The three posts ( one , two , three ) below form a set of posts on the topic of serendipity in languages. Please do send me a note if you have thoughts or comments on this.
I really do prefer to use the terms 'pseudoreduplication'/'pseudoreduplicates' although I am proposing the (somewhat less elegant) terminology of 'serendreduplication'/'serendreduplicates'. Also, I would prefer to use the terms 'pseudohomophones'/'psuedohomophony' where I am using the (somewhat less elegant) terminology of 'serendhomophones'/'serendhomophony'. I will be using these terms interchangeably. The use of 'serend-' terminology is there to disambiguate matters if any confusion should arise.
Update: Thanks to my friend Krishna Kunchithapadam for his suggestions, including and particularly the point on retaining the use of the terms 'pseudoreduplication' and 'pseudohomophony'.
Second Update: Note that 'pseudohomophony/'serendhomophony' is a statistical concept. If a significant fraction of speakers of language B think that word XB in language B is a homophone of the word XA in language A, then XB and XA are 'pseudohomophones'/'serendhomophones'.
Third Update (Nov 30, 4:34 pm): Note that my definition of 'serendhomophony' is general enough to cover the case where word XB in language B is a homophone of the word XA in language A for a particular performance, for instance, for a single Youtube video of a song. A different Youtube video of the same song could lead to a different word XB-1A in language B to being a homophone of the word XA in language A.
You could think of the performance of a song performance you listen to on Youtube as a process. It is a process (cf. stochastic process) producing words XA1, XA2, .. XAn in language A. Simultaneously, it is also a process producing the words XB1, XB2, ..., XBn in language B for speakers of language B serendhomophonous with XA1, XA2, ..., XAn. And equally simultaneously, it is also a process producing the words XC1, XC2, ..., XCn in language C (again serendhomophonously). And so on. (You could cover every language and dialect on earth).
A different performance of the same song on Youtube will be a process producing words XA1, XA2, .. XAn in language A. But simultaneously, it will also be a process producing the words XB-1A, XB-2A, ..., XB-nA in language B for speakers of language B (serendhomophonously). And equally simultaneously, it is also a process producing the words XC-1A, XC-2A, ..., XC-nA in language C (again serendhomophonously). And so on.
Note that there are two levels of homophonousness here : per performance and per language.