Two phonological processes (?) that are widely used in Indo-Atyan philology, viz, Compensatory Lengthening (CL) and deaspiration (D) are problematic concepts at least in the context of describing the ‘development’ of Bangla language. In the case of CL, the change of [a] >[A] has nothing to do with compensation, rather, it is a substitution of one sound by another. On the other hand, the change of [dh] > [d] is a case of D, but mh> hm > mm/m is not a case of D. It is, of course, a case of h-deletion. Author pointed out that the earlier interpretations of such changes are due to the overemphasis on the writing system and are guided by the memory of cultural system of writing.
These fallacies were initiated by S.K. Chatterjee and widely followed and taught in linguistics departments in India where Indo-Aryan philology is practiced following the age-old neo-grammarians’ tradition. Consider the following examples of change, generally known as CD: janma >jamma >jAm “birth”; agni > aggi >Ag “fire”, madhya> majjha> mAjh “middle”.
All these changes are a case substitution of [a] by [A]. The interpretation of [A] as a lengthened form of [a] is the result of a cultural convention of writing. When Bengali children are taught the Bangla script, they learn that /a/(low back vowel ) -e akare A ( low central vowel ), i.e. , as per visual presentation in writing, if you put secondary graphemic symbol of A beside the grapheme a, you will get the grapheme A. However, in the case of this type of substitutional sound change, there is no phonetic or phonological evidence of lengthening per se.
Let us consider the following examples for the case of so-called D: Chatterjee (1926:441, 1950:160-78) and Sen (1987:183) described the middle Bangla changes like krsna > kanha > kAnay, brahma > bramha > bramma as cases of D like dh > d. However, phonetically speaking, these are not at all a case of deaspiration. These cases show clearly that the h is deleted and there is no question of loss of aspiration. This description of D is guided by an MIA sound change in which one gets dh> h-like changes. Here, dh is substituted by h. I think the memory of Roman script with diagraph leads to such non-phonetic description of sounds of Bangla.