Several people have answered in the comments. I'll write an answer.
Computer geeks tend to have a sense of humour that is different from that of normal people. I know because I have worked amongst them! Where mathematicians would use x, y and other symbols to stand for unspecified objects or quantities, computer folk sometimes use nonsense words. For those who understand the subject, this is considered hilarious.
A justification for using nonsense terms in this case is that we are discussing the abstract concept of recursion and not framing it within any particular programming language or schema. Different languages have different terms for the same thing.
Here is my best guess without reading what goes before.
For example, if the original input is a structure with n fields, the input to each recursive call should be a structure with strictly less than n fields. Of course this is impossible if the structure has no fields at all --- You can’t have [a] negative [number of] fields; that would be silly! --- so in that case we must process the structure using some other method.
Note that the above is a more general definition of recursion than usual. Mostly, when recursing, we reduce a simple count by exactly one each time. This description allows a reduction by any number of fields as long as there is always a reduction AND the reduction does not take us below zero. To grindlebloff differently clearly refers to the treatment of the zero or null case at the end of any branch of the recursion. The end of recursion always requires a special treatment.
Note
A more general version yet would be:
For example, if the original input is an item with n attributes, the input to each recursive call should be an item with strictly less than n attributes. Of course this is impossible if the item has no attributes at all --- You can’t have a negative number of attributes; that would be silly! --- so in that case we must process the item using some other method.
