0.1.89 • Public • Published


@zsnout/ithkuil is a package containing several utilities for working with New Ithkuil text. It can generate romanized text from JSON objects, parse text into JSON objects, and write in Ithkuil's writing system using SVG paths.


This package is in active development, so expect change. That said, this package is completely suitable for use, and has even been tested somewhat.


@zsnout/ithkuil can generate (as of May 29, 2023), parse (as of July 6, 2023), create block script for (as of August 1, 2023), and gloss (as of August 25, 2023) every kind of word in Ithkuil, including:

  • Formatives
  • Specialized Cs-root formatives
  • Specialized personal-reference formatives
  • Single- and dual- referentials
  • Combination referentials
  • Affixual adjuncts
  • Bias adjuncts
  • Modular adjuncts
  • Parsing adjuncts
  • Register adjuncts
  • Suppletive adjuncts
  • Referentials with suppletive adjuncts as heads

It also has many, many more functions. It can:

  • Map the values of various categories to their names
  • Provide descriptions for certain forms
  • Check whether a consonant form is phonotactically legal
  • Insert glottal stops into vowel forms
  • Compute Ca forms and their geminated counterparts
  • Validate incoming objects containing Ithkuilic word data

@zsnout/ithkuil has many more functions. However, these are too numerous to place into a single document. To experiment with all of @zsnout/ithkuil's functionality, install it and use your code editor's tools to see the available functions it exports.

Important Note

Unlike most packages available online, @zsnout/ithkuil does not have a top-level export. That is, directly importing from @zsnout/ithkuil will fail.

import { parseWord } from "@zsnout/ithkuil"

Instead, always make sure to import from the proper sub-package. Currently, there are four sub-packages:

  • @zsnout/ithkuil/data, which has JSON information about all roots and affixes.
  • @zsnout/ithkuil/generate, which generates romanized text.
  • @zsnout/ithkuil/gloss, which turns Ithkuil words from the JSON format into gloss strings.
  • @zsnout/ithkuil/parse, which parses romanized text.
  • @zsnout/ithkuil/script, which generates SVG block script.
  • @zsnout/ithkuil/zod, which provides Zod validators for this project.
import { parseWord } from "@zsnout/ithkuil/parse"


If you're generating block script using the script module, you may find it helpful to use our mini JSX library to make things easier to work with. It's completely optional, but it makes writing code much simpler. Here's an example without JSX.

import {
} from "@zsnout/ithkuil/script"

function displayText(text: string) {
  const svg = document.createElementNS("http://www.w3.org/2000/svg", "svg")

  const result = textToScript(text)

  if (result.ok) {
    const row = CharacterRow({
      children: result.value,
      compact: true,

    const anchored = Anchor({
      at: "cc",
      children: row,

  } else {
    const text = document.createElementNS("http://www.w3.org/2000/svg", "text")

    text.textContent = result.reason




displayText("Wattunkí ruyün")

And here's the same example with JSX:

import {
} from "@zsnout/ithkuil/script"

function displayText(text: string) {
  const svg = (
        ok={(value) => (
          <Anchor at="cc">
            <CharacterRow compact>{value}</CharacterRow>
        error={(reason) => <text>{reason}</text>}
  ) as SVGSVGElement



displayText("Wattunkí ruyün")

Before setting up JSX, make sure you're in a TypeScript project.

If you don't already have a JSX transform, add these lines to your tsconfig's compilerOptions to enable JSX:

  "compilerOptions": {
    "jsx": "react-jsx",
    "jsxImportSource": "@zsnout/ithkuil/script"

If you're already using a JSX transform, write this at the top of any files you want to use @zsnout/ithkuil's JSX in.

/* @jsx react-jsx */
/* @jsxRuntime automatic */
/* @jsxImportSource @zsnout/ithkuil/script */

That's it!

Note that the JSX runtime is intentionally very minimal. There are no event listeners, no signals, no hooks, very few type definitions, and it only generates SVG elements.

Example 1

This example compiles a simple formative.

import { formativeToIthkuil } from "@zsnout/ithkuil/generate"

const result = formativeToIthkuil({
  root: "kš",
  type: "UNF/C",

// kšala

Notice how the function formativeToIthkuil is able to infer default values for most of its arguments.

Example 2

Just being able to convert formative roots into Ithkuil isn't very exciting. Let's try an example with many options.

import { formativeToIthkuil } from "@zsnout/ithkuil/generate"

const result = formativeToIthkuil({
  type: "UNF/C",

  // Type-2 concatenation
  concatenationType: 2,

  // Completive Version
  version: "CPT",

  // Stem II
  stem: 2,

  // "kš" root
  root: "kš",

  // Dynamic Function
  function: "DYN",

  // Objective Specification
  specification: "OBJ",

  // Amalgamative Context
  context: "AMG",

  slotVAffixes: [
    // Referential Affix
      // 1m:BEN Referent
      referents: ["1m:BEN"],

      // Ergative Case
      case: "ERG",

  ca: {
    // Multiplex/Fuzzy/Connected Configuration
    configuration: "MFC",

    // Coalescent Affiliation
    affiliation: "COA",

    // Graduative Extension
    extension: "GRA",

  // Repetitive Phase
  vn: "REP",

// hwikšöeroeržžgeiha

Notice that in the above example, we didn't have to say whether the Vn slot was a Valence, Phase, etc. We simply specified "REP" as the Vn slot and @zsnout/ithkuil figured out the appropriate category automatically.

In addition, we only specified part of the Ca complex. We left out Monadic Perspective and Normal Essence, but @zsnout/ithkuil inferred those.

Example 3

Not only can @zsnout/ithkuil generate formatives, but it can also parse them using the parseFormative function. Let's try it out.

import { parseFormative } from "@zsnout/ithkuil/parse"

const result = parseFormative("malëuţřait")


// {
//   type: "UNF/C",
//   concatenationType: undefined,
//   shortcut: false,
//   stem: 1,
//   version: "PRC",
//   root: "m",
//   context: "EXS",
//   specification: "BSC",
//   function: "STA",
//   slotVAffixes: [],
//   ca: {},
//   slotVIIAffixes: [
//     { type: 2, degree: 5, cs: "ţř" },
//     { type: 2, degree: 1, cs: "t" },
//   ],
//   mood: undefined,
//   caseScope: undefined,
//   vn: undefined,
//   case: undefined,
//   illocutionValidation: undefined,
// }

As you can see, parseFormative quickly turns formative strings into programmatically analyzable objects.

Note that concatenated formatives must be parsed separately from their parents, as shown below with the example "hlarrau-laza" (a tribe of people owned by a cat).

import { parseFormative } from "@zsnout/ithkuil/parse"

// Correct:

const result1 = parseFormative("hlarrau")
const result2 = parseFormative("laza")

console.log(result1, result2) // { type: "UNF/C", ... }, { type: "UNF/C", ... }

// Incorrect:

const result = parseFormative("hlarrau-laza")

console.log(result) // undefined

Note that parseFormative can output three types of values.

  1. If the formative doesn't have a valid slot structure (e.g. an invalid number of C and V forms), undefined is returned.

  2. If the formative has a valid slot structure but its slots are filled with invalid values (such as üö in the Vc slot), an error is thrown.

  3. If the formative is a valid formative, the parsed formative is returned.

Example 4

@zsnout/ithkuil doesn't just handle formatives; it can work with all types of Ithkuilic words. Let's try generating an affixual adjunct, as those are used frequently to move affixes out of formatives.

import { affixualAdjunctToIthkuil } from "@zsnout/ithkuil/generate"

const result = affixualAdjunctToIthkuil({
  affixes: [
      type: 1,
      degree: 2,
      cs: "c",
  scope: "VII:DOM",

// äce

Example 5

This example creates a referential (1m.BEN-CTE-GID₁/3).

import { referentialToIthkuil } from "@zsnout/ithkuil/generate"

const result = referentialToIthkuil({
  referents: ["1m:BEN"],
  specification: "CTE",
  affixes: [
      type: 1,
      degree: 3,
      cs: "c",

// raxtec

Example 6

This example shows the glossing capabilities of @zsnout/ithkuil.

import { glossWord } from "@zsnout/ithkuil/gloss/index.js"
import { parseWord } from "@zsnout/ithkuil/parse/index.js"

const result2 = parseWord("wetace")

if (result2) {
  const { short, full } = glossWord(result2)

  // S2-‘that one’-‘female’₁-ABS

  // stem_two-‘that one’-‘female’₁-absolutive

As you can see, @zsnout/ithkuil provides simple glossing functionality for Ithkuil words.

Example 7

Every data form has a corresponding Zod parser for it, which allows for quick and easy validation of it.

This example validates that an object is, in fact, an Ithkuilic adjunct. This can be useful when integrating this project with external sources.

import { adjunctToIthkuil } from "@zsnout/ithkuil/generate"
import { adjunct } from "@zsnout/ithkuil/zod"

const myAdjunct = getAdjunctFromSomeInternetSource()

try {
  const realAdjunct = adjunct.parse(myAdjunct)

  const result = adjunctToIthkuil(realAdjunct)

} catch (error) {
  console.error("The adjunct was malformed.", { cause: error })



  • Breaking change: Multiple referents may now be stored in a referential affix. As such, all referential affixes now use the referents property instead of the singular referent property, and take a ReferentList instead of a single Referent.

  • Breaking change: parseReferentialAffixCs has been merged into parseReferentListAndPerspective. As such, parseReferentListAndPerspective now takes a boolean argument indicating whether it is parsing a referential affix.

  • Breaking change: The vnToAffix, toAffix, and mergeAdjunctsAndFormative functions have been moved from @zsnout/ithkuil/script to @zsnout/ithkuil/generate to allow for them to be imported without needing an SVG document in place.


  • Fixed issues with examples.


  • Massive breaking change: @zsnout/ithkuil/generator's exports are now only available under @zsnout/ithkuil/generate. They are no longer exported from @zsnout/ithkuil.

  • Massive breaking change: @zsnout/ithkuil/parser's exports are now only available under @zsnout/ithkuil/parse. They are no longer exported from @zsnout/ithkuil.

  • Breaking change: The DFT level has been correctly updated to DFC.

  • New feature: The @zsnout/ithkuil/script sub-package has been created, and may be used to generate Ithkuil script as SVG paths.


  • Fixed punctuation in README.


  • Output change and bug fix: Referentials that have multiple referents are now parsed correctly.


  • Breaking change: The Referent type, the ALL_REFERENTS array, and the REFERENT_TO_ITHKUIL_MAP and REFERENT_TO_REFERENT_OBJECT_MAP objects have all been reordered. Where their unions, elements, and keys were previously ordered 1m:NEU, 2m:NEU, 2p:NEU, ..., 1m:BEN, 2m:BEN, ..., 1m:DET, ..., they are now ordered 1m:NEU, 1m:BEN, 1m:DET, 2m:NEU, 2m:BEN, etc.

  • Breaking change: Referentials can now take multiple referents in their second referent slot. As such, the referent2 property of referentials has now been pluralized and changed to the referents2 property.

  • Breaking type-level change: The CN type union now includes MoodOrCaseScope instances. As such, it no longer only contains string literals.

  • Output change: Passing differently ordered lists to referentListToIthkuil now gives the same output.

  • Added MoodOrCaseScope class to represent Cn forms in Modular Adjuncts.

  • Added functions to lex and parse adjuncts

  • Added parseWord to parse formatives, referentials, and adjuncts all at once

  • Added wordToIthkuil to turn generic words into Ithkuil

  • Added word Zod parser to validate generic JSON objects representing words

  • Fix several issues involving the parsing of formatives with specialized Cs-roots and specialized personal-reference roots

  • Allow multiple referents in C2 of dual referentials

  • Added SingleRegisterAdjunct to indicate individual registers, such as DSV:START, CGT:END, and so on. Note that SingleRegisterAdjunct does not include NRR:START, NRR:END, or END:START.


  • Output change: Formatives with specialized personal-reference roots now properly take a+Ca shortcuts

  • Bug fix and output change: Formatives with specialized Cs-roots or personal-reference roots whose final Slot VII affix is NEG/4, DCD/4, or DCD/5 keep their proper Vv values instead of having them replaced with Slot VII affix shortcut values

  • parseFormative now successfully parses formatives with specialized Cs-roots and specialized personal-reference roots


  • Breaking change: Changed WithWYAlternative.precededByW to WithWYAlternative.valueAfterW

  • Breaking change: Changed WithWYAlternative.precededByY to WithWYAlternative.valueAfterY

  • Breaking change: Renamed export ALL_DIPTIONGS to ALL_DIPTHONGS

  • Breaking change: Zod validators are now only exported from @zsnout/ithkuil/zod under the direct names aspect, caseScope, partialFormative, etc., and are no longer available from the core @zsnout/ithkuil.

  • Output change: All generated words now end in a vowel form, fixing issues with External Juncture.

  • Output change and bug fix: Vv forms are now properly generated. Previously, a Stem 3 PRC word may have outputted a Stem 0 CPT Vv form, and vice versa.

  • Bug fix: Moved Zod validators into a separate file. This fixes an issue where imports were resolved in an order that caused many Zod validators to be undefined or in a TDZ when they were expected to have values.

  • Added STANDARD_VOWEL_TABLE, an array of vowel forms indexed as would be appropriate for humans. Where the previous tables had unexpected behavior (ZERO_INDEXED_VOWEL_TABLE[2][9] = degree 0 of sequence 3), the new table behaves exactly as would be expected (STANDARD_VOWEL_TABLE[3][0] = degree 0 of sequence 3). The old tables ZERO_INDEXED_VOWEL_TABLE and ONE_INDEXED_VOWEL_TABLE will remain for performance- and TypeScript-related reasons.

  • Use STANDARD_VOWEL_TABLE instead of ONE_INDEXED_VOWEL_TABLE in generator/affix/index.ts.

  • Add deepFreezeAndNullPrototype helper

  • Add insertGlottalStopIntoPossiblyWithWYAlternative helper

  • Change formatting to insert commas at the end of multiline function calls

  • Add regular expression builder

  • Add basic regular expression forms such as V, C, H, CG, CNG, VG, and VNG

  • Add regular expression for non-shortcut formatives

  • Add parsers for affixes, Ca forms, case scopes, cases, formatives, illocutions and validations, moods, referential affixes, and Vn forms

  • Add transformWord

  • Add VowelForm for systematic parsing and stringification of vowel forms

  • Add parseFormative to parse formatives

  • Add Slot VIII Mood/Case-Scope shortcut generation


  • Breaking change: Fixed spelling mistake (concatenatenationType --> concatenationType)


  • Added case-accessor and inverse case-accessor affixes


  • Added an array constant that contains all cases arranged in order and that properly skips the 8th degree of the last four case categories. This new array, ALL_CASES_SKIPPING_DEGREE_8, has a length of 72 and contains four empty cells: one before each of the RLT, VOC, NAV, and PLM cases.

  • Fixed Ca gemination algorithm (geminates involving N/RPV and A/RPV are complicated, and many Ca forms simply cannot be geminated according to Ithkuil's rules)

Package Sidebar


npm i @zsnout/ithkuil

Weekly Downloads






Unpacked Size

4.15 MB

Total Files


Last publish


  • zsakowitz