A Writing System Within a Writing System
When Hong Kongers type on their phones or read a tabloid, they encounter characters that no Mandarin dictionary contains. These are the 固有字 (guyouzi) of Cantonese — Han characters invented specifically to write Cantonese grammar and vocabulary that have no equivalent in Mandarin.
The five workhorses are:
- 嘅 ge3 — genitive/attributive particle (Mandarin uses 的 de)
- 喺 hai2 — locative copula 'to be at' (Mandarin: 在 zài)
- 唔 m4 — negation prefix (Mandarin: 不 bù / 沒 méi)
- 咁 gam3 — 'so / such / like this' (Mandarin: 這樣 zhèyàng)
- 啲 di1 — plural/partitive 'some / a bit' (Mandarin: 一些 yīxiē)
These characters are not typos or slang abbreviations; they are fully codified in the Unicode Han Unified repertoire and appear in Hong Kong print every day. Apple Daily and Oriental Daily carried 粵語白話文 (Cantonese vernacular writing) for decades in their columns, entertainment and gossip pages and racing sections — the news pages themselves stayed in Standard Written Chinese, with Cantonese surfacing there mainly in headlines and quoted speech — and Facebook posts from Hong Kongers overwhelmingly use them.
The mainland PRC, however, leaves most of them out of its standard character lists: 嘅, 喺, 咁 and 啲 are all absent from the 通用规范汉字表 (2013), and entering them in many mainland input methods returns a blank. 唔 is the exception — it sits at no. 4500 of that list's second-level table, but as a Mandarin interjection ('hmm'), not as the negator Cantonese uses it for. This reflects a broader political and linguistic tension: writing Cantonese in Cantonese challenges the implicit assumption that Mandarin is the universal form of Chinese.
Cantonese also preserves the final stops -p, -t and -k that Mandarin lost, and it split its tones into upper and lower registers when the Middle Chinese voiced initials devoiced — part of why even shared vocabulary is unrecognisable across the two. 嘅 likely derives from 個 (gè) narrowed semantically; 唔 is attested in Song-dynasty colloquial texts. These are not modern inventions but centuries-old oral forms finally given their own graphic home.