Great post, Bryan! It's often just assumed that science makes progress, and this "fact" is used to contrast with philosophy, which supposedly never makes any progress at all (because we don't agree on everything, or something like that). I think you do an excellent job of turning that idea on its head.

Thanks! I find that students and even non-philosopher colleagues often assume a quasi-positivistic position, without appreciating the seemingly intractable problems of going from pure sense-data ("red, here, now") to actual scientific theory ("protons are composed of two Up quarks and one Down quark").

This is a painfully simplistic straw-man of the process, aims, and products of science.
For starters, If science were as you describe it here, it would have no predictive power at all, and would have fallen out of fashion long ago, as a tool for understanding the universe. Secondly, you seem to be offering an objection to an argument that nobody is making, by pointing out, via Copernican Revolution, that chronologically later explanations are not necessarily improved explanations. Thirdly, you claim that more precise and nuanced explanations of observable phenomenon are not progressive, merely because you find them "too complex", and not symmetrically pleasing. 
And all this occurs just in the first two paragraphs. You go on to conflate early attempts at natural philosophy with modern science, equivocate theories with descriptions of observable phenomena, confuse interpretation with inference, and you conveniently ignore the fact that the content of a failed theory doesn't get discarded wholesale. The stuff that works gets brought forward. The stuff that doesn't gets discarded. 
It seems your preference for really simple answers to complex problems extends to your understanding of the progress of science, and the progress of humanity that has in fact resulted from it. Rather than building up these cheap targets to fire at, why not instead make a positive case for philosophy as a discipline? If Stephen Hawking, et. al. are indeed so wrong, then it should be a matter of course to point to the things in philosophy that demonstrate just how wrong they are.
And in doing so, you could even be using a form of the scientific method. Which is perhaps the first thing you could put on that list of philosophical advancements.

Although interesting, and however tempting it is to say that scientists aren't talking about the same things as they 'progress,' it seems to me more that Galileo learned Aristotle's vocabulary, found it to be lacking in explanatory power, and so edited or improved upon the vocabulary. Newton then learned Galileo's vocabulary in some sense as well, and, again, edited or improved upon the vocabulary to increase its explanatory power. The same is done in the case of Einstein towards Newton, it would seem, in accomodating for special cases near the speed of light. So, it seems to me that the validity of the argument that scientists aren't even talking about the same things is contingent on an assumption that they developed these vocabularies independent of each other: Newton without Galileo, or Einstein without Newton. This is of course false, and so because their vocabularies evolve out of one another, they do seem to be in legitimate conversation with each other. 

Isaac Newton:
"If I have seen further, it is by standing on the shoulders of giants." [Translated into Modern English]
But if those scientific giants where looking the wrong Way then surely those standing on them looked the wrong Way too. 
Thoreau and I:
If you are going to build a castle in the sky make certain the foundation is absolute. 
As for the current state in physics of probable mechanics, there sure is a lot of smoke in their mirrors!. And sadly science continues to expound on it: ..News at 11: The god particle has been divided in a super duper collider and the holy ghost has been found. The scientist who predicted the ghost has named him Casper (after himself of course) the friendly ghost. 
Thanks for the interesting subject,
Isn't science funny? =

Of course, there is the little matter of whether time is progressive at all. There is no mathematical model which necessitates this, and so it might seem to be largely a human perspective, or, if this kind of reasoning is permissible, it requires humanity as its proof. One commentator on progress, whose name I do not recall, made a distinction between invention and innovation. Invention, he explains, is true creation as if from nothing, whereas innovation takes the invention, the new idea, and develops and ramifies it. This distinction throws a rather different light on the issue of progress in science. Brilliant minds are often deeply conservative and cannot outstrip convention, but are past masters at exploring and filling in the possibilities of existing thought and methods. But true invention tends to buck the trends. Innovation, in this sense, is always very much of its time, invention untimely.

In each case the more recent explanations are more general (apply to more phenomena) and more parsimonious (they explain the wider scope of phenomena without simply tacking on multiple ad hoc explanatory entities).  Universal gravitation works for projectiles and pendula in addition to planets.  Relativity extends to high velocities, as acknowledged in the original post, and adds an upper limit on velocity.  Modern macromolecular biochemistry explains many processes in living organisms, not only the inanimate material treated in older chemistry.  Whether increased generality and parsimony count as 'progress' could always be debated, of course, but I think we should be clear that this is the direction scientists take to be desirable for their fields.   For example,I believe it was this desire for increased generality and parsimony that led to the "intent to measure" mass, not just some random culturally-relative urge.
With regard to philosophy, deniers of progress in that field may be using a particular idea of what philosophy is, like some of our students ("We all had papers to write and problem sets to finish, so after dinner we procrastinated for hours by sitting around philosophizing," by which they mean "Chewing the fat not about less-important things like whether Madonna or Lady Gaga is sexier, but rather about Big Questions like the Meaning of Life").  However, if we take philosophy to be the attempt to discover and clarify values and assumptions and to analyze how our beliefs depend on those values and assumptions, and if we accept that better-organized puzzlement is a big improvement over badly-organized puzzlement, then the denial of progress in philosophy is easily refuted.
- Steve George

Thanks to everyone for the many comments, each of which advances the dialogue in some way.  Allow me to make a general observation, and then address some of the specific points that were made.
GENERAL
If you think later scientific theories incorporate (in some sense) the observations or theoretical claims of earlier theories, have you actually read the earlier scientists?  In other words, have you actually read Newton's Principia, or Galileo's Dialogue concerning the Two Chief World Systems, or Copernicus's On the Revolutions of the Celestial Spheres?  If not, you really don't know whether science advances.  Later textbooks misrepresent earlier theories in an unconscious effort to paint a picture of scientific advance.
Similarly, have you read works by the (ex-physicist) Thomas Kuhn on the history of science, most importantly The Structure of Scientific Revolutions?  I have a guide to some of the chapters in this work here:  http://faculty.vassar.edu/brvannor/Kuhnhelp.pdf If you have not read Kuhn, or some of the work inspired by him, you do not have an informed opinion on the history of science.
Just as a sample, consider the following different views that have been held of atoms:
?Democritus (5th cent BCE) ? indivisible particles, with shapes that explain properties
?John Dalton (early 1800s) ? particles that differ in ?mass? and ?complexity?
?J.J. Thompson (1897) ? particles with electrons embedded in them: ?plum pudding model?
?Niels Bohr (1913) ? particles with electrons orbiting around nucleus: ?solar system model?
?Erwin Schrödinger (1926) ? particles with electrons in indeterminate orbital zones
?Otto Hahn (1938) ? particles which can be divided: nuclear fission is possible
 
The preceding descriptions transparently do not apply to the same entities.  (An indivisible particle is obviously not the same as a divisible particle, for example.)  Hence, the ontology of science is not converging on some one theory. 
To Greg Gauthier
?Perhaps I am wrong, but you seem to take it for granted that there are observed phenomena that are neutral between theories.  But this is precisely what several hundred years of research (in both philosophy and psychology) have shown to be highly implausible.  What would a pure observation statement look like?  Perhaps "red, here now"?  How can we get from a statement so parsimonious to the actual content of real scientific theories like "protons are composed of two Up quarks and one Down quark"?  Real science is conducted in statements that already embody theoretical comments, even at their most basic, like "the current through a conductor between two points is directly proportional to the potential difference across the two points."  (Notice that none of the terms in that sentence is a direct observation term.)
To Truman Chen
Later scientists certainly do engage with the vocabulary of their predecessors.  However, the question is whether the content of the earlier vocabularies are preserved in the later vocabularies, and this seems false.  "Mass" for Einstein simply does not mean what "mass" meant for Newtwon.  The radical nature of the meaning-change is obscured by the fact that Einstein uses Newton's term. (See also my earlier example how "atom" has meant very different things in the history of science.)
To Steve George
"Generality" and "parsimony" are either too vague to be of use in theory choice, or they are routinely violated.  Galileo's physics and Darwin's evolutionary theory are both much LESS general than their Aristotelian predecessors, because they drop problems that the earlier theories were intended to solve.  (For example, Galileo had no plausible explanation for why the earth wasn't torn apart by the radical motion it was undergoing.)  "Parsimony" is an aesthetic property, not a precise mathematical one.  Do we prefer theories that are committed to fewer entities, or fewer KINDS of entities?  Do we prefer fewer entities or fewer mathematical laws?  Basically, "parsimony" is just a psuedo-precise way of saying, "This theory looks more plausible, given my tastes and those of the current community of scientists."

Thanks to Bryan Van Norden for this interesting initial post and responses to comments.   In his response, he provides a substantial reading list and says that if participants haven?t done this reading, they ?really don?t know whether science advances,? and they ?do not have an informed opinion on the history of science.?  Even if there is some truth to this, such a criterion for legitimacy in participating in the discussion seems out of place in a public forum like this one, which invites people to register and participate without stated academic prerequisites.  Imagine if a working scientist reading this website were to post this: ?No one who hasn?t actually done science can really know what it is about.  If someone hasn?t completed a multi-year scientific project working daily in the lab or in the field, he or she does not have an informed opinion on anything about science.?  Even if there is some truth to it, it wouldn't seem appropriate here.
Van Norden?s original post was clearly and concisely stated in plain language, and included not a single citation or link to other work, which I thought was a strength.  At the very least, if there are prerequisites for making legitimate comments, this should be stated in the original post, so people don?t waste time formulating responses that will be considered invalid based on criteria other than the actual content of the response.  Perhaps what Van Norden intended was to recommend useful reading for those interested in the topic, though it didn?t come out that way.
Much writing about science identifies parsimony as one characteristic of scientific explanations, and this rings true to most scientists.  Certainly the notion of parsimony is not instantly clear and needs expansion.  ?God wills it so? is supremely parsimonious in the number of words used, but doesn?t cut it as scientific explanation.  From the fact that parsimony is a complex property that needs to be explicated in the context of science, however, it does not follow that it is necessarily only ?aesthetic? and just a matter of the ?tastes and those of the current community of scientists"  as Van Norden claims.  Perhaps the whole notion of what ?explanation? is could be the subject of another post on this very useful website.
- Steve George

If you see science as a project of eliminating doubt, then you will naturally see it as progressive. But if you see it as a process of raising doubts, you might think again! As for philosophy, if you see it as progressive you're not reading what I am of it, and I read, pretty much, everything. Frege set philosophy on a track of retrenchment, not enlightenment, that carries on very much today. Bob Dylan says you can't criticise what you don't understand, I'd add to this you can't enthuse about it either. I'm with Bryan about the reading. If you suppose philosophy advances, try reading Time in the Ditch, by John McCumber, or Against Method or The Conquest of Abundance, by Paul Feyerabend.

Well. This blog post has truly generated voluminous opinion and utterly philosophical dynamic tension. It seems to me unassailable to say that science is a progressive discipline because it builds upon itself through both successes and failures. Development of thermonuclear weapons was an astounding success for science an a resounding failure for mankind--but, on the bright side, it led to peaceful applications such as power stations and nuclear medicine. The other commenters have said pretty much everything else worth saying. So, I'll sign off,
 
Cordially,
Neuman.

Unassailable? Is this a philosophy discussion or a knitting bee?
The question raised might just as well be phrased: "Does logical inference reference its antecedent term?"
The question then becomes, what is the pertinence, if any, of the question? And what if the pursuit of wisdom is regression, if only seen from the perspective of a dogmatic faith in reference?
Modern manufacture required a whole array of machine tools. But before it could even get started it needed a way to make things perfectly round, easy enough, but also perfect flat, so as to use as a standard for creating all the other devices needed. This was a more difficult problem, delaying the modern world for some time. The Greeks found a solution, as can be seen in a cursory look at Greek columns still surviving. How? How do you mill a surface flatter than the human eye can judge? To close enough tolerances to become the basis of modern milling methods? Such breakthroughs are, of course, prerequisite to "progress", like clarifying the butter before making the roux, but this does not mean that there is an irreversible accumulation of capabilities. Sometimes we gain capabilities only at the cost of others we thought we had secured.   
 

I wonder whether there is a relatively straightforward verbal issue that underlies disagreement in this discussion so far.  If words like "advancement," "progress," and "cumulative" are taken to mean "steady change always in one direction in a single, monolithic effort," then it sets up science (and philosophy too) to be seen to have failed to advance, progress, or be cumulative.  However, that's not how those words seem to be normally understood.  One's savings and investments, as assessed today, may have progressed and advanced cumulatively over decades in spite of having undergone a decrease during the recent recession.  A tree may advance or progress in height in the long run even if some branches are lost, and even if the main trunk breaks off and another branch eventually becomes apically dominant. 
In science and philosophy, lots of people are engaged in the effort at any one time, pushing things in different, perhaps mutually inconsistent, directions.  Many of these efforts will never show up centuries later as specific entries in the current state of knowledge in the field.  However, if those efforts failed because they stimulated other people's successful attempts to find weaknesses in them, they will have contributed greatly to long-term cumulative progress.  (I realize there's a school of thought that says this characteristic of "falsifiabilty" is the only measure of worth, and I don't know enough to argue for or against that extreme view, but at least it seems to be one aspect of how these fields progress.) 
Maybe Frege's work on developing systems of formal logic will be eventually be seen as a backward step in philosophy, as Gary Washburn provocatively argues, but set theory turns out to be a very important human intellectual tool, whether or not one defines it as being inside or outside of philosophy.  It led to the efforts of Bertrand Russell and others to find flaws.  Perhaps the resulting discovery of paradoxes in the system could be said to have invalidated the effort as judged by its initial goals. But all that intellectual ferment is surely related, if only indirectly and perhaps non-monotonically, to later attempts to formalize systems.  From this came very significant results, such as showing that reasonably complex formal systems must necessarily be incomplete.  To me this is great overall progress, individual reverses notwithstanding.
Descartes argued for the primacy of reason and logic in justifying all our beliefs.  That clear, extreme statement provoked Hume and others to find flaws, and they seem to have found some.  For example, if we believe that past experience, even when incorporated into well-established scientific laws, is a basis for expectations about how things will be in the future, that's fine but such a belief apparently can't be grounded on reason, evidence, or any combination of the two.  We can write all we want about "Descartes' Error" etc. but that in no way detracts from Descartes' importance in what I would call "progress" in this field.
Same in science.  One could go through the list of changing ideas about atoms in Bryan Van Norden's reply on July 31. That list to me is a great example of scientific advancement and progress in the sense I'm arguing for here - not that it was steady, monotonic, or monolithic . The critique was that these people (from Democritus to Hahn, over a span of 2,500 years, weren't all taking about the "same entities."  What I believe they were asking about was, "What would happen if we start with apparently homogenous material and keep dividing it into smaller and smaller pieces?  Could we keep going indefinitely, or would we reach a piece that, if divided further, no longer retains important characteristics of the original material?"  (If the word "indivisibility" was used, it turns out that the above sense of indivisibility, rather than absolute indivisibility, proved productive.)  Observations and experiments led to the concepts of mixture, compound, and element, each with its own characteristically smallest piece - "atom," in the case of elements.  If Thomson hadn't formulated his plum-pudding model (note: "model!" 1 Right, not final absolute reality!) of the atom, Rutherford wouldn't have put an undergrad on the project of measuring large-angle scattering of alpha particles by gold foil.  Since some alpha particles did bounce almost straight back, the plum-pudding idea model doesn't work. (Not that Rutherford predicted that outcome, but the whole thing wouldn't have been tried without the provocation of Thomson's model.)  The fact that Thomson's model isn't foremost in anyone's mind when planning experiments on atoms today doesn't negate its importance in stimulating progress in our understanding of atoms. 
Irregular movement with fits and starts, big jumps, sometimes steady advances, and temporary reversals can still lead to real, overall, long-term progress.
- Steve George

Is evolution "progress"? Or just change? There is no "evolutionary" reason why it should be anything more than a rolling alteration, without any direction at all but survival in response to other altering conditions. However, there is a wrinkle in this. The more an organism develops along a certain line, the further it gets from its original form. Wings will probably never be grasping limbs again. Quadrupeds will not likely have six or eight legs, as their earlier forms did. Cells that have become dedicated to support tissue will not likely ever be reproductive cells, or independently replicating cells as in bacteria. I suspect, in fact, that some cells do revert in this sense. We call this "cancer". But is the analogy useful? Is the alternative to progressive knowledge cancerous?
Consider the case of plate tectonics. Its discoverer was ridiculed for decades. Don't many "discoveries" come through a struggle against conventional wisdom nominally dedicated to the progressivity of knowledge? Is this discussion hampered, even hamstrung, by such conservatism? Frege would get no objection from me if he did not saddle philosophy with the burden of a conviction in positivism. The work of philosophy is not to build a store of possessions, but to clear away obstacles to a different kind of change and development. The reason Frege's logic does not work is because ideas, and the mind simply, simply are not constituted as they would need to be if his premise were true. The premise, that is, that propositional conditions are hermetic, so that the reverse can be used as a measure of truth. One thing logicians keep getting wrong (obstinately?) is that logic is about truth. It is not. Logic is the science of validity of inference, not truth. Just a science is about fact, not the founding conditions of knowledge. Immodesty has no excuse in either. I suppose facts are cumulative, if they do not denature themselves as we accumulate them. But the terms logic famously do alter as we progress from a premise. So much so that a long enough sequence of inference can yield a term contrary to its premise. Rather like a message that gets garbled when passed down a line by whispers. This does not mean that the method is corrupted the theory, it might just as well mean the theory itself is corrupt. If so, the difference might be resolved by the explanatory power of the one over the other. That is, the presumption of progress does not explain what its inverse needs and does explain. For instance, the terms of the proposition must be hermetically independent for the system of inference to control the validity of inference, and all inference must simply restate the premise to be valid. But these prerequisites invalidate the linguistic content of the terms comprising the original proposition. If meaning doesn't leak a little, or isn't "labile" (as the existentialists like to put it) language would be impossible, either as a matter of human development or as one of formal necessity. But this would seem to suggest that every logical term is anomalous to logical form. However, what if this is not the initial condition of the proposition, but the most exquisite rigor in the sequential inference from it? Then what? How much in variance to its original term does it take to convince us that the change is the most real and most rigorous origin of the ability to reason? And if the least term of that change is the most conclusive proof of the corrupted state of the conviction in reasoning from a premise or antecedent term, then the conviction in the constancy of terms reveals that corrupted origin, not as the corruption of its methods, but as the inconstancy of its conviction. Hence all too human moods. Evidence, empirical fact, demonstrating the supremacy of human frailty over the progressivity of the accumulation of knowledge.