As a kind of dialect dictionaries, a monolingual dictionary of Cantonese contains a wide range of colloquial vocabulary. Lexicographers tend to assemble a large body of dialect words that occurs in relation to the head-character for one entry in the dictionary. As for reflecting the various pronunciations for one character, different pronunciations of lexical items are often presented in parallel columns. However, in the event of colloquial-vs.-literary readings, all the parts of the entry would be mostly placed under the combination of several different pronunciations. As a result, dictionary-users might think that colloquial pronunciation is used in oral communication while literary one is used in formal expressions. In fact, a single character could be pronounced using both literary and colloquial pronunciations, whereas words (character combinations) might be read as either the literary pronunciation or the colloquial one. Hence, in most monolingual dictionaries of Cantonese published in the past, dictionary-users have great difficulty deciding whether to use the literary or colloquial pronunciation when reading words out. Sound changes in expression are initiated by conforming to ancient sounds and/or language contact. These causes leading to the phenomenon of literary and colloquial readings have been studied by many scholars. They agreed that without context, it is hard to determine whether a character should be pronounced in the literary or colloquial way. However, when the Chinese characters are combined to form lexis, it is easier to determine its pronuncation. Such expression in speech is understood and adopted by numerous people in Cantonese-speaking areas. The dictionary, Guangzhouhua Zhengyin Zidian (Dictionary of Standard Cantonese Pronunciation), is employed as a case study to explore these issues of literary and colloquial readings of lexicography. This paper attempts to seek a way for a clear policy to indicate accurately the literary-vs.-colloquial alternation by using suitable indicative labels. Moreover, recommendations for the case-based compilation are given in the discussion.

作為方言字典的一種，廣州話單語字典收錄不少口語詞。編纂者往往為每個字目組織了大量與字頭有關的方言詞。為表示某個字有多個字音，音項每每分立以表示該字在構詞上的不同發音。 然而，字典對“文白異讀”音卻多同時並列。人們只認識到讀書時用文讀音，口頭語則用白讀音。事實上，字音可有文、白讀音之分，唯詞音可能只有單種讀法。若依照過去單語字典這種編排，讀者們是難以決定某詞之正確發音的。 因為要迎合古音、或因語言接觸而導致了語音的改變，學者們早有論述這種產生“文白異讀”的現象。他們的研究顯示了在字的層面上文讀音與白讀音的分工並不明確，但在詞的層面上，這種不明確性就大大減少了。方言地區的人們在使用習慣上正反映著這種差別。 本文以《廣州話正音字典》作為例子揭示在字典編纂上有關“文白異讀”的問題，於解決方法上提出了用標示號指明單讀音的建議，並對該字典於“文白異讀”有關的編纂做法給予意見。