It's 2D and it doesn't include text-based positioning (or, indeed, any kind of freetext scene description -- you just 'summon' objects by typing their attributes+name), but its database of nouns and adjectives is notably vast, and it also includes a fairly sophisticated trait-based interaction system.
E.g. a 'hungry boy' will always seek food (that is, any object in the scene which has 'edible' trait). Combine that with the game's physics engine and recreating the classic 'carrot-on-a-stick' setup is few clicks' work.
To more directly compare the two:
Taking a random featured scene (https://www.wordseye.com/view-picture/23507), here is my quick attempt at recreating it in Scribblenauts Remix (Android): http://i.imgur.com/vkq3ha8h.jpg
Parts list: bike (x2), colossal white spotty sphere, cherry, immobile beagle, immobile strawberry, huge red sphere, huge red puddle.
The 'immobile' adjectives are used because the beagle would destroy everything by happily running around, and the almost-round strawberry would inevitably fall off the round top of the sphere.