ยป Grammar in Context
Natural languages are ones that humans naturally speak. Constructed languages like Klingon, Esperanto and computer programming languages are not natural. Languages are how people do speak, not how some experts think a proper person should speak.
There’s a theory that the grammatical rules of natural languages only take local sentence context into account. For example subject and verb must agree in a clause, but if an independent clause is grammatically correct, then it can’t become grammatically incorrect in a compound sentence.
A number of counter-examples have been given over the years. From what I can find in Google Scholar, the argument ended in 1985 in a draw. Many contemporary linguistics and computer science textbooks say in passing that English has non-local grammatical rules. But all the examples of non-local rules I can find seem to have one of two flaws:
- Underestimate how big “local” can be: “Is that carrot cheesecake or carrot CAKE-cake?”
- Confuse grammatical rules with meaning: “Alice, Bob and Malory are two of my favourite people.”
Languages with only local grammatical rules are known as “context-free” and they can be formally specified using a context-free grammar. Computer scientists use formal grammars to program computers to understand languages. Processing constructed langauges like programming languages, mark-up languages (eg: HTML) and domain-specific languages (eg: mailing addresses) is a big part of what computers do. And it would be nice if computers could also understand natural languages.
Actually, Esperanto, the constructed language, is a natural language. It started out artificial, but its inventor set it free soon after publishing it in 1887. People picked it up and ran with it, and the rest is history. Like every other natural language, it is used in all facets of life, and evolves as dictated by the community of speakers. Its design was sound, however, so it hasn’t evolved radically in 123 years, remaining as simple, logical and regular as when it was published.
The best way to see this is to try it out. Lernu.net offers multiple free self-study courses (many Esperanto speakers these days are self-taught, and it’s easy enough that you can learn it effectively that way.
Michjo
22 Aug 10 at 10:18 pm
The reason I wanted to exclude constructed languages from this discussion is that it would be easy (?) to construct a language that is not context-free. I assume Esperanto is context-free, so it doesn’t really matter here.
That being said, Michjo, perhaps you should take up your beef with Wikipedia:
Jared
23 Aug 10 at 9:57 am
Thanks for pointing out the Wikipedia article on natural languages. I guess I’m of the philosophy that natural is as natural does; if it waddles like a duck and quacks like a duck, it’s a duck. Having experienced Esperanto, I have seen first-hand that it behaves in every way like languages everyone would consider natural and living, and so it, too, is natural and living. The only thing artificial with regard to Esperanto is the taxonomy that declares it to be something it is not. But enough of my ranting
.
Esperanto grammar is actually surprisingly informal. The formally specified grammatical rules are only 16 in number; most of what we now call “Esperanto grammar” is intuited from the usage observed in collected writings by its inventor and subsequent Esperanto speakers. While Zamenhof et al. strived for simplicity in their writings, imbuing the grammar with the same simplicity, because it is determined largely by usage as opposed to decree, I’m not sure it’s impossible for correct, proper Esperanto to be context-sensitive.
Michjo
23 Aug 10 at 7:36 pm
[...] Recent Comments Grammar in Context (3) [...]
MentalPolyphonics » Generating Random Sentences from a Context Free Grammar using VisualBasic for Excel
24 Aug 10 at 10:08 am
Wow, Wikipedia makes Esperanto seem much more elegant than evolved* languages (“all verbs are regular”).
Whether Esperanto is context-free or not probably rests on the same arguements as for English. For example, whether respektive (respectively) is a syntactic keyword or serves as just another adverb at the grammatical level.
The idea of deriving grammar from published texts for a supposedly “constructed” language is fascinating. Thank you for bringing it to my attention.
* A less weasely word than “natural”.
Jared
24 Aug 10 at 2:03 pm