I'm really sensitive to the beginning of a motif or a phrase or something that's kind of the backbone or becomes kind of the spine that you grow muscle tissue onto. You know from that, if you have that good beginning, it's like everything that grows off it often has potential. Maybe I'm good at that early bit of recognition of pieces of potential. I'm not sure.

I download music just like anybody else, but it's a weird relationship when you're a musician.

I'm not a peak oil person. I'm not a biohazard apocalyptic kind of freak. I don't have a supply of weapons or gold bars under my house.

For me it's the hypnotic simplistic reduction of hooks into some form of prayer wheel or something. I really appreciate people who can work in those quiet environments making really precious latticework like lace in a weaving but, for me, I like it a bit more hairy and like psychedelic Peruvian knits or something! You know what I mean? I go into fabric analogies, you know.

I try to structure albums in a pattern, like in a way where there's a motif that runs throughout or some kind of conceit that informs it in a general way. Maybe it's in a harmonic key. I like to go metastructural sometimes, like look at more than the three-minute passage and how that interacts with other pieces. And I've been increasingly interested in false starts and fraudulent beginnings, and things that don't reach their implied conclusions. I take an album and I kind of start moving things around like Jenga.

There's a limit of any form of representation; it's the same about writing about visual art. I still think it's useful for people to think through things in a deeper way, and use adjectives even if they're not sufficient, you know? I always find it interesting what terms they use to refer to the work. It's always different, and that's kind of intriguing. Sometimes it's clichés, but often it's really creative ways of paraphrasing or reformatting what to mean seems something else. I like that, personally.

People still talk about sampling as this new, progressive problem in music. There are technologies now where you can glean the polyphonic information out of a sample and then put that back in and then score it for five instruments. You don't need digital audio to sample; you can rewrite things.