The contemporary coexistence of fronted and plosive reflexes of ğīm excites much scholarly interest: does the voiced velar stop [g] reflex prevalent in some urban areas of Egypt constitute a continuation of the Proto-Semitic phoneme /g/; or is it a recent phenomenon, the result of a process of ‘de-affrication’ which stabilized in Cairo and its surrounding provinces as late as 1800–60 CE? Following the influential work of Blanc (1969, 1981) and Hary (1996a), the latter interpretation has become the established consensus among scholars of Judaeo-Arabic. In relation to Judaeo-Arabic, Blanc's thesis relies on (i) the use of the diacritic with gimel denoting ğīm and (ii) occurrences of assimilation, metathesis and substitute graphemic representations of ğīm. In questioning the assumption of the diacritic's phonetic significance, demonstrating its unreliability as a source from which to reconstruct the historical phonetic reflexes of ğīm and re-examining oft-cited instances of assimilation, metathesis and graphemic substitutions in light of new evidence, this paper establishes that Judaeo-Arabic orthography makes a much more limited, albeit valuable, contribution to the field of historical Arabic linguistics in this regard than has previously been suggested.