- Science cannot test the proposition that biological features are designed.
- Darwinism explains the appearance of design in biology not as actual design but as the product of natural selection and random variation.
- Darwinism is science.
While I think there's something to this objection, I would also reject premise 1 as stated. Surely there are scientific mechanisms for distinguishing natural features from artifacts of the sorts we are aware of (e.g., forensic science can distinguish at least some murders from deaths by natural causes). What science cannot distinguish is a hypothesis that biological features are the product of evolution from the hypothesis that a divinity created biological features that look just like the product of evolution. Similarly, science cannot distinguish automobiles that are created by people from automobiles that are created by a divine being to look just like automobiles built in a human-built factory, nor can it distinguish human beings who were born of a man and a woman from human beings who are directly created to look exactly as though they were born of a man and a woman (Omphalos included). (In other words, God could choose to work directly, simulating evolution, or indirectly, using evolutionary mechanisms or setting up the initial conditions and letting evolution run its course, and those hypotheses are empirically indistinguishable. Some religious believers advocate a view where events have natural causes, yet are also caused by supernatural beings such as Satan. In such an anti-Ockhamite, unparsimonious view, there is no scientific way to distinguish an event with both natural and supernatural causes from one which didn't have the latter.)
If a God-based hypothesis can be formulated in such a way as to have empirically testable consequences which are distinguishable from evolution, I don't see why it couldn't be science. This means there *could be* an "intelligent design" that qualifies as scientific--but what's been promoted in Dover is simply a renamed creationism, rather than a new field with any scientific content.
The real problem for such God-based hypotheses is that there really are no limits or definitions around what God does or would do--no empirical evidence is ever considered to be evidence against God by the advocates. I think there actually is empirical evidence against many specific gods which have been endorsed through the millenia, including commonly held views of contemporary monotheism. If you say that humans are psychologically similar to God (being created in his image), that God is perfectly rational and desires particular outcomes, then actions (or inaction) inconsistent with those desires, intentions, and facts of the world are evidence against such a God's existence. This gives evidential weight to atheistic arguments such as the argument from evil, the argument from (reasonable) non-belief, arguments based on the dependency of consciousness on physical brains, the facts of evolution, religious disagreement, and on the distribution of religious beliefs (indicative of cultural transmission rather than supernatural intervention).