Depending on how you look at it, I think the answer to questions 1 and 2 is either "
NO, BUT" or "
YES, BUT."
The PSR-E443 has no MIDI ports, so there's
NO way to directly connect it to another MIDI keyboard and use it as a slave unit of the other keyboard--
BUT if you connect both keyboards to a computer then you can do what you're asking about in questions 1 and 2 by using a DAW.
So in other words,
YES, you can do what you're asking about--
BUT you must use a computer and a DAW to do it.
I haven't tried it yet, but I think you should be able to sync the PSR-E443's arpeggios to the tempo and beats of your PSR-2000-- assuming the PSR-2000 is sending that data to the DAW-- by setting the PSR-E443's "External Clock" function to "On."
As for playing in parallel 4ths and 5ths (or other intervals), you'll be able to do this if your DAW provides that feature. I've done it in Acoustica Mixcraft, and it's incredibly easy to do. You can layer up to 32 different outputs-- because that's the PSR-E443's maximum polyphony-- although you're actually limited to some number between 16 and 32 layers depending upon which voice(s) you use. Note that those numbers are based on the assumption that you're playing only one key at a time on the PSR-2000 or other controlling keyboard, since the number of layers must be multiplied by the number of keys being played:
number_of_keys_played X number_of_layers <= 32 = maximum_polyphony
So if you intend to play only 1 key at a time (as you'd play a monophonic synthesizer) then you could layer up to 32 outputs. But if you intend to play up to, say, 5 keys at once then you could layer up to 6 outputs (because 5 X 6 = 30 <= 32 which is okay, whereas 5 X 7 = 35 > 32 so layering 7 outputs is not okay).
These numbers also assume that you aren't using dual voices, split voices, or styles on the controlling keyboard, because if you are then you must take them into account when figuring your numbers.
As I said above, the PSR-E443's maximum polyphony is 32 notes (i.e., it contains 32 independent sound generators), although it actually varies between 16 and 32 depending upon the voice(s) used. Some voices use 1 sound generator (or element) per note, so you can play 32 notes at once using those voices. However, other voices use 2 sound generators (elements) per note, so you can play only 16 notes at once using those voices. I'm not talking about two voices layered together (Main and Dual, or R1 and R2 as they're called on some Yamaha keyboards), just a single voice. If you look in Yamaha's XG documentation you'll see a tabular listing of XG voices, and it includes a column labelled "E" (for "elements") that indicates whether a voice uses 1 or 2 elements. For example, the "Grand Piano" and "Percussive Organ" voices both use 1 element, so you could play 32 notes at once using one or the other of those voices-- or if you layer them together then you could play 16 notes at once (because for every key you press you're actually getting 2 notes). On the other hand, the "Clavi" and "Rock Organ" voices both use 2 elements, so if you play either one by itself then you can play 16 notes at once, or if you layer them together then you can play only 8 notes at once (because each key you press plays 2 notes and each of those notes uses up 2 elements, hence a total of 4 elements per key). Likewise, if you layer a 1-element voice with a 2-element voice then you can play 10 notes at once (because each key you press uses up 3 elements, and 32 elements divided by 3 is 10 with a remainder).
Obviously, if you use Mixcraft or another DAW to layer multiple outputs together, you must take into account how many elements each output will use up. For example, if you layer "Grand Piano," "Clavi," "Percussive Organ," and "Rock Organ" together then each key you press will need 6 elements to play the desired notes, hence you could play up to 5 notes (keys) at once.
Anyway, when you layer multiple outputs together in Mixcraft, you can use the same voice or different voices, and you can transpose each output however you want. For example, you could layer two outputs using the same voice, with one output transposed 0 semitones and the other output transposed +5 semitones, such that each key you press will play that particular note plus the note that's a fourth above it.
By the way, you'll also need to take into account how the voices will be selected. The easiest way to do it is to have the DAW simply pass along whatever Bank Select and Program Change events are generated by the master keyboard, in which case you'll (probably) be layering the same voice on top of itself-- and you'll also want to be sure that the selected voice is available on the slave keyboard. An alternative is to have the DAW send its own Bank Select and Program Change events to the slave.
Another thing you need to remember is that each voice will require a MIDI channel of its own. For example, if you layer three outputs of the same voice together, then you can send all three on the same MIDI channel. But if you layer three different voices together, then you must use three different MIDI channels to send them.
I know all of that sounds pretty complicated, but it's actually very straightforward and logical once you "get it."
