Have you had an experience you couldn’t quite put to words? Or understood some things that cannot be described well, and everything you tried to describe it in feel like poor analogies of it? Or that you even have to resort to using analogies to begin with?
And then someone mentions a word that sounds familiar, and suddenly, the connection makes sense. It made sense for the word to mean the experience/series-of-events/phenomena that you had experienced/understood.
Earlier this afternoon I had that experience. I had experienced something that is really difficult to describe, and put to words. I took a lot of notes about it, but I wasn’t able to accurately or satisfactorily explain it with words. What the experience was and the topics it surrounded is not of much importance, nor is it profound because I spent the rest of the afternoon obssessing about the fact there are no names to describe exactly what I had experienced.
In fact, the whole meta-ness about names makes even writing this blog post a little difficult, but I hope I am able to express what I mean quite clearly.
Names are pretty important, because without them, we do not understand the world. In fact, when you name a colour, you actually start perceiving the colour as a separate colour, as did the Chinese and Japanese discovered when they named the colour blue.
Given that names are pretty important, there are a lot of problems with names.
The Problem With Names
To talk about names clearly and unambiguously, some definitions are in order. Magritte’s Treachery of Images would be best used to illustrate my case of what a name is. A name, by my definition for the purpose of this blog post, is simply a reference. Just as the painting of the pipe is referring to the “idea” of a pipe (such ideas may be concrete or abstract, but I am getting ahead of myself), a name simply refers to something else. What this “something” is, can be concrete (keys, car, house), of it can be abstract (Einstein’s Mass Energy equivalence).
Both nouns and verbs are names, for different things. Nouns are typically names for real-life objects and abstract ideas, while verbs are names for actions and activities. For this part, I’m going to mainly concentrate on using nouns as an example, but you can easily extrapolate my arguments for verbs.
Ambiguity, Part I
The first problem with names is this: The thing that the name refers to is ambiguous in most situatons.
Arguably, (in-situ) context helps. For example, if I were in a completely empty room with you with a pipe in the middle of the room, if you mentioned “pipe” or “the pipe”, it would occur to me that it refers to the pipe in the room. Otherwise, a “pipe” simply refers to the my idea of what a pipe is, which could be any one of these:
Or a whole other lot of things and ideas.
And that brings to conclusion part I of Ambiguity – a name can refer to a lot of very different things.
Ambiguity – Part II
Even if we add a lot of context to a piece of communication (for example, you may refer to smoking certain herbs that activate the cannabinoid receptors in the brain* because let’s face it, who uses tobacco pipes nowadays? ), the name could still be highly ambiguous.
Imagine if you will, you encountered a tribe, C whose language has no name for the word “pipe”. But yet they clearly do have a device they smoke pot out of. Imagine this tribe has to regularly build and dispose of such pipes on the fly, as such they have no need to give it a name (let’s say they use apples to make the pipes). They have a name for apples, and despite using apples as pipes, they have no words for a pipe. They instead, call it by either what it is (an apple), or by its function (i.e. “giver of smoke” or “smoke gift” * You’ll see later that these two are essentially the same, and not a noun at all )
The key idea is that people of Tribe C knows what it is, but have no specific name for it.
If on your first contact with tribe C, you pointed out that the thing they use to smoke out of is not only called an apple, it’s also called a pipe, they may not recognize it as a distinct noun. Afterall, it is in their culture to consume the pipe after combustion* mmm, munchies , and they never needed a name for it because it’s a throwaway thing.
Even so, there has to be a shared understanding of what a “pipe” refers to. If you were making first contact with the tribe, and showed them your pipe, and called it a “pipe”, they may think that only that is a pipe.
This exactly is the problem that artificial intelligence faces – a “pipe” does not only refer to one thing. It refers to a set of things that share certain features with each other (the basic pipe is simple: 1 hole for the herb, 1 hole for the smoke to get out of, and optionally one carb). It’s the idea, or concept of a pipe that is shared.
A “pipe” can essentially mean any of these and more:
The above ambiguity can of course be cleared up further, using deictic demonstratives like this or that; or definite articles like the. But not all languages have them.
My Issues With Naming
Often when I am explaining things, I will slip into jargon. Friends have often told me that I’m always arguing semantics. And this is exactly the problem I have with naming things.
Most things are poorly defined. To the point that at the beginning of each discussion I have, I tend to actually lay out the definitions clearly and unambiguously (which, try as I can, isn’t as easily achievable as it seems). This makes me sound rather pedantic and nitpicky.
It also makes me look rather inconsistent. The definition of things change from discussion to discussion – sometimes the scope of definition is wider, and sometimes it is narrower, depending on the context of the discussion. And yet, they share the same name.
One of the most interesting development in computer science is content addressable storage. Every file has a unique hash to it (made from the contents of the file obviously), and the file’s name is its hash, which is its contents. If the file changes, a copy is made with a new hash. For all intents and purposes (barring strange collisions), the hash is the content, and the hash is the name of the file.
This of course, cannot work in real life. There is no hashing function in real life that can describe an action, or even an object. And it would be laborious if we were to call everything by the contents of the object (content can be features of the object, or even the action of the object). But let’s imagine it can. Instead of a hash, the name of a thing in the real world is called by its features and properties. This leads to one of two possibilities: Nounless languages, and verbless languages* There has to be some level of abstraction. I think humans are comfortable with the abstraction at the physical layer of life .
A nounless language would be quite hard to imagine. But apparently they do exist. The Salishan group of languages are apparently nounless (it’s a debatable feature), as are the language of Riau in Indonesia. In those languages, a duck is “one that exhibits duck like behaviour”. Think off it as duck-typing of human languages.
A verbless language is even harder to imagine. I don’t think there are any in real life. But let’s imagine how some of the language would occur. Instead of saying “I eat food”, you’d probably say something along the lines of “I mouth food”.
What To Do About Naming
To be honest, I don’t think there is a good solution at all. Creating a language that is precise and unambiguous is nearly impossible. Programming languages come closest to being precise and unambiguous, but communicating in code would be a nightmare.
Everything I wrote so far can actually be applied to programming languages. Nounless languages are essentially functional languages (well, kind of. The analogy breaks down upon a lot of further thought). Verbless languages are essentially object oriented programming languages. But these are mere analogies. There things that are yet to be named, that we do not know of.
TL;DR – Names are a pain in the ass, especially when you think about it. Now, tell me what you think