution is set up as a dichotomy between a “natural” world and the world as it is. This has always struck me as a bit strange. If event attribution is to be used to determine culpability (in some sense) then one can hardly blame those who burned fossil fuels before the deleterious effects of greenhouse gases were known. If there was a choice to go on burning fossil fuels (or not) after we knew of these effects then the question really ought to be what harm the additional greenhouse gases caused. Rather than comparing a “natural” (slightly odd0) world with the modern world, it would make more sense (to me at least) to compare the modern world with a realistic world that might have been3. What that looks like though, I dunno0.

Anyway, this paper takes a different tack. It analyses the notion based on a particular definition of moral responsibility:

To say that an agent is morally responsible for something is to say that that agent is an appropriate target, in principle, of request for justification regarding that thing and that she is eligible, in principle, for a variety of moral responses depending upon how well or poorly she meets the justificatory request.

Smith 2015 (I’d give you the link but the reference is missing from the version of the paper I read)

The example offered to illustrate this is a truck driver1 who runs over and kills a child when the child leapt out from behind a rock. There was no chance for the driver to anticipate the situation, manoeuvre out of the way or stop. In that case, while the driver might not be responsible for the death of the child, they are still expected to offer their version of events (justification) and there may be further expectations based on what they say (moral responses). Even in the situation as described, where the driver is clearly not at fault, we might expect them to express regret or sadness.

In a similar way – and leaving aside the question of economic reparations, which are a separate issue – while historical emitters of greenhouse gases might not have known better, they are still responsible in the above sense. There are all manner of legitimate questions that could be asked and these might lead to a variety of symbolic moral responses, or so argues the paper.

I’m simplifying the argument here. It’s considerably more nuanced than my summary (as you might imagine). On the other hand, I’m not convinced it really engages with the full difficulty of the situation either.

The illustrative example is straightforward, but reality is considerably more complex. We’re not dealing with individuals, we’re dealing with almost everyone who ever lived. There’s a footnote which points out that the “agents” here are considered to be States, but it strikes me that the rules for states are different to those for individuals. The causal chain we’re dealing with is considerably longer and is part of a broader network of causes and effects and choices. In this situation, individual choices do not inevitably lead to particular outcomes and it is not clear how to quantify the connections.

Anyway, more reading for me.


-1 I find the exact term rather pleasing: the excusable ignorance objection. Like a Robert Ludlum novel.

0 If a storm floods a city and someone does the standard event attribution on that event, then in the “natural” world there is no city to flood. Or there is a city, but it somehow reached a population of millions without emitting a single molecule of carbon dioxide but is at exactly the same level of flood preparedness2. In a sense, event attribution usually dodges this criticism by focusing on the meteorological aspects of the event, but the current preferred approach chooses an event class that is impact focussed.

1 You can add details to this, I suppose: driver was driving below the speed limit, driver was sober, driving conditions were fine, truck was roadworthy etc.

2 This is why we need science fiction writers to collaborate with us on event attribution papers6

3 though for other applications, there might be better alternatives. The OG event attribution paper by Stott et al. on the European summer heatwave of 2003 compared the 2003 summer with typical future summers. When we’re using event attribution to inform adaptation decisions5, the choices we have aren’t between a “natural” world and the world as it is, but rather between different attainable futures. We’re not going back to the “natural” world any time soon4 and greenhouse gases are very likely to continue accumulating.

4 Not for many thousands of years it seems.

5 Another thing that seems strange to me. We shouldn’t just adapt to things that have happened. We should take into account things that could happen based on a assessment of the probability, impacts, vulnerabilities etc. If we can do all that math, but wait around till something happens to deploy it then I feel like we’re doing something wrong.

6 Though imagination of another type might be needed for philosophy papers on the same subject1.