for the software running on the Raspberry Pi these are 2 completely independent sound cards. While it might be possible in general that a program outputs in parallel to 2 sound cards, I'm not aware of a simple solution for this. Creating a virtual sound card with Pulseaudio might work, but this is not an easy task. I don't think that you will find a pre-configured solution for this.
Hm. Ok, thanks. If I put the amp+ output to a splitter box, like so: http://www.monoprice.com/product?c_id=109&cp_id=10903&cs_id=1090305&p_id=9995&seq=1&format=2 and had the output go to my speakers, as well to another output, is it possible to attach speaker wire to a headphone jack? I basically want to read the audio signal on an arduino so I can create a graphics equalizer display. Will I run into problems because of the power on the speaker wire?
This will create several problems: 1. The output is not ground-referenced, but floating. This might create potential differences that can destroy your circuit. 2. The output will still have some residual switching frequencies which might result in incorrect results of your DA conversion.
You would need a differential amplifier for this. This is quite complex. However, why using an external circuit for this? You could directly to the analysis on the Raspberry Pi. This will need some programming, but it should be possible. The ALSAeq ALSA plugin might be a good starting point for this. It implements an equalizer, but I guess, you want to implement a spectrum analyser - right? It should be relatively easy to implement this with alsaeq as the basis. You will need some experiences with FFT, but I guess you have this. Otherwise it will be also very hard to implement this on a micro controller.
If I were to have an external circuit, I'd imagine the DAC+ would be better with an external amplifier, correct? That way I could put the circuit in line.