After reading Marco's query the first possibility that occurred to me was that of ;s.r:nkhalaa. In the time I had available, I compiled the sets of .Rgveda sentences that you see in the attachment. If you concentrate on what I have put in boldface type, you will notice that the order in which the boldface words occur is the same as that of the actual or seeming words in Marco's first concatenation. So, there is some promise in the exercise I attempted, but to determine the exact principle (or the major principle among possibly several principles) behind the nonsensical sentences or chunks cited by Marco, much more collecting of .Rgveda-pada-paa.tha data will be necessary. 

One difficulty that will have to be overcome is presented by the places at which Marco's nonsensical sentences occur. To see the pattern I have mentioned above (namely the order agreement pattern), I had to go beyond the padas of the first five verses, but Marco's first ;s.r:nkhalaa is reported to have been written after the pada-paa.tha of the first five verses. The same oddity may exist in the case of the nonsensical sentences occurring after the fifteenth verse. Why the particular placing if the scope of the sentences is not found to be restricted to the preceding five or ten verses? We do not notice such intriguing extent of coverage in the case of the A.s.taadhyaayii or Nirukta ;s.r:nkhalaas. Their coverage does not extend beyond the sections at the end of which they occur. Perhaps a study of more .Rgveda-pada-paa.tha manuscripts will lead us to a solution. Perhaps the order agreement I have seen created a false hope. 

In the case of one of the sentences cited by Marco, we have  atra nirnnayam, which in itself is not nonsensical. Are such tails or appendages found in the case of other sentences that sound like ;s.r:nkhalaa sentences?

a.a.