Maybe if you could unpack this a bit.

By keyboard input do you mean just the keybed (piano keys) or other controllers as well?

By "dynamic" do you mean amplitude or just anything that isn't static (like any controller button or MIDI message)?

All of the expressiveness possibilities with the keyboard. Keyboards having, in the case of the MODX, velocity and legato states, what can be done with these, in terms of dynamics?

Montage adds Aftertouch, obviously enough. But perhaps put that aside to just focus on what's possible in terms of everything that can dynamically respond to velocity and legato-ness state.

If I'm understanding the question correctly, the per-part dynamic response from the keyboard is controlled by the Velocity Depth and Velocity Offset parameters.

Velocity:

1) (AWM2) Determine which element(s) are triggered
2) Invokes accent phrases with arpeggios that support these (when threshold is met)
2a) Note: accent phrases may be (time) quantized so the actual accent may lag when the threshold is met in order to align the accent with the given quantized value
3) Optionally (can be turned on/off) controls the volume of "Random SFX" (release sound) that are contained in arpeggios that support these
4) Can scale/set the velocity of notes inside an arpeggio
5) Can scale the amplitude of element(s)/operator(s)
6) Can determine if an arpeggio is triggered or not (arpeggio velocity limit)
7) Can determine if a motion sequence starts (velocity threshold). This includes super knob automation
8) Can scale the pitch (pitch/vel)
9) Can scale the cutoff (cutoff/vel)
10) Can scale the resonance (res/vel)
11) Can scale the PEG depth (PEG depth/vel)
12) Can scale the FEG time (FEG time/vel)
13) Can scale the FEG depth (FEG depth/vel)
14) Can scale the amplitude (level/vel)
15) Can impact level of an element around its velocity limits (vel cross fade)
16) Can determine if an individual key bank within a single waveform sounds or not (key bank defined velocity limits) - you can only edit these for user waveforms. Preset waveforms contain these and are "hard wired"
17) Velocity related scaling can be scaled by the position along the keyboard you strike a note. Example: PEG Time Key Follow Center Key can determine which note is not scaled for PEG Time and then the "Time/Vel" scalar sets the slope of the line impacting how much velocity impacts the PEG Time according to the distance from the center key on each side (one side positive one side negative unless the slope is 0. So these settings would influence the velocity passed along to a certain key position and the scaled parameter (like PEG Time in this example)
18) Speaking of PEG Time - you can define which "segment" of the PEG is impacted. Segments of PEG are attack, attack+decay1, decay (I assume decay1+decay2 - although I'd need to look it up to be sure), attack+release, or all for all segments. The same is true for FEG.

... is this the kind of thing you're looking for?

Left off legato for now although off the cuff I think the XA legato and what mono vs poly "do" are the only times legato comes into play. Although not necessarily legato - drums will allow for you to cutoff a drum key if another key in the same group is triggered. For say closing a hihat - the open hihat will stop sounding when the close hihat is triggered according to the grouping rules. I say not legato because drums wouldn't necessarily need to have you hold down one key and play the next one (in the same group) in close proximity to holding down the first key. You could play detached and the sound would still be cut off of the first note assuming the sample hasn't completed already. For shorter samples this will start approaching "legato" because you'd have to play the 2nd note quickly after the first one to have this happen (you're racing the sample stopping). You're racing if your goal is to invoke forcing the first sample to stop playing earlier than the end of the sample.

Brilliant. Thank. you.

Am trying to figure out how expressive I can make a sound, for playing, with a key focus on Legato as a separate technique for extracting differing expressiveness versus staccato play.

And will then figure out what can be modded by the Assignable Buttons 1 & 2 for differing 'modes' of expressiveness.