Grammatical categories (categories like ‘noun’, ‘verb’, ‘adjective’, etc.) may seem mundane, but they actually bring tremendous power to human language.
In fact, without categories our languages would be puny, boring, and—most disturbingly—impossible to learn. This post explains why.
Note: this is the second in a series of five posts on the topic ‘Capturing Infinity in Natural Language’
Language without Categories?
Suppose, as a kind of thought experiment, that there were no categories like ‘noun’, or ‘verb’. How would you learn to talk? Specifically, how would you learn to order your words into sentences?
Well, without categories there could be no word-order patterns—no generalizations about how types of words fit together. Thus, to learn to make sentences, you would have to memorize them one by one, like this:
This account may sound appealing, but as a model for language learning it has two fatal flaws:
- First, to learn a language in this way you would have to memorize a very large list of sentences. In practice, this list would have to be billions of items long—and in principle (as the previous post explains) it would have to be infinite. Either way, it’s too much to memorize.
- Another problem with this model is that it only explains your knowledge of sentences that you have heard someone else say. But how can we then explain your creative power with language? After all, you can make any number of sentences that no-one else has ever said—but you can’t memorize something you’ve never heard!
The Power of Categories
So what did you learn, when you learned to make sentences?
Well, obviously, you learned patterns. That is, you learned generalizations about how to order words together. And, in every language, those generalizations—key parts of any complex linguistic system—are all based on grammatical categories. That’s why grammatical categories matter.
For example, at an early age you learned to recognize these categories (among others):
- Noun (‘N‘)
- Verb (‘V‘)
Then, once—and only once—you could ‘see’ words in terms of these categories, you gained the ability to recognize sentence patterns like these templates:
- N V
- N V N
We’re going to modify the format of the templates shortly to make them even more powerful, but even the simple format shown lets you make very large numbers of sentences—including brand new ones. It works like this:
How many sentences can you make? It depends on how many words you know, and the precise formulation of the templates—but the effect is always multiplicative. And, as the table below illustrates, it’s pretty easy to multiply your way up to a billion:
|# Nouns||# Verbs||# of sentences you can make with template N V
(=N x V)
|# of sentences you can make with template N V N
(=N x V x N)
There’s no way you could ever memorize billions of individual sentences. But you can make billions of sentences—and this is why.
As you add more vocabulary and templates, the number of sentences in your power grows exponentially, easily to many billions, or even trillions—yet, the number of things you have to learn remains quite manageable. And all because of grammatical categories!
With the format for the templates given above (N V, N V N, etc.), you still do reach a limit—albeit a massively large limit—to how many sentences the system captures. To get to infinity, we need to modify the format of the templates, and that will be the topic of the next three posts in this series.
But the point still stands: grammatical categories are the foundation for any kind of sentence pattern—and this makes them the first step to capturing natural language infinity.
- Post 1 – World’s Longest Sentence? (Introduction)
- Post 2 – The Power of Grammatical Categories (This post)
- Post 3 – The Infinite Power of Generative Rules (Coming soon)
- Post 4 – The Power of NP (Coming in a while)
- Post 5 – VP, S, and Infinite Syntax (Coming last in the series)
Because grammatical categories bring such power, all human languages have some version of them. Languages differ somewhat in exactly which categories they have (for example, German does not distinguish ‘adverbs’ from ‘adjectives’, as English does)—but there are also universal features to the categorial systems, plausibly based on innate structures.
Some languages make it very easy for children to ‘see’ the grammatical categories of words, by for example having a special prefix on nouns, another prefix on verbs, etc. In English, children need to cue in to more subtle features, such as the fact that only certain words can bear tense marking (which for English defines the class ‘verbs’).
If you are interested in language acquisition, you should be aware that children do go through a period where they learn just by memorizing full sentences—but they soon move on to generalizations based on categories. Surprisingly, when they start to generalize children’s speech errors often go up, because they can over- (or under-) generalize the patterns—which is why advances in learning sometimes give the superficial appearance of going backwards!
Recognizing the power of grammatical categories is not exactly a new discovery: in fact, it was one of the earliest things linguists talked about. Sanksrit grammarians in ancient India were the first to introduce the idea, and the ancient Greeks refined it into the categories we still talk about today.
Bakker, Egbert J. “The Birth of Grammar in Ancient Greece.” A Companion to the Ancient Greek Language. Chichester, West Sussex, U.K.: Wiley-Blackwell, 2010. Print.
Eisenbeiss, Sonja. “Syntax and Language Acquisition.” Syntax and Language Acquisition. Academia.edu. Web. 31 Jan. 2015. <http://www.academia.edu/1220666/Syntax_and_Language_Acquisition>
Lasnik, Howard, and Marcela A. Depiante. Syntactic Structures Revisited Contemporary Lectures on Classic Transformational Theory. Cambridge, Mass.: MIT, 2000. Print.
Robins, R. H. A Short History of Linguistics. Bloomington: Indiana UP, 1968. Print.
Wiltschko, Martina. The Universal Structure of Categories: Towards a Formal Typology. Cambridge, 2014. Print.
Zoomer Zuppies available here.
Copyright © 2015 Dexterous Tongue