Definition, Abstraction, and Language’s Limitation

Language is a strange thing, simultaneously a powerful tool and, perhaps, a limiting one.  As one of my favorite quotes asserts, which I have certainly deployed in posts previously, we’re trying to plumb the secrets of the universe using a tool intended to tell each other where the best fruit is.  This particular reflection is prompted by a conversation I had recently with my brother about how we define things and the abstractions involved.  Where I asserted that abstractions result in language’s limitations, my brother argued that language’s limitations cause its abstractions.  That, however, is getting ahead of ourselves, because we have to begin with a discussion of the nature of definition.

Consider a rock.  Now, as a (presumably) fluent English speaker, I’m confident that you have some idea of what I mean when I say “rock,” because we have a shared understanding, a shared definition, of this concept of a rock, this arbitrary sound which we’ve established to refer to the concept it labels; however, I am also confident that you and I are not imagining the same rock, because the term “rock” is an abstract one, referring to a broad category of objects which are considered rocks despite having vastly varying properties.  If I wanted to ensure that we are thinking of the same rock, I would have to better define the rock.

Well, science, that mixture of rationalism, empiricism, and skepticism, offers us the tools to define a rock objectively.  We can measure the rock’s properties: its mass, its elemental composition, the properties of the subatomic particles which compose it, the electromagnetic reflectivity.  Measuring these properties will define the rock in a way will allow anyone to share the same definition of the same rock, regardless of who or what is involved.  A rock’s mass will be the same whether it’s being measured by a human or an amorphous dark matter entity hundreds of lightyears away.  Thus, we remove the abstract quality of the term “rock,” and can offer a precise description of “a rock.”

It is still only a rock, though, and it does not encompass all rocks.  The term “rock” is really a shorthand to allow us to express concisely a broad category of objects with some shared qualities.  We can probably agree on what a rock is, but not on what it isn’t, and establishing a rigorous, universal definition for “rock” is impossible without relying upon abstractions.  This is why those annoying CATCHPA tools work(ed) – computers have difficulty understanding abstract concepts and identifying them in ambiguous contexts.  This should remind you of Platonic forms, which was in a respect intended to address this very property of language by supposing that there is a universal essence of concepts like rocks and chairs that can take so many different forms and diverse specific qualities.

This becomes a limitation of language because it requires us to deal in abstract terms, because when I say “rock,” we imagine different rocks.  It leaves ambiguity – but that ambiguity is also language’s strength.  It means that I can say “rock” and we will both reach a similar understanding of what I mean, even if the exact form is different, without needing to rigorously define the exact properties of the rock (or of all rocks).  Think of it like the strength of a phonetic alphabet compared to a symbolic or pictographic alphabet.  Each letter in a phonetic alphabet stands for something abstract, a name for a sound that may be different in different combinations, but only a few are required to express a nearly infinite variety of ideas and concepts.

Alternatively, from the philosophical standpoint from which my brother approached this discussion (as opposed to my physics-based approach), abstraction can be said to arise from the limitations of language.  This is more challenging to understand, and a more absolutist position, claiming that language itself and its in-build limitations give rise to abstractions like the one we encounter when seeking to define “rock.”  It tends to presuppose that it is impossible to define anything in a non-abstract manner, which I do not believe to be true.  Since language is fundamentally referential, no universal definition such as described above can be established.  Quantum physics’ observer effect lends some credence to this perspective.

Language, abstraction, definition: are these limits, or are they capabilities?  Are they features, or are they bugs?  This might be the more important question to consider than the one my brother and I discussed and which I explored in this post.  The abstractions required for definition in language might appear to be limitations, but they allow me to say that “Raven stood upon a rocky promontory,” without having to provide an exhaustive list of definitive, quantitative and qualitative measurements to fully constrain the system for you.  Maybe this is language’s weakness, but it is also it’s power.  It is why language is not a static equivalency table, but a living, organic system, not something that can be employed in a vacuum, but something that is employed for conversation, dialogue.  What I say and how I say it matters, but what you hear and how you hear it matters, too.  That’s how meaning arises, and it is more strength than weakness.