Ideograms are "too far gone" to be recognisable to Western readers now - too evolved away from their roots, when different writing instruments were used. So FWIW I would expect that your hypothetical future emoji conlang will not resemble them. Maybe it will look like an AI has redesigned airport symbology ?
At which point it would just be NIH to not just use Chinese. As I understand it, the Americas had vastly more linguistic diversity prior to European contact than the rest of the world combined. I don't know about further south but in at least much of North America sign language was used as a common language. I'd guess if you want to create what could be a universal language you would want to try to align written, spoken, tactile, and signed forms as much as possible, which would suggest directly aligning the signed and written forms as much as possible since both are (at least mostly) visual. And if possible keep the tactile version as close as possible as well. The spoken aspect seems like the hardest and maybe better to not bother with that at all.
Also, IMO emojis are too small. I've been wanting a 4x size for emoji (double width and height). With more compact writing systems you can make everything bigger.
> The spoken aspect seems like the hardest and maybe better to not bother with that at all.
You'd probably find that this is precisely where linguistic evolution is fastest (kids, teens, etc.) and it would quickly get out of step with other modes.
I remembered after a bit that some people teach their infants a simplified sign language to introduce them to language faster, so to the extent that spoken language evolves faster I'm guessing it is just because it is most widely used.
