Field checking doesn't have much to do with accessors, or ISA checks. Reading the rest of your thread, I'd like to summarize.

package Dog;
use fields qw/intelligence/

sub new {
    my $pkg = shift;
    my Dog $self = $pkg->fields::new();
    $self->{intelligence} = shift;
    $self;
}

sub intelligence {
    my Dog $self = shift;
    $self->{intelligence};
}
[download]

That's the code you're responsible for. Lets say the accessor for 'intelligence' had a typo, and accessed the field 'inteligence'. Without using fields and typed variable declarations, you would not get any erros.

This won't compile, and is broken:

sub intelligence {
    my Dog $self = shift;
    $self->{inteligence}; # typo
}
[download]

And normally people would write

sub intelligence {
    my $self = shift;
    $self->{inteligence}; # typo
}
[download]

and the code would compile correctly, but is just as broken.

useing fields is meant to help you write your classes a bit cleaner, and is a bit like trying to catch typos in variable names. Just using $foo wouldn't work normally if there's no such variable, yet perl will only complain with use strict. Using a variable to store the intelligence value, $intelligence, and typoing it the next time around to $inteligence is the best analogy I can find for getting your field names wrong, and that's where use fields; helps.