Aistritheoir

What is this?

After seeing some cool work by @nmashton with GF, I was super excited to start hacking around in it...however, the syntax is kind of opaque and I couldn't quite figure out how to get it onto the web. So this is an attempt to take the idea of describing a natural language grammar programmatically to the web. At the moment it's mostly being used to help me learn Hungarian (and Test Driven Development/Coffeescript, but those aren't nearly as important), so approach with caution.

Onwards and upwards!

As I mentioned above, it's currently oriented towards Hungarian, but the goal is to make it language agnostic as I use it to help me learn the grammar of more languages. It would be cool to be able to take the Language object and generate an interactive website from it.

Language feature support

• Orthography: minimal
• Words: minimal
• Inflections: probably just enough to frustrate

How to use it

This set up is taken directly from the tests.

Create a language object - this stores all of the information you generate

hungarian = new Language("hungarian");

Define an orthography - this is accessible by the word. Surround n-graphs (digraphs etc.) with parentheses.

 hungarian.orthography({
    "vowels": {
      "back": "aáoóuú",
      "front": {
        "rounded": "öőüű",
        "unrounded": "eéií"
      },
      "long": "áóúőűéí"
    },
    "consonants" : "bc(cs)d(dz)(dzs)fg(gy)hjkl(ly)mn(ny)prs(sz)t(ty)vz(zs)",
    "sibilants": "s(sz)z(dz)"
  });

Create some words - make sure to use the dictionary form (also known as the lemma). The "VERB" is the POS. The inflection scheme is chosen automatically based on the part of speech and matching rules defined in the .

hungarian.word("ért","VERB")
hungarian.word("tanít","VERB")

The most interesting bit is creating inflection rules. It takes a grammatical descriptor (e.g. "1sg" or "DAT") and the morphological form ("+Vk"). The morphological form is described by what replacements need to be done: in this case "V" is replaced by the vowel harmony type (index to the array) and the '+' means it's a suffix. You'll notice that there are some more verbose conditions. These are described with a (at the moment) reduced grammar:

• after means that what follows must appear at the end of the word
• orthographic lists are accessed with their path in quotes, e.g. 'vowels.back' will match any aáoóuú
• use or to have multiple conditions
• strings of words are denoted with an xN, where N is the number of times it should be repeated
• + is basic concatenation

For example, after 'consonants' x2 or 'vowels.long' + t expands to /([b|c|...|zs]{2}|[á|ó|...|í]t)$/

hungarian.inflection({
    "schema": ["back","front.unrounded","front.rounded"],
    "name": "VERB",
    "1sg": {
      "default": {
        "form": "+Vk",
        "replacements": {
          "V": ["o", "e", "ö"]
        }
      },
      "-ik": {
        "form": "+Vm",
        "replacements": {
          "V": ["o", "e", "ö"]
        }
      }
    },
    "2sg": {
      "default": {
        "form":"+sz",
        "replacements":{"V":["a","e","e"]}
      },
      "after 'consonants' x2 or 'vowels.long' + t": {
        "form":"+Vsz",
        "replacements":{"V":["a","e","e"]}
      },
      "after 'sibilants'": {
        "form":"+Vl",
        "replacements":{"V":["o","e","ö"]}
      }
    },
    "3sg": {
      "form": "+",
      "replacements": {}
    },
    "1pl": {
      "form": "+Vnk",
      "replacements": {
        "V": ["u", "ü", "ü"]
      }
    },
    "2pl": {
      "default":{
        "form": "+tVk",
        "replacements": {"V": ["o", "e", "ö"]}
      },
      "after 'consonants' x2 or 'vowels.long' + t": {
        "form": "+VtVk",
        "replacements": {"V": ["o", "e", "ö"]}
      }
    },
    "3pl": {
      "default": {
        "form": "+nVk",
        "replacements": {"V": ["a", "e", "e"]}
      },
      "after 'consonants' x2 or 'vowels.long' + t": {
        "form": "+VnVk",
        "replacements": {"V": ["a", "e", "e"]}
      }
    }
  });

A more advanced example, using the past tense marker.

Make sure you list the markers you want to use, the order they should be added to the root. In this case, the indefinite past tense requires just the 'PST' marker

 hungarian.inflection({
    "schema": ["back","front"],
    "name": "VERB-PST",
    "markers": ["PST"],
    ...

As it stands, the syntax for markers is decently simplified - or at least doesn't require as much typing as a full inflection. The conditions are written using the same restricted English as described for the inflections (Marker is a subclass of Inflection). You can also include exceptions, which are words that don't meet the general regex condition, but should still be marked a certain way. Within conditions, you can also have overrides based on number/person marking ("1sg" etc.).

  hungarian.marker({
    "schema": ["back", "front.unrounded", "front.rounded"],
    "name": "PST",
    "after 'consonants' x2 or 'vowels.long' + t": {
      "exceptions": ["fut","hat", "jut", "köt", "nyit", "süt", "üt", "vet"],
      "form": "+Vtt",
      "replacements": {"V": ["o","e","ö"]}
    },
    "after 'palatals' or +ad or +ed": {
      "exceptions": ["áll","száll","varr","forr"],
      "form": "+t",
      "replacements":{}
    },
    "default": {
      "exceptions": ["lát", "küld", "mond", "keyd", "függ", "fedd"],
      "overrides": {"3sg" :{
          "form": "+Vtt",
          "replacements": {"V": ["o","e","ö"]}
        }
      },
      "form": "+t",
      "replacements": {}
    }
  })

Testing

Here there be tests. Mocha and Chai that sholdier boy.