But you can form a continuous set of linear combinations of these things, just as you can with gluons. Indeed, what the article calls W and Z bosons (and photons) are just such linear combinations--the ones that appear in the low energy limit after the electroweak phase transition occurs. Before that phase transition, different linear combinations (i.e., a different basis of the electroweak vector space) are the ones that naturally appear. So saying that there are two W, one Z, and one photon is really counting basis vectors in the electroweak vector space, just as saying there are 8 gluons is really counting basis vectors in the gluon sector of the strong interaction vector space.