European word translator: an interactive map

Original Article ↗ Hacker News Discussion ↗

Overall reception

  • Many find the map delightful and novel: a clever way to visualize vocabulary across Europe and a fun exploratory tool.
  • Users enjoy trying words like “pineapple,” “turtle,” “girl,” “tea,” etc., to see unexpected patterns and oddities.

Data source, accuracy, and limitations

  • Translations are from Google Translate circa 2014 and are pre-cached, not live; users note machine translation has improved since.
  • Numerous errors are reported:
    • “She runs” is wrong or shifted in several languages (e.g., “she functions,” “she walks”).
    • “An example” is mistranslated in Finnish and Hungarian; Italian “the example” is also off.
    • “Cross,” “lead,” and “folk” show inconsistent meanings across languages (noun vs verb, religious symbol vs verb, etc.).
  • The dictionary is very limited: many common phrases, month names, country names, and “small {noun}” combinations return nothing.
  • Users observe Google’s tendency to pick arbitrary cases or genders for Slavic and gendered languages, especially when pivoting through English.

Word meaning, polysemy, and design issues

  • Multiple commenters argue that 1–2 word mappings are inherently misleading because:
    • Words are polysemous and don’t map 1:1 across languages.
    • The site usually shows only one sense and one form, ignoring other valid options and grammatical distinctions.
  • Suggestions:
    • List all major translations or at least match part of speech (noun vs verb).
    • Allow phrases/sentences with highlighted key words.
    • Use newer translation tech (LLMs) which seem better at context and sense selection.

Dialects, variants, and coverage

  • Tool mostly treats each state as a single language, ignoring strong internal variation:
    • German (Germany/Austria/Switzerland), Dutch/Belgian Dutch, Italian regions, and Spanish regional languages are cited.
    • Swiss German and other Alemannic varieties lack a standard written form, complicating support.
  • Examples of fine-grained dialect maps (German, Swiss German, Chinese) are shared; people wish this tool had zoomable dialect-level detail.

Pronunciation and phonetics

  • For Cyrillic- and Greek-based languages, users want Latin transliteration; phonetic transcription is especially requested for Ukrainian/Russian.
  • Several propose coloring by phonetic or string similarity (e.g., Levenshtein distance) instead of fixed country colors.

Etymology, number systems, and fun observations

  • Users discuss:
    • Ancient core vocabulary (“iron,” “stone,” “cow,” “sun,” “salt,” “tea,” “mama/papa”) and what it reveals about language families and trade routes.
    • The “tea/cha” split and exceptions like Portuguese and Russian-influenced Polish words.
    • Humorous compound words: “shelled frog”/“shield-toad” for turtle, “spike-pig”/porcupine, “flying mouse” for bat, Japanese “tree child” for mushroom.
    • Number systems: French base-20 forms, Belgian/Swiss septante–huitante–nonante, German/Dutch reversed order, and speculation (mostly dismissed) about effects on math ability.