Loading ...
Sorry, an error occurred while loading the content.

204131Re: Universal language for communicating with space aliens

Expand Messages
  • Logan Kearsley
    Jul 25, 2014
    • 0 Attachment
      On 25 July 2014 13:41, Patrik Austin <patrik.austin@...> wrote:
      > [Logan:] > Consider "A forest is fallen in by a tree." In this English sentence,
      > "forest" becomes the subject, while "by a tree" is an adverbial
      > phrase. The translation into logical form, however, is the same. It's
      > still "tree PAT forest LOC fall EVT". You could even say "There is a
      > fall by a tree in the forest"- now you've made "fall" the subject, but
      > it's still represented in weird-syntax-lang by "fall EVT".
      >
      > That's all cool stuff. You're working on the logic of language to express one event in different logically plausible ways – it's somehow similar to analogous expressions in formal logic. That's exactly what I'm looking for.
      >
      > However, what you're doing is you're demonstrating how universal logic of language works when you were supposed to prove there's no such thing.

      I am doing no such thing. You *cannot* do any such thing without
      access to a large sample of actual language-using aliens whose
      languages you can survey for structural and functional properties. All
      I'm doing is demonstrating that language must conform to some rules of
      logic, which is universal. If "the universal logic of how language
      works" is just "how logic works, generically", then it's a useless
      concept. Just say "alien languages must be logical, because
      information theory", and leave it at that. Until we've actually met
      aliens, you cannot make any assumptions about how their languages
      might work beyond the basic restrictions of mathematics and logic, and
      what I'm trying to demonstrate is that that allows for way more
      variation and weird stuff than the strictures of natural human
      language do (with a side-order of "human languages can actually be
      weirder than you think").

      For reference, here's another (much better developed than me thinking
      about it for 5 minutes) conlang based on a very different approach to
      formal logic that also does not have clearly identifiable nouns,
      verbs, or adverbs, although in a much different way than
      weird-syntax-lang doesn't have them:
      http://conlang.org/cms/wp-content/uploads/zweaver_lcc5_slides.pdf

      > To go back to the analysis, 'forest LOC' can be analysed as a noun and a postposition that together form an adverbial. If you mean to have it all as one unit, you could spell it as one word: tree-PAT-forest-LOC-fall-EVT, as you might do in an agglutinative language.

      It can, but so can *every* other clausal argument. You can also
      interpret *every* clausal argument as being a verb, or as a noun
      phrase. Those distinctions do not exist in the language- they only
      appear when you try to translate it into a language that *does* have
      those distinctions.

      > However, I'm still going to break it down to smaller parts and look for the subject, verb, object and adverbial. You do this with Finnish which is a synthetic language, which makes it a lot weirder, but you do it all the same.

      You're doing it wrong. If you assume that these categories must exist,
      then you *will* find them, no matter which categories you choose. The
      cost of finding them is that your analysis will be inconsistent, and
      result in an incorrect understanding of how the rest of the language
      works. Finnish actually *does* have subjects and verbs, but you can
      discover that by examining Finnish entirely on its own terms, without
      previously assuming that those categories must exist. I could find
      articles in Russian or numerical classifiers in English if I wanted
      to- doesn't mean they'd actually be there. This is the kind of
      thinking that leads to people claiming that "you can't split an
      infinitive", even though English speakers do it all the time with
      perfect comprehensibility, or to Anglophone students of Russian
      embarrassingly ending their sentences with prepositions- it's applying
      the rules of one language to another, without regard to whether or not
      they actually make sense. If you insist on analyzing alien languages
      like this, good luck ever being able to decipher enough to hold a
      proper conversation with them!

      > Now I see that you didn't mean 'fall' to be a verb, but an event i.e. a noun. That's then the subject of the sentence whether or not it includes the words in front of it. In that case IMPL is the verb.

      Why does that one have to be the subject? You can translate it lots of
      different ways, so what language-internal criteria result in the
      conclusion that the phrase "fall" or "fall EVT" are the subject? In
      common parlance, a grammatical subject is defined as the thing that
      performs the action of the verb. That can be easily contradicted, so
      for Real Linguistics we're happier to define it as an argument that
      occupies a particular syntactic position with respect to the verb and
      is assigned to a particular slot in the verbs argument structure.
      Here, however, there is no other verb, thus nothing whose argument
      structure can be used to assign roles, and the phrase has no
      privileged syntactic position with respect to any other constituent of
      the clause. In that case, how can you reasonably define anything as
      being the subject?

      To support the fact that there is no verb- neither that phrase not any
      other clausal argument forms a syntactic head. It doesn't have an
      argument structure nor assign roles. It *can* be used directly as an
      argument to a different predicate, which is very noun- or
      pronoun-like. Basically, it doesn't have *any* of the syntactic or
      semantic features that commonly identify verbs in most human
      languages.

      IMPL actually has the best claim to being a verb, but it translates to
      "if", a subordinating conjunction!

      > The sub clause has the verb CAUSE. You're probably thinking of a word with a performative function (see J. L. Austin's Performative utterance). So it's a verb although the verb could even be 'CAUSE sound' depending how you make the rules.

      Sure it doesn't. I am not thinking of a performative- I'm thinking of
      a postposition that assigns the thematic role of causer. Having an
      argument structure and assigning a role sounds kind of verblike, but
      no more so than English prepositions are! Also no more so than every
      other postposition in the sentence. No, there are languages that allow
      multiple finite verbs per sentence (see "serial verb construction"),
      so that's not necessarily an argument against, but we can still see
      that none of the postpositions form syntactic units that dominate
      anything like a "verb phrase", and they don't assign roles to any
      other arguments. So, we're still on pretty shaky ground calling those
      verbs, either. Analyzing 'CAUSE sound' as a verb is particularly
      problematic because it's not even a coherent syntactic unit!

      For further reference, to demonstrate the constituent boundaries (and
      introducing NOT as a prefix unary logical connective), the following
      are all valid equivalent renderings of the test sentence in
      weird-syntax-lang:

      CLS tree PAT forest LOC fall EVT END IMPL CLS it CAUSE sound PAT
      future TIME END INT
      NOT CLS it CAUSE sound PAT future TIME END IMPL NOT CLS tree PAT
      forest LOC fall EVT END INT

      CLS tree PAT fall EVT forest LOC END IMPL CLS it CAUSE sound PAT
      future TIME END INT
      NOT CLS it CAUSE sound PAT future TIME END IMPL NOT CLS tree PAT fall
      EVT forest LOC END INT

      CLS forest LOC tree PAT fall EVT END IMPL CLS it CAUSE sound PAT
      future TIME END INT
      NOT CLS it CAUSE sound PAT future TIME END IMPL NOT CLS forest LOC
      tree PAT fall EVT END INT

      CLS forest LOC fall EVT tree PAT END IMPL CLS it CAUSE sound PAT
      future TIME END INT
      NOT CLS it CAUSE sound PAT future TIME END IMPL NOT CLS forest LOC
      fall EVT tree PAT END INT

      CLS fall EVT tree PAT forest LOC END IMPL CLS it CAUSE sound PAT
      future TIME END INT
      NOT CLS it CAUSE sound PAT future TIME END INT IMPL NOT CLS fall EVT
      tree PAT forest LOC END

      CLS fall EVT forest LOC tree PAT END IMPL CLS it CAUSE sound PAT
      future TIME END INT
      NOT CLS it CAUSE sound PAT future TIME END IMPL NOT CLS fall EVT
      forest LOC tree PAT END INT

      (I got tired of doing the inverted variations after the first six....)

      CLS tree PAT forest LOC fall EVT END IMPL CLS it CAUSE future TIME
      sound PAT END INT
      CLS tree PAT fall EVT forest LOC END IMPL CLS it CAUSE future TIME
      sound PAT END INT
      CLS forest LOC tree PAT fall EVT END IMPL CLS it CAUSE future TIME
      sound PAT END INT
      CLS forest LOC fall EVT tree PAT END IMPL CLS it CAUSE future TIME
      sound PAT END INT
      CLS fall EVT tree PAT forest LOC END IMPL CLS it CAUSE future TIME
      sound PAT END INT
      CLS fall EVT forest LOC tree PAT END IMPL CLS it CAUSE future TIME
      sound PAT END INT

      CLS tree PAT forest LOC fall EVT END IMPL CLS sound PAT it CAUSE
      future TIME END INT
      CLS tree PAT fall EVT forest LOC END IMPL CLS sound PAT it CAUSE
      future TIME END INT
      CLS forest LOC tree PAT fall EVT END IMPL CLS sound PAT it CAUSE
      future TIME END INT
      CLS forest LOC fall EVT tree PAT END IMPL CLS sound PAT it CAUSE
      future TIME END INT
      CLS fall EVT tree PAT forest LOC END IMPL CLS sound PAT it CAUSE
      future TIME END INT
      CLS fall EVT forest LOC tree PAT END IMPL CLS sound PAT it CAUSE
      future TIME END INT

      CLS tree PAT forest LOC fall EVT END IMPL CLS sound PAT future TIME it
      CAUSE END INT
      CLS tree PAT fall EVT forest LOC END IMPL CLS sound PAT future TIME it
      CAUSE END INT
      CLS forest LOC tree PAT fall EVT END IMPL CLS sound PAT future TIME it
      CAUSE END INT
      CLS forest LOC fall EVT tree PAT END IMPL CLS sound PAT future TIME it
      CAUSE END INT
      CLS fall EVT tree PAT forest LOC END IMPL CLS sound PAT future TIME it
      CAUSE END INT
      CLS fall EVT forest LOC tree PAT END IMPL CLS sound PAT future TIME it
      CAUSE END INT

      CLS tree PAT forest LOC fall EVT END IMPL CLS future TIME it CAUSE
      sound PAT END INT
      CLS tree PAT fall EVT forest LOC END IMPL CLS future TIME it CAUSE
      sound PAT END INT
      CLS forest LOC tree PAT fall EVT END IMPL CLS future TIME it CAUSE
      sound PAT END INT
      CLS forest LOC fall EVT tree PAT END IMPL CLS future TIME it CAUSE
      sound PAT END INT
      CLS fall EVT tree PAT forest LOC END IMPL CLS future TIME it CAUSE
      sound PAT END INT
      CLS fall EVT forest LOC tree PAT END IMPL CLS future TIME it CAUSE
      sound PAT END INT

      CLS tree PAT forest LOC fall EVT END IMPL CLS future TIME sound PAT it
      CAUSE END INT
      CLS tree PAT fall EVT forest LOC END IMPL CLS future TIME sound PAT it
      CAUSE END INT
      CLS forest LOC tree PAT fall EVT END IMPL CLS future TIME sound PAT it
      CAUSE END INT
      CLS forest LOC fall EVT tree PAT END IMPL CLS future TIME sound PAT it
      CAUSE END INT
      CLS fall EVT tree PAT forest LOC END IMPL CLS future TIME sound PAT it
      CAUSE END INT
      CLS fall EVT forest LOC tree PAT END IMPL CLS future TIME sound PAT it
      CAUSE END INT

      -l.
    • Show all 118 messages in this topic