In order for something to be perceptible, it must have sufficient size and exist concurrently with the perceiver. if something exists concurrently and is of sufficient size, then it is fit to be perceived by the external organs. And if one rejects composite wholes such as pots, then there would be nothing fit to be perceived by the external organs. And because there would be no perception, inference and the other pramanas could not work....We perceive. Thus the composite, the whole, exists in its own right.
[The Nyāya-sūtra: Selections with Early Commentaries, Matthew Dasti and Stephen Phillips, trs. Hackett (Indianapolis: 2017) pp. 101-102.]
I found this interesting, because the kind of position the Nyaya philosophers had in view here -- that because composites, like pots, are made of ultimate constituent bits, it is the constituent bits that exist, not the composites -- is one that one finds occasionally these days, and Uddyotakara captures perfectly what I've sometimes thought the problem with it: our knowledge of the reality of these elemental bits is dependent entirely on our knowledge of the composites, by inference, and thus there must be composites to know, and from whose reality we can derive the reality of the ultimate parts, or nothing ever gets off the ground. And, as the Nyaya-sutra indicates, this can be pressed even further.
The translation of Dasti and Phillips is fairly nice, incidentally (I can't speak to accuracy, but it is certainly readable); I don't like that they rearranged the work thematically, but they at least provide an outline of the Nyāya-sūtra itself, which makes up for part of that.