My understanding was that at least some sub-vocalisation is a normal part of reading and writing. The wikipedia article on subvocalisation cites several sources supporting that claim. The article also claims that there is no evidence to suggest that speed reading training that involves suppression of sub-vocalisation is effective. There is also evidence that suppressing subvocalisation can in some instances result in decrements in performance (see Baddeley et al, 1981).
There is some evidence to suggest that degree of subvocalisation is related to complexity of the reading passage. Baddeley et al 1981 summarise the use of
the electromyographic monitoring of the muscles involved in
vocalisation in order to study the role of articulation in reading
(e.g. Sokolov, 1966). Using this technique, Hardyk and
Petrinovitch (1970) concluded that subjects were able to read and
comprehend simple prose without the occurrence of
electromyographic activity, but that such activity did occur in
the case of more complex passages.
One explanation would be that degree, prominence or frequency of subvocalisation increases as the reading task gets more difficult, where experienced difficulty would vary both based on the individual's ability with the language and with the objective difficulty of the reading passage.
References
- Baddeley, A., Marge Eldridge & Vivien Lewis (1981): The role of
subvocalisation in reading, The Quarterly Journal of Experimental Psychology Section A: Human Experimental Psychology, 33:4, 439-454
- HARDYK, C. D. and PETRINOVITCH, L. R. (1970). Subvocal speech and comprehension level as a function of the difficulty level of reading material. Journal of Verbal Learning and Verbal Behavior, 9, 647-52.
- SOKOLOV, A. (1966). La parole intbrieure dans la pensbe concrete. In Recherches Psychologiques en U.R.S.S., les Editions du Progrks, U.R.S.S.