Veganism in a world of causally efficacious consumption
Recall that in the Introduction, I mentioned two ways of responding to the causal inefficacy problem. One of these was to accept causal inefficacy and try to argue that consumption of animal products is nevertheless immoral. Thus far in the discussion of this chapter, this strategy has seemed to be quite difficult to pursue. We’ve considered a complicity-based version of this strategy^ (pursued by McPherson) and a respect-based version of this strategy (pursued by Hereth) and they have both run into serious problems. My own considered opinion is that this general strategy simply cannot succeed. 1 believe that if consumer behavior really is causally inefficacious in the sense that it truly makes no difference whatsoever to the lives of animals, then it cannot be wrong.
So now let’s consider the other strategy that we discussed in the Introduction. That strategy is to try to overturn causal inefficacy and show that our consumptive behaviors do in fact make a difference. In what follows, I’ll explain how we can do this and then I’ll discuss what it means for veganism.
The threshold model
The most widely discussed and influential response to the causal inefficacy problem derives from what is known as the threshold model.The threshold model is illustrated in the following vignette offered by Shelly Kagan:
[T]here are, perhaps, 25 chickens in a given crate of chickens. So the butcher looks to see when 25 chickens have been sold, so as to order 25 more. Here, then, it makes no difference to the butcher whether seven, 13, or 23 chickens have been sold. But when 25 have been sold this triggers the call to the chicken farm, and 25 more chickens are killed, and another 25 eggs are hatched to be raised and tortured.12
If you purchase a chicken from this butcher every day for a long enough period of time, you will on average trigger the birth, suffering, and killing of one chicken for every purchase.
The key claim of the threshold model is that the butcher scenario is structurally analogous to meat purchasing in the real world. Specifically, the claim is that whenever we purchase meat, we take a small chance of triggering a large increase in production—and the probability of triggering such an increase is inversely proportional to the size of the triggered increase; i.e., any given purchase has some small chance of being what we may call a triggering purchase—a purchase that triggers some large increase in production and (hence) causes a lot of animal suffering.
Importantly, this model accommodates uncertainty. The model is silent about how probable it is that any given purchase is a triggering purchase, and equally silent about how large the effect of any given triggering purchase will be.
If the threshold model describes consumption in the real world, then it answers the causal inefficacy problem. That’s because the threshold model implies that some of our purchases, namely triggering purchases, do indeed have victims, contrary to what the argument that generates the causal inefficacy problem implies. A consumer will typically be unable to know which of her purchases are triggering purchases but, if she purchases animal products regularly, she’ll have reason to believe that some number of her purchases are triggering purchases and thus have victims. Moreover, the view implies that when a purchase does have victims, it has many victims, and thus does a lot of harm. On this type of view, the decision to purchase animal products is risky from a moral point of view. Similarly, given that (under normal circumstances) we have no weighty reasons13 to purchase meat, it is wrong to purchase meat (under normal circumstances) because doing so risks causing great harm, according to a view that I will call the market threshold idem.
The market threshold view is not just a tidy story. There is good reason to believe that it is at least approximately true, as Steven McMullen (an economist) and Matthew Halteman (a philosopher) have explained.14 After critically examining the arguments of philosophers who have promoted the causal inefficacy problem, they conclude by endorsing the standard threshold view:“if there is some probability (1/n) that any given purchase will occur on a threshold [i.e., will be a triggering purchase], then the threshold action will trigger a reduction in production of around n units, yielding an expected impact equal to l,”and so, for example,“there will be a close to 1-1 relationship between the purchase of a chicken and the expected impact on production.” In short, on a long enough time scale, a consumer of meat can expect to cause the existence, suffering, and death of about as many animals as she consumes, and similarly for consumers of milk, eggs, and other animal products.
Additionally, McMullen and Halteman observe that shifting consumption away from animal products can be especially beneficial due to several different additional causal roles of consumption beyond those specified in the threshold model.These include effects on supply chains (when the number of animals killed decreases, this also decreases the number of animals needed elsewhere in the supply chain, e.g., animals used for breeding purposes), network effects (consumers’ choices affect other consumers’ choices: e.g., rising tofu consumption makes tofu more visible and available, thus potentially causing additional consumers to consume more tofu and less meat), and scale effects (as the market for vegan alternatives increases, economies of scale kick in, prices drop, and veganism becomes more affordable, hence more attractive).