>The Lossy Codec

"I-Ching" is not the title — it's a lossy re-encoding of a sound the original system never made.

When the core text of the 易經 was being compiled — roughly 800 to 200 BCE, across the Western Zhou and the Warring States — the two characters were pronounced something close to *lek k-lˤeŋ. That is the reconstruction from William Baxter and Laurent Sagart, the current standard in historical Chinese linguistics. Zhengzhang Shangfang's independent reconstruction lands in the same neighborhood: *leɡs keːŋ.

Say it out loud and nothing modern comes back. It is not "ee-ching." It is not "yee-jing." The first glyph, 易, opened with an *l- and closed on a hard -k stop. The second, 經, fired a velar k- at the front of the syllable. Mandarin has since dropped both features. The name everyone reads on the spine is a re-render of a re-render — the audio passed through a codec that quietly threw away the high-frequency data.

The compression had a cause, and it was geographic. Mandarin is not simply "Chinese" — it is specifically northern Chinese, and northern pronunciation was reshaped by centuries of contact with the languages of the steppe: Jurchen, Mongol, Manchu. The linguist Mantaro Hashimoto named the process the "Altaicization of Northern Chinese."

The corruption was systematic, not random. Northern varieties, pressed up against the toneless Altaic languages, shed tonal complexity and dropped their final stop consonants — the -p, -t, and -k endings Old and Middle Chinese ran on everywhere. That whole class of syllables, the entering tone (入聲, rùshēng), simply vanished from the northern register. The southern varieties — Cantonese, Hokkien, Hakka — sat outside the contact zone and kept the archaic bits intact. So Cantonese yik-ging (易經) still carries the -k stop on 易 and the velar k/g- on 經. Mandarin yì-jīng has lost both. The court of Zhou would have heard something nearer the south.

There is no need to reconstruct anything to prove the lost sounds were real. They were backed up — exported to neighboring systems and write-protected. When Chinese characters entered Korea and Japan, mostly across the Tang and Song dynasties, they carried that era's pronunciation with them and then froze in place while the mainland kept overwriting itself. The Sino-Korean and Sino-Japanese readings are fossils: snapshots of how Chinese sounded a thousand years ago.

易經 in Japanese is Eki-kyō (えききょう). The "eki" preserves the final -k Mandarin dropped; the "kyō" keeps the velar k- Mandarin palatalized into j-. In Korean it is Yeok-gyeong (역경) — same -k ending on 역, same g- initial on 경. The pattern repeats across the simplest tokens in the set, the numbers:

Every column except Mandarin holds a final consonant Middle Chinese had. Korean and Japanese do not "sound like Cantonese" — they each independently preserve the older Chinese that the north overwrote. Cantonese is just the living dialect nearest to what the loanwords froze in place.

So it is no surprise the English-speaking world cannot agree on the filename. The major encoders:

• I-Ching — Wade-Giles (1859), built by Thomas Wade and refined by Herbert Giles. The English standard for over a century. Encodes Mandarin.
• Yi Jing / Yìjīng — Hanyu Pinyin (1958), the PRC's official system, now the academic default. Also encodes Mandarin.
• Yi-King — the EFEO transcription used by early French sinologists working with southern speakers. The quiet irony: their spelling sits closer to the ancient sound than either of the standards.

There are more — Yale, Gwoyeu Romatzyh, a scatter of missionary spellings that never stabilized — each trading phonetic accuracy against learnability and politics. None of them recovers the original audio. They are all transcriptions of an already-compressed signal.

The compression was eventually ratified by committee. In 1913 the new Republic of China convened the Commission on the Unification of Pronunciation (讀音統一會) in Beijing — forty-four delegates from twenty-six provinces, tasked with fixing a national reading for some 6,500 characters. The northern Mandarin faction under Wang Zhao (王照) had the votes; more provinces spoke Mandarin variants. The southern delegates fought for the entering tone and the -n / -ng distinctions their dialects had kept from Middle Chinese.

It broke down physically. Wang Rongbao (王榮寶), leading the south, used a Shanghai colloquialism that Wang Zhao misheard as a Mandarin curse — and Wang Zhao attacked him and chased him from the hall. The commission shipped a compromise, the 老國音 ("old national pronunciation"), with some entering-tone distinctions left in. It satisfied no one. By 1932 it was quietly replaced by the 新國音 — essentially Beijing Mandarin. The entering tone was formally dropped. The sounds the south and the fossils had preserved were declared irrelevant to the standard.

The romanization is the door, not the room. "I-Ching" is the cached read; the original signal is older than any way of spelling it.