The difference I would explore is the very difference between the physics of music and the psychology of music.
Good, although I would point out that I also (briefly) explored it above in differentiating between (1) scientific principles behind what constitutes a performance of music and (2) a listener's perception of those principles (perception I take to be similar to your use of "psychology"). Mathematical and scientific principles are, in fact, behind every quantitative and qualitative perception that one has. For a second example, consider another artistic medium, such as painting. Painting also has two basic elements that are behind every brush stroke: shape and color. The shape is the summation of the geometric elements that comprise it -- lines, points, etc.; color exists according to a spectrum, affected by light properties as interpreted by the eye. Both of these can be explained through argument. This brings me to your next point:
I would offer that your approach, since admittedly scientific, does not answer using musical arguments.
I'm not entirely clear on what you mean by this, but if it's what I think, then I believe it is a mistake, a conflation of two ideas. It isn't possible to make a deductive argument in defense of either psychology (for a discussion of why psychology is not evidence-based and thus has no valid arguments behind it, I would recommend reading Paul Lutus's excellent blog) or a listener's musical perception. This is simply an emotional reaction to what one hears, affects one's appreciation (i.e., liking or disliking) of the music (as you say, "artistic understanding"), and is neither right nor wrong. This is not to say that one cannot examine one's perceptions and use them as the point of departure for developing an argument. But an argument's purpose is, in the end, to discover truth (i.e., distinguish truth from falsehood) and such an exercise is always rooted in the scientific method of investigation, as (largely) developed by Aristotle.
2. If the rhythm is changed, then it's notation is necessarily changed.
No, this is a nonsequitur. Notation simply refers to the system a composer uses to mark his score. It exists independent of anything the performer does. The performer, in manipulating tempo and rhythm (i.e. duration), does so on a continuum. Just as musical pitches exist on a continuum (Bartok I believe even wrote a piece in quarter tones), so too does rhythm, and to apply rubato is of necessity to change rhythm within that continuum (e.g., something in between an 8th and a 16th in the example of the Chopin prelude above). This is what can make performances infinitely interesting: since numbers and their variations and divisions can go on infinitely, so too is a performer's ability to manipulate them infinite. I think the best adn most original performers are aware of more degrees of the variations within that continuum.
A conductor only manages tempo (from your duration point of view), yet the orchestra maintains ensemble. The easiest (Occam's Razor) explanation for this reality is that there has been no change in the rhythm of the individual parts. Otherwise, is to believe an enormous amount of coincidences that have a probality approaching zero.
I think the conception of Occam's Razor is a philosophical copout. We shouldn't be self-consciously simplifying our assumptions, because in doing so, we might make a mistake or ignore an important piece of evidence. We only want to use just as many assumptions as we need to make our points. I would quote Wikipedia: "The simplest available theory need not be most accurate" and also "However, on many occasions Occam's razor has stifled or delayed scientific progress. For example, appeals to simplicity were used to deny the phenomena of meteorites, ball lightning, continental drift, and reverse transcriptase. It originally rejected DNA as the carrier of genetic information in favor of proteins, since proteins provided the simpler explanation." This is clearly a very harmful way of thinking since it can ignore elements of the evidence solely to try to twist the "available evidence" to one's own ends.
In your example, an orchestra's maintaining of ensemble is a different issue altogether: synchronization. This wouldn't mean that the entire orchestra, playing as one (just as the pianist plays as one), is attempting to follow the instructions of the conductor and in doing so, is attempting to achieve rubato, very difficult to bring off, but possible (within a reasonable margin of error again, since again, perfect synchronization, especially if a mathematical approach to rhythm is deviated from, is never possible for human beings).