I've only glanced at the paper, but it looks to me as if he's using it in another perfectly usual way, namely referring to situations where the wavefunction and

I've only glanced at the paper, but it looks to me as if he's using it in another perfectly usual way, namely referring to situations where the wavefunction and/or the Hilbert space it lives in can be written exactly or approximately as a product over simpler things. There's nothing wrong with that.