I think atrapado hit the nail on the head when he wrote:
From this quote what I gather was that Dawkins is saying you cannot define the essence of rabbit based on this. I see no proof that you cannot define a rabbit just that his method didn't work. All I was trying to say was "sure we don't have a definition or a definition that we agree" but that doesn't proove that a definition doesn't exits or that someday we'll find or agree on one.
The issue here is one of nomenclature -- "what is a species?" I brought up a related problem, which is "when does a species stop being one species?" If, for instance, sickle-celled humans could not reproduce with round-celled humans. This is easy to imagine, even though it is not true. Imagine that the developing embryo of a hybrid human always died early in development, as it was developing its circulatory system, because it was receiving conflicting instructions to create red blood cells with mutually incompatible traits, or because its mother's circulatory system was not compatible with it and thus the womb was inhospitable.
This would mean that we had two distinct groups of humans in the world. We would have to create a new species name, like H. secula (Latin for "sickle"), for the sickle-celled humans. According to this scenario, the first person who was born with sickle cells was the first member of that species, even if it was not recognized at the time they were born. Where this gets confusing is that, if a person is born who cannot reproduce with any other kin of his family, no matter how closely or distantly related, then his line would immediately die out.
Therefore, there must be a gradually declining reproductive compatibility in nature between diverging groups. We could measure this as a percentage of fetus viability. The first generation that evolves in some way may have 99.5% viability with his ancestors, and establishes offspring of his own which evolve a tiny bit more. They have 99% viability with the first generation's parent generation. The third generation has 98.5% viability, and so on.
According to this reasoning, one could at least define a genetic point where offspring were statistically less likely to be possible between two groups (<= 50%). If we defined that line as the demarcation between species, then absolutely there was a first human, who could not reproduce with contemporaries from related families who were slightly more simian than he was. Fortunately he had enough relatives who were closer genetically, but not too close, that his line was viable.
Now, this is all following the notion that species are defined by their ability to reproduce with each other, and looking at Wikipedia, I see this is only one way to define a species. But according to this long-accepted definition, I fail to see why there cannot be a first member of any given species. This is not questioning evolution, this is questioning the system of nomenclature's limits and definitions.