First, how do you "split the signal" coming out of the preamp. Do you simply use a y-xlr cable (if there is such a thing) or is something more complex needed to maintain signal integrity?
Nothing complex, there are several ways to do this. First the electrical part: Your preamp may have an output impedance of 80 Ohm. Your interface-, controller- or mixer inputs may have 10 kOhm. If you connect the output to two inputs, your preamp output sees still 5 kOhm and will be happy with that.
After the preamps you get a healthy line signal, preferably balanced, not easy to disturb. The impedances are ok, therefore you can safely distribute one output to two inputs without degrading the quality.
You can make your own cables or use a standard half-normalled patch panel or solder custom connectors. It does not matter as long as you connect one output to two or maybe three inputs. Don't try this in the other direction! Never connect several outputs to one input because this may destroy the output stages of your equipment.
Another method are split boxes, they are used to split microphone lines before the preamps. They work similar to passive DI-boxes but overall balanced. They have one input and two or more outputs per module. The input goes straight through, parallel outputs go through transformers which isolate the original signal from the clones. A solid solution but needs a lot of space and is not necessary to clone a line signal.
The third and widely used method is to split electronically. Many audio interfaces can do this and the manufacturers call it "direct monitoring". The name is often misleading. If the analog signal gets converted to digital, then splitted and routed to an output where another A/D conversion happens, the signal flow is by no means "direct". There is the latency of two A/D conversions. It can be very short and many people live with it. It is not a good solution if you want to monitor many signals because you would use all of your interface outputs. If you have an interface with an ADAT output, you can use an ADAT converter for monitoring. However, in my opinion splitting before the interface is the way to go if your want zero latency.
Secondly, as you're recording into the preamp, the backing tracks are being played through Logic, right? Aren't those effected by latency?
Yes they are but it does not matter. That's the magic of delay compensation in a DAW. Logic measures all plugins and calculates the maximum delay time. Let us look at an unrealistic extreme situation: Due to a special plugin, your playback latency is 20 minutes. Means, when you press the play button, you have to wait 20 minutes to hear the first tone. What are the musicians doing during this time? They drink a beer or two because they hear nothing and cannot play. Then the music comes and they play along. Finally Logic says, "What was the latency? Ah, it was 20 minutes. So I put the files 20 minutes before they were recorded and everything is fine."
Got it? As long as your interface latency is correctly set, which is really important, you wont have a problem. You play along the playback and Logic knows where to place the recorded parts.
To check and adjust the latency compensation of your audio interface you can follow John Pitcairn's instructions on his web page:
http://www.opuslocus.com/logic/record_offset.php
It is an old article but still valid.
Me too, most of the time.