What Does Shit Smell Like?
Or: is your shit the same as my shit?
If flies are conscious, I suspect that shit smells quite pleasant for them—perhaps somewhat like freshly fried falafels or steaming brownies. Maybe the smell is something completely unimaginable for us humans, but it sure must smell nice, judging from how much they love the stuff! In fact, even our best friends, the dogs, love to gobble up nuggets of brown gold (as I’m sure they would call it), so I assume it also has an appetizing smell for them. This suggests to me that shit doesn’t have a smell, or well, that it doesn’t have one smell; there is no real smell of shit.
Maybe the pleasantness of a smell can vary completely independently of the quality of a smell, such that the quality of a smell has nothing to do with how pleasant it is. If this is true, then flies and dogs could have the exact same smell of shit as me and still find it pleasant. This seems implausible from reflection on the quality of smells in that the pleasantness of a smell seems to simply be constituted by the quality of it—but that is certainly only a fallible guide. It also just seems plausible that flies can and do smell shit differently from how we do. In any case I hope you will at least indulge me for the duration of this post in supposing that the smell has a different—and quite pleasant—quality for flies, dogs, and other creatures who eat shit.1
If there is no one smell of shit, then is it really shit that were smelling? It appears to me that this poses at least a small problem for naive realism.2 Naive realism is the view that in perception, we are directly presented with objects in the world—that the object of perception is the object in the world. So when I see an apple on a table, what I see is the apple, and not some representation of it. Likewise, when I smell shit, I smell shit. This means that smelling like shit is a property of shit. An implication of this seems to be that if both I and a fly veridically smell shit and it leads us to perceive different properties (call them H-shit and F-shit respectively), then shit has contradictory properties! After all, H-shit is not F-shit, and they are each the full smell of shit—the real deal. It looks like we are in deep shit.3
The naive realist may liken this to another problem: when I look at a table from the side its top appears to be a trapezoid, but when I look at it from the top it appears to be a rectangle (i.e., not a trapezoid). This means the table has contradictory properties! The answer here is to say that the appearing-to-be-trapezoidal is only a relational property, and so the table has no contradicting non-relational properties: it only has the non-relational property of being rectangular. The problem with this is that we would then have to say that smell is a relational property, but that is very implausible on naive realism—shit really smells like something (namely shit). I think the disanalogy is that shape is a primary quality (things like size, shape, position, etc.) while smell is a secondary quality (things like color, taste, smell, sound, etc.). Although the appearance of primary qualities can be relational, even under veridical perception, the appearance of secondary qualities cannot. If I veridically perceive a red apple, then the apparent size and shape may vary depending on where I am standing, but it will always appear red to me, and no matter from what angle I veridically smell a nice big pile of defecate, it will smell like shit, as I am sure you are well aware.
A more apt analogy here, then, is color. At least on the face of it, naive realism tells us that objects have colors—apples are red and shit is brown (if everything goes well). After all, at least some of the appeal of naive realism is that it preserves our perceptual appearances, a part of which are that objects actually have color. Likewise, objects actually have smells. This means that the analogy should instead be something like this: Two people, A and B, are looking at the same wall. A sees it as white, and B sees it as orange. On this sort of naive realism, both A and B cannot be perceiving veridically (perhaps B forgot to take off their orange sunglasses), since something cannot be both white and orange at the same time in the same respect.
A way to settle disagreements about color on this sort of view would be to say that the color of the wall is the color it appears to have under ideal viewing-conditions (that means off with those sunglasses, B). But what are the ideal smelling-conditions of shit? Is the ideal shit-smeller a fly or a human, or even a dung beetle? (those guys seriously love shit!) It would seem that we would have to adjudicate this if the smell differs between us, meaning I or the fly (or both), contrary to what we stipulated, are not veridically smelling shit. I don’t know about you, but I think flies know their shit a lot better than humans do! This seems to lead to a strong sort of smell-skepticism: how can I know that my shit is the real shit?
It looks like a fruitful route for the naive realist may be to abandon the idea that objects have real secondary qualities that we can perceive, or that they are perceiver-relative. This mat open the door too much for the veil of perception for the naive realist’s liking, however. After all, if my perceptual appearances of secondary qualities are so strong—it strongly appears that the piece of shit in front of me is brown and that it smells like… well, shit—and they are not actually of objects in the world, then why trust my other perceptual appearances; if by smelling shit I am not actually smelling a real property out there, it becomes harder to see what the motivation is for thinking that a creamy, brown turd is the object of perception when I look at it.
Or formulated in another way: the naive realist will have to give an account of what it is we are perceiving when we perceive secondary qualities—if the smell of shit is not something inherent in the anal discharge we find around us, or only exists relative to smellers, then there must be some additional element other than shit being the object of my smell-perception. If it was just shit, then there should be no difference between flies and me, since the existence and qualities of shit itself do not depend on someone being there to smell it, given realism. Perhaps it is merely some sort of mental representation that we map onto the primary qualities that we perceive. But once we allow that our regular perception is pervaded by this sort of “perceptual mediator”, it is just no longer clear what is the appeal of naive realism is anymore. You directly perceive the shape of the turd, but the only way you do this is by perceiving the contrast of the hazelnut-brown of it to the off-white of the public toilet you just desecrated. And so it is just not clear what role the turd itself is actually playing except a causal one. More generally, if color, taste, smell, sound, etc., are all perceived through an indirect mode of perception such as mental representation, then where are the ordinary objects to be found in our perceptual experience—where is the actual turd in your visual field behind all the colors shaped turd-wise?
I think the best way to view the argument I have given here might not be as a rebutting argument against naive realism. Rather it should be seen more as undercutting it: while naive realism can plausibly accommodate the challenges here, doing so seems to lead to a place where it is no longer obvious that the view is true to our original motivations. If we have to abandon the view that shit really has a smell, and concede that our visual experience is completely filled with things that are not the objects perceived, it is no longer so plausible that a naive realist view is very attractive anymore.
The conclusion here is in a sense a welcome one: your perceptual experience is never constituted by shit. So perhaps naive realists should not be so opposed too giving up their views. I think philosophy of perception is maybe one of the hardest areas of philosophy, though—combining some of the hardest parts of philosophy of mind and epistemology—so I don’t have very strong convictions either way. But these considerations at least seem to pose some problem for naive realism. Regardless of what you think, I hope we can all agree that it is just fun to talk shit sometimes.
Regardless, there are other lines of evidence that strongly point to the relevant variation, for example color blindness, or studies like these.
I originally said that direct realism was my target, but that is a more general genus of views, which includes naive realism—thank you to Alex Popescu for clearing up my confusion on this!
I’m sorry.
Substack, asking the important questions
A quick pedantic comment: direct realism (as treated in the literature of philosophy of perception), is just the thesis that our perception puts us into direct contact with external properties/objects. Exactly what this entails will vary from account to account, but it’s generally understood to entail that the contents of perception (as well as the contents of our perceptual beliefs + discourse) are about external properties and objects. Note that this says nothing about whether perception is essentially representational or not. Contrast this with indirect realist accounts like the sense data theory, which claims that the contents of perception literally are sense data. When you think about an entity that you perceive, you’re just thinking about a sense datum, not some object out there in the world.
So you can believe our experience is just a representation of what an external object is, but still believe in direct realism. In fact, representationalism is billed as a form of direct realism. It is naive realism which is the view that external objects and/or qualities literally constitute the objects and/or qualities in our experience, which the representationalist denies.
What you’re really getting at here is whether the qualities and properties of our sensory experience are *worldly*. That is, whether they are shared in common with the external objects being represented. Many representationalists hold that the answer is yes, that when we represent an external entity, we somehow incorporate its actual properties and qualities into our experience.
I think this is very dubious (at least if you buy into a naturalist account of representation), not least for the reasons you mention here. But even if we reject the worldly properties assumption, that doesn’t necessarily entail we have to abandon direct realism, since we might think that the contents of perception are still directly about the external entities themselves. For example, you might be an epistemic and semantic externalist who thinks that perceptual content is just determined by our brain states being in the right sort of causal contact with these external entities, irrespective of whether perceived entities and actual entities happen to share properties in common.
Hope this helps!