I'll start with the last, namely that this all started when someone compared science to detective work and I said that it was not due to the fact that detective work deals with unobservable particular phenomena whereas science deals with observable universal phenomena. Someone posited that Darwin's theory is a counter claim and I objected. So that's how this all got started. I assumed by defending the person I was arguing with that you were defending his position, which, apparently, you were not.
Expansion of space: You still did not address how the ability to observe a table is any different from observing these phenomena. We observe the redshifting of light (granted, not through our eyes, but we still directly observe it) and infer the only conclusion that is logically allowed by observing that phenomenon, that the universe is expanding. Similarly, we observe the light reflected from a table and infer the only conclusion that is logically allowed by observing that phenomenon, that a table is right there. Perhaps our difference is in how we are using the term observable - I'm using it to refer to anything that can be experienced, either through it's continuing occurrence or through replication. I'm essentially using it as a means of contrast to unobservable phenomena (i.e. things that can no longer be experienced due to their having finished, namely events of the past). Now, some events of the past are observable in ways that most aren't - namely the way in which astrophysicists can still observe the direct after effects of the "Big Bang". No one can observe the after effects of, say, Caesar's decision to cross the Rubicon.
I think we are using the term "observable" in the same sense -- meaning the capacity of a thing to create an impression on us directly through our senses. In the strictest sense, I suppose neither a galaxy nor a table is "observable", since it is the light emitted from the one and reflected by the other, rather than the thing itself, which creates the impression. However, I don't think there is much profit in drawing such fine distinctions, so at odds with common usage. As I use the term, both galaxies and tables are observable.
There is another class of things, however, which are not observable in the same way, and may be perceived only indirectly. Our perceptions of such unobservable phenomena consist of inferences drawn through the application of reason to sensory impressions of other phenomena which are observable. In this unobservable class, I would include the expansion of space, perceived indirectly through the redshifting of galactic light, and dark matter, perceived indirectly through the rotational speeds of galaxies as revealed by their shapes.
(I agree that the "aftereffects" of Caesar's crossing of the Rubicon, whatever they might be, do not permit the same sort of inference of cause that redshifted light permits. Our knowledge of Caesar's doings stand on the far shakier foundation of human account, perhaps buttressed here and there by archaeological findings. All is not lost to vagary, however, as the event itself may still be observed from a proper vantage point in space a little over two thousand light years away.)
On Darwin: But you seem to be avoiding the point that experimentation is necessary to make science sound. You seem to be claiming that any theory based on the analysis of the world is science. So, is the claim that the stars and planets revolve around the earth science? It was a claim made by observations of the empirical world that was falsifiable. Is the theory of spontaneous generation science, since it too is empirically based and falsifiable? I would say no. There must be some sort of experimental rigor in order to argue for or against a point.
As I have previously stated, when possible, experimentation is an excellent means of examining the reliability of a theory. Unfortunately, experimentation is not always possible when a theory is advanced, given the constraints imposed by the existing state of the technological arts. This does not mean, however, that we must necessarily give up on empirical verification of that the science cannot be made "sound". Frequently, the discipline offered by experimentation may be exacted through comparison of theoretical predictions with observations of natural phenomena, as when Arthur Eddington's photographs of the apparent displacement of stars during the 1919 solar eclipse confirmed relativity's predictions about the Sun's warping of space.
No, I am not claiming that any
theory based on the analysis of the world is science. I won't bother to rehash the features which make a theory science, as another poster in this thread (it might have been Demalachine) did an excellent job of it some days ago.
Where you and I differ is that you seem to require that a theory be proved right before it can be called science. In my view, however, science is a methodology, not a result. It is as valuable when the light it sheds illuminates a dead end as when it shows the way forward. Science's greatest strength is its capacity to jettison those of its products (like Lamarck's theory of inheritance of acquired characteristics) which eventually fail empirical tests or are eclipsed by more persuasive approximations of the observable world. I think science is best though of as a process by which approximations of the world are successively refined or replaced with better ones.
I don't know enough about the history of the Ptolemaic, geocentric model of the heavens to tell you whether it was a product of the scientific method, though the development of the scientific method is generally credited to a much later time. The Ptolemaic model was at least a step in the direction of science, as it was obviously based on painstaking empirical observation of the sky, did a remarkably accurate job of predicting the apparent movements of heavenly bodies, and made falsifiable predictions. If it fell short as a piece of science, it was not so much in its being wrong, as it was in its failure to provide any reasoned explanation for all the little curlicues and reversals of direction in the paths of the stars and planets required to make the thing work.
Though replacement of the Ptolemaic with the Copernican, heliocentric model was a signal achievement, the latter was not without its own serious faults, chief among them its claim that the planets moved in circular orbits about the Sun (as Kepler later demonstrated, the orbits are elliptical). And, of course, the Copernican model was "incomplete" for its failure to explain the mechanism that kept the planets in their orbits, in much the same way Darwin's theory of evolution was "incomplete" for the absence of genetics. But, surely you would not say that, for the error of circularity and the failure to explain gravity, what Copernicus did was not science.
Similarly, Newton's conception of gravity as a force exerted by masses upon each other instantaneously across space has turned out not to be as accurate an explication of falling objects as Einstein's warping of space. Was Newton's work not science?
The problem with your definition, I think, is that there will be no science until science has attained the certainty of religion.