"The only words for smells we have are comparisons.
Compare our words for other perceptions: We have basic words for colours which are not dependent on the knowledge of specific substances or objects for their understanding. Red, for example. It is not a simile or a metaphor. All it means is red. Similarly with touch: we have smooth, rough, spiky, soft. With smell we must say rose, sweat, tarragon, concrete, wet dog.
We simply have no words for smells. We simply have no words for smells. Acrid, I suppose, comes close to being an abstract word for a smell, but it's more like a kind of smell than a specific one: It's more about an unpleasant burning sensation caused by a smell. We have lots of modifiers like that: Smells are pungent or heady or nauseating. But those words aren't specific smells in themselves.
We have poetic approaches too. Writers will say a warm smell, for example, but their synesthesiacal attempts just underline the paucity of available vocabulary."
I don't totally buy his absolutist claim that "the only words for smells we have are comparisons," at least not if we allow that there are truly words for other senses, but there might be something to this. The normal human eye has only three different types of photoreceptor cells that respond to different colours. These are the cones responding to short (blue), medium (green) and long (yellow-red) light. The tongue similarly has only four different types of taste buds that respond to five different kinds of flavours, though this isn't the simply one-to-one correspondence that exists in the eye.
I have not been able to turn up any such limit or classification for olfactory receptor neurons. And it makes sense that things with a limited pallet have specific labels but that those with more diverse basic components be described in other ways.
I'm less sanguine about Smith's analysis of touch. It seems to me that the textures he lists are kinds of texture rather than a specific ones. And he has carefully avoided discussing sound.
Anyhow, what about his claim that we have no words for smell? I propose at least three: sweet, sour, and bitter.
These are activated by G-channel receptors, which exist both in the mouth and in the nose. It is well acknowledged that our sense of taste is largely influenced by our sense of smell. So I'm not sure that we can clearly say that these are primarily tastes and only synesthesiacally smells.
But even allowing that they are only tastes, the word sweet meant "pleasing to the senses, mind or feelings" in Old English, and bitter comes to us from biting. These are the poetic approaches that Smith dismisses. Here his argument breaks down; either you accept that words based on the poetic approach qualify for both taste and smell, thus accepting words like: acrid, pungent, sharp, green, and fresh or you reject the poetic approach for taste and say that we don't really have sense words there either. Similarly, if salty qualifies as a taste, why would we not accept words like musty, fruity, flowery, oily, and metallic as smells?
But perhaps the most basic problem with the proposal is the idea that a word for something is somehow more genuine than a metaphor or paraphrastic description and that a language that is missing a particular word for a given concept is somehow impoverished thereby. Geoff Pullum exasperates: "Everybody thinks that the key thing about a language is which words it has." No! "Natural languages are much better thought of as systems of conditions on the structure of expressions (words, phrases, sentences). Some of the well-established conditions apply to word-sized units (it really is well established that dog denotes Canis domesticus, and that the is the only acceptable form for the definite article), but the constraints do not entail a roof on the number of words or prescribe which ones are genuinely in the language."