Whenever I observe my fellow Brazilian countrymen learning to speak English, a clear sound change pattern stands out:
[θ] → [f] [ð] → [d], syllable-initial [f], syllable-final
So, for example, think is usually pronounced [fĩk], then is pronounced [den] and with, [wif].
Since Brazilian Portuguese does not have dental fricatives, this kind of sound change is much to be expected. What is interesting about it is the regularity of the pattern. Among other "likely" candidates, such as [s] or [t], which are also sounds found in BP, [f] is — almost always — chosen to replace [θ]. Similarly with [ð], that could as well be replaced with [z] or [v].
Even those who have never heard other people speak that way will make this choice. So, my question is: can this sound change be predicted from Brazilian Portuguese phonology? What else could explain its regularity? In general, if a phonologist is presented with two languages, A and B, can they predict what sound changes are most likely to occur if the native speakers of A start speaking B as a second language?