The closest attribute available for a spoken recording of a song is “a cappella”, but this is a type of performance that still performs music (therefore composition credits are relevant), just without instruments. There should be a “spoken” attribute for these recording–work relationships (which would suggest to ignore composition credits).

Example.