As soon as a speaker becomes bilingual, the novel language will very subtly affect the native one, even if it is not much employed. This is how first language attrition may begin. Most attrition research handles spoken language; however, as in writing people have more time to think and revise, written data is presumed to reflect language skill more clearly. This inquiry explores immigrants’ first language attrition in written texts. Data was a corpus of 71,848 words taken from the weblogs of a number of Persian speaking immigrants to Australia that were written monthly up to the fiftieth month of their immigration. It included lexical, morphological, syntactic and semantic deviations of the native language of these immigrants. Their own written production two months before immigration is employed as control. The result validates the vulnerability of lexicon, as in former research. There is a decreasing, and finally stabilized state in some cases of attrition; this may be accounted for by the acculturation phenomenon. Lots of interpersonal variations have been detected which may either be gender-related or a matter of personal preference. Hence, this inquiry calls for further psycho-sociolinguistic investigations to explain the language attrition.