According to the wikipedia entry of Mutatis mutandis, "Mutatis mutandis is a Medieval Latin phrase meaning 'with things changed that should be changed' or 'having changed what needs to be changed' or 'once the necessary changes have been made'."
In Classical Latin, the subject of the Ablative Absolute clause mutatis mutandis would be the nominalized verbal adjective/gerundive mutandis 'the things to be changed' and its predicate would be the perfect passive participle mutatis '(having been) changed'. One could wonder if a second (though less natural) reading of this Ablative Absolute could also be possible, a reading where mutatis would be the subject ('the things that were changed') and mutandis would be the predicate ('having to be changed'). From what I discussed in a related post this second reading is not expected to be possible in Classical Latin but perhaps it could be possible in Late Latin and/or in Medieval Latin. For related discussion see also this post. It seems that Joonas thinks that this Ablative Absolute construction is potentially ambiguous. According to him, the two readings are: (1) "having changed what had to be changed" (i.e., the conventional reading) and (2) "by changing what had been changed" (in my view, a controversial reading, at least in Classical Latin). He says "Both readings are reasonable". I don't think both readings are possible in Classical Latin. Only the first one (the conventional one) is possible. But what about in Medieval Latin? I was wondering if at this stage the gerundive could function as the predicate of an Ablative Absolute construction.
NB: To avoid possible confusions, let me clarify that I'm saying that mutatis mutandis cannot be interpreted as an Ablative Absolute (AA) construction (in Classical Latin) iff mutandis is the predicate of the AA. The predicative reading of mutandis is indeed possible in a sort of (syntactically lower) dominant-like gerundive construction like mutatis mutandis tribunatum gessit ('he spent the tribunate in changing what had been changed'. In syntactic terms, the AA is a sentential adjunct, whereas the dominant-like gerundive construction is part of the Verbal Phrase (VP). In the AA reading of mutatis mutandis in an example like mutatis mutandis tribunatum gessit only mutatis can be the predicate of the AA (in Classical Latin): 'having (been) changed what needed to be changed, he spent the tribunate'. Note that an obvious prediction of my analysis above is that both constituents (AA and predicative gerundival constructions) could coincide, i.e., something like urbe capta, patribus insectandis tribunatum gessit is perfect. Perhaps some of you will insist in labeling ("analyzing") both urbe capta and patribus insectandis as two examples of Ablative Absolutes on the basis that both involve a clausal subject-predicate relation. Fine but remember that they are two different syntactic types of clausal constituents. After all, Grammar is not just a matter of putting labels/taxonomy but is a matter of understanding!