This doesn't explain how the mapping from Latin characters was originally determined. I thought it might just be the result of a font for a legacy character enc

This doesn't explain how the mapping from Latin characters was originally determined. I thought it might just be the result of a font for a legacy character encoding being used with ASCII data, but that doesn't seem to be the case. Of all the legacy encodings (shift-jis, gb18030, big5 etc.) my Python installation supports, the only encoding that produces CJK for bytes < 128 is UTF-16: import encodings import unicodedata def try_default(default, f, *args, **kwargs): try: return f(*args, **kwargs) except: return default b = bytes(range(128)) for encoding in set(encodings.aliases.aliases.values()): s = try_default('', b.decode, encoding) if any('CJK' in try_default('', unicodedata.name, c) for c in s): print(encoding, ''.join(c for c in s if c.isprintable())) prints utf_16_be ȃЅ؇ࠉฏထሓᐕᘗ᠙ᨛᰝḟ‡∣␥⠩⨫Ⱝⸯ〱㈳㐵㘷㠹㨻㰽㸿䁁䉃䑅䙇䡉䩋䱍乏偑剓呕噗塙婛屝幟恡扣摥晧桩橫汭湯灱牳瑵癷硹穻籽繿 utf_16_le Ā̂Ԅ܆ईଊഌ༎ᄐጒᔔᤘᬚᴜ℠⌢┤⤨⬪⼮㌲㔴㜶㤸㬺㴼㼾䅀䍂䕄䝆䥈䭊䵌低児卒啔坖奘孚嵜彞慠换敤杦楨歪浬潮煰獲畴睶祸筺絼罾 utf_16 Ā̂Ԅ܆ईଊഌ༎ᄐጒᔔᤘᬚᴜ℠⌢┤⤨⬪⼮㌲㔴㜶㤸㬺㴼㼾䅀䍂䕄䝆䥈䭊䵌低児卒啔坖奘孚嵜彞慠换敤杦楨歪浬潮煰獲畴睶祸筺絼罾 Those probably have too many strokes to look good in a tattoo.