5 ms·

Ten things all languages have in common

12 points by adrian_mrd 4y ago

tgv 4y ago
All languages? Of course not. There are small languages which don't fit one or two categories. E.g., how can you have a dialect if there's only a handful of speakers in the same village left? Languages may have the capacity to develop dialects, but they don't necessarily have them. And not all languages can communicate all ideas. Piraha is a famous example of a language without counting, and which also lacks abstract words.
- ohCh6zos 4y ago
  To take this a step further, it seems easy to build a conlang that lacks one of these.
schwartzworld 4y ago
I wonder, in programming we have this concept of Turing completeness to describe the ability of languages to express a common standard of functionality. Is there an equivalent for communicative languages where all complete languages can express the same ideas and concepts as other complete languages?
- lupire 4y ago
  Define complete. Turning machines are only complete on a self-referential way, within the limited scope of computation. Human language is far more complex than a Turing Machine language or any other computational language.
- xaedes 4y ago
  They claim "All languages can communicate all ideas and feelings", but I think it is wrong. First of all there are ideas and especially feelings which can not be directly communicated with language at all. Surely you can label them and hope for the best, that the other side can relate to it, but you can not actually transmit the true idea or feeling you have. If language could truly communicate all ideas it should also be able to communicate all tacit knowledge, i.e. the term for the concept of "tacit knowledge" wouldnt be necessary to begin with. You can try to explain, but the explanation is only the map, not the territory. Or to relate it with computing terminology: Language is digital but existence is analog.
  - mcphage 4y ago
    They don’t mean that there’s a single word in every language for every feeling, but that you can use the language to express it—or at least get asymptotically close.
- xigoi 4y ago
  There's the idea of “semantic primes”: https://wikiless.org/wiki/Semantic_primes?lang=en https://wikiless.org/wiki/Semantic_primes?lang=en
  - jkingsbery 4y ago
    This link would be useful to see on first page of HackerNews. The Duolingo article - not as much.
noname120 4y ago
PR drivel.
- Frotag 4y ago
  It's duolingo. Article is about as surface level as any of their courses.
- londgine 4y ago
  The title made it seem like there would be interesting content like theory of grammar, but rather stupid (if not false) "facts" like all languages are equally valuable and equally complex.
xigoi 4y ago
I think they should've specified “natural languages”. Some of these aren't necessarily true for constructed languages, let alone programming languages.

12 points by adrian_mrd 4y ago

tgv 4y ago

All languages? Of course not. There are small languages which don't fit one or two categories. E.g., how can you have a dialect if there's only a handful of speakers in the same village left? Languages may have the capacity to develop dialects, but they don't necessarily have them. And not all languages can communicate all ideas. Piraha is a famous example of a language without counting, and which also lacks abstract words.

ohCh6zos 4y ago
To take this a step further, it seems easy to build a conlang that lacks one of these.

schwartzworld 4y ago

I wonder, in programming we have this concept of Turing completeness to describe the ability of languages to express a common standard of functionality. Is there an equivalent for communicative languages where all complete languages can express the same ideas and concepts as other complete languages?

lupire 4y ago
Define complete. Turning machines are only complete on a self-referential way, within the limited scope of computation. Human language is far more complex than a Turing Machine language or any other computational language.
xaedes 4y ago
They claim "All languages can communicate all ideas and feelings", but I think it is wrong. First of all there are ideas and especially feelings which can not be directly communicated with language at all. Surely you can label them and hope for the best, that the other side can relate to it, but you can not actually transmit the true idea or feeling you have. If language could truly communicate all ideas it should also be able to communicate all tacit knowledge, i.e. the term for the concept of "tacit knowledge" wouldnt be necessary to begin with. You can try to explain, but the explanation is only the map, not the territory. Or to relate it with computing terminology: Language is digital but existence is analog.
- mcphage 4y ago
  They don’t mean that there’s a single word in every language for every feeling, but that you can use the language to express it—or at least get asymptotically close.
xigoi 4y ago
There's the idea of “semantic primes”: https://wikiless.org/wiki/Semantic_primes?lang=en https://wikiless.org/wiki/Semantic_primes?lang=en
- jkingsbery 4y ago
  This link would be useful to see on first page of HackerNews. The Duolingo article - not as much.

noname120 4y ago

PR drivel.

Frotag 4y ago
It's duolingo. Article is about as surface level as any of their courses.
londgine 4y ago
The title made it seem like there would be interesting content like theory of grammar, but rather stupid (if not false) "facts" like all languages are equally valuable and equally complex.

xigoi 4y ago

I think they should've specified “natural languages”. Some of these aren't necessarily true for constructed languages, let alone programming languages.