This paper tries to reveal the constraints on the use of the present perfect progressive (PPP) in combination with numerical object NPs. Existing accounts tend to take situation type classes as a basis for the description of the PPP. It is shown that such an approach does not yield adequate results. (Un)boundedness (Declerck, 1991; Depraetere, 1995) plays an equally important role as (a)telicity in determining whether the progressive can be used or not. (Un)boundedness, as opposed to (a)telicity, is concerned with actual terminal points (of situations referred to) rather than potential (inherent) endpoints. It will be shown in this paper that, in some cases, the conflict between the unboundedness inherent in the progressive form and the boundedness often brought about by numerical object NPs that are used in nonstative sentences results in unacceptability. Considerable attention is first paid to the constraints on the use of the past progressive with numerical object NPs. The second part of the paper focuses on the PPP: apart from (un)boundedness and (a)telicity, the type of perfect and our knowledge of the world also play their part in determining whether or not the PPP is acceptable in sentences with a numerical object NP.