The determinations can be necessarily close calls when sometimes all you have is a jaw or a foot, and some are questioned: H ergaster (mentioned in the article) and H rudolfensis spring to mind. But the fossil record helpfully seems likely to throw up more "classic" than intermediate specimens because it's the former in which adaptation to their environment and way of life are more fully developed: nature abhors a half-adapted population. In practice a truly intermediate form unclassifiable as one or other known species would be more likely to be labelled a newly-discovered species related to both of its neighbours: out of the tangle a clearer picture seems to be evolving than was available only decades ago, though our classification of discrete human species may emerge blurrier than in the past when we had a few australopithicine types, H erectus, H habilis, H heidelbergensis, neanderthals and us and little else that I recall.