If you implement a matrix as an array of arrays, it can be either row or column major, depending on whether you define the first index to be a row or a column index.

Arrays are just arrays (and hopefully sequential in memory).

There's some part of the array that is sequential in memory, and that's the "C" array containing pointers to the SVs holding the structures holding the values of the elements. The SVs themselves could be anywhere.

Two-dimensional arrays in Perl are implemented either as arrays of array references (where the references may reference arrays that are stored at any arbitrary memory location), or using hashes where there are special separators in the hash keys.

Neither of these methods guarantees the order in which the elements are stored in memory, and there is no straightforward way to iterate over the elements in sequence. Therefore the concept of row major cannot be applied imho.

If one needs a two-dimensional structure on a one-dimensional array, it's usually done by hand as in $array[$row * $columncount + $column]. Maybe one could use operator overloading to access these kinds of arrays with the usual $array[$row][$column] syntax.