Rubato where there is no indication of rubato

Question

Unless I'm crazy (not impossible, tbh) every performance of Messiaen: Prelude No. 7, Plainte calme that I have heard does not seem to follow the sheet music.
There's a passage where the time signature changes from 3/4 to 6/8, but I'm yet to find a performance where I can hear the change in tempo. Also, it seems that people play this song almost entirely with rubato, even though there is no rubato indication.

This rubato thing is not exclusive to Messiaen: many times I've heard great pianists employing rubato to pieces where the piece has no indication for it. People playing Chopin in particular, seem eager to add their own pulse.

So I guess my question is:
Where do these liberties come from; is this a matter of interpretation (player's choice), or is there something guiding these performances, like "composer X's pieces should be played more freely"?
And if there are guidelines, where can I learn more about them? (suggestions of book, authors, etc. are welcomed)

Answer 1

I'll address the general questions:

Many times i hear great pianists employing rubato where the piece has no indication for such thing. ... Where does these liberties come from: is this a matter of interpretation (player's choice) or is there something guiding these performances, like "composer X piece should be played more freely"? And if there are guidelines, where can i learn more about it?

Yes, if "rubato" means "not playing with perfectly strict metronomic rhythm," then this is a liberty that can be afforded to the player. And in fact, often it is not explicitly called for because it is implicitly expected. The sheet music doesn't tell you the audience to clap or you to bow, but these are conventions. In fact, I would guess that any music that expects no rhythmic freedom is more rare. I remember an impassioned lecture by a piano professor titled "Please Don't Play in Time," amounting to "Look, just because the music doesn't tell you how to bend time doesn't mean you shouldn't; rather, it's a clue that you should."

But how much freedom you give yourself, and the particular ways you exercise it, should not just be based on your own whim, but on informed "performance practice." Which is just an official phrase for "How people did it [in a given time, or place, or genre]." Some composers might actually expect absolutely strict metronomic rhythm; I'm not sure Steve Reich has strong feelings on the subject, but I wouldn't attempt his phasing compositions without keeping a steady tempo. Note, the entire notion of being "metronomic" is not a given for all periods of music. I mean, the metronome wasn't even invented before a certain point, of course; but even after that, there was a societal sea-change to embrace mechanization in the 20th century, and prior to that, a "mechanical" tempo might be not only undesirable but un-conceptualized.

The amount and "style" of rubato I use would depend on what I'm playing. In the case of Chopin, there is a lot written about his playing (I think, without bothering to look up documentation, that this is the origin of the term "rubato"?). In particular, contemporaries make note of the alterations of tempo in his right hand, but how the left hand was unaffected by them. Meanwhile, a classical or baroque work might manipulate tempo or rhythm according to slightly different conventions. Another factor that matters a lot is whether you are playing alone or with other people. A baroque solo-instrument "prelude" or "fantasia" explicitly expects a quasi-improvised freeness of tempo, that is perhaps different each time (but often following the same principle as Chopin: The "big beats" are spaced regularly even if what comes between is mutated greatly). But if you play as an ensemble, and especially if your material shares its rhythm with another player, then you can't just do whatever you want on the spur of the moment. (Though, even in a large ensemble, there can be fluctuations of tempo as long as they're agreed on and expected. Some such are ingrained: a waltz band could probably play the Blue Danube, with it's gradual accelerando, without a conductor. Some are familiar rhetorical gestures, like the slight ritardando or fermata going into the "big finish" of a musical number. Similarly, some historical ensemble music might make similar fluctuations without being told, simply because it's "how it's done.")

So how do you find out how rhythm and tempo are handled in the relevant performance practice for your piece? You have to do some research. Ultimately, to be a performer, you also have to be to some degree a musicologist. The most useful kind of research might be listening to other performances, but beware; some might be, themselves, uninformed performances; others might perpetuate a "canon" of practice that is different from the composer's intent. There are certain turns of timing that are expected in Gershwin's Rhapsody in Blue, but the original 1924 recording, with Gershwin at the piano, is lacking in them (as well as taking some radically different approaches to articulation and phrasing):

It also behooves us to read whatever there is to read about our genre, both primary sources and secondary sources who argue over how to interpret the primary ones. This can be problematic; what does it mean when a pre-metronome writer says something is "free"? How free?—but it does at least give us some groundwork.

Answer 2

In the case of Messiaen's works, there are usually recordings either by the composer himself, or by his wife Yvonne Loriod, her sister Jeanne Loriod, or by pupils and proteges (Pierre Boulez, Myung-whun Chung, George Benjamin, Peter Hill, Pierre-Laurent Aimard, ...) that were done under his supervision. These will give you a clear idea of his intentions. Specifically for the Preludes, there is a recording by Yvonne Loriod.

Messiaen can be a challenge for piano players who are used to a 19th century repertoire. During the Messiaen centenary in 2008, touring pianists would add a Messiaen number to their concerts, and I saw Jean-Philippe Collard do this (I don't remember the exact piece), and I thought his interpretation sounded very un-idiomatic and unnecessarily 19th-century-ish.

In Messiaen's works of the 1940's and thereafter, he used the idea of "valeurs ajoutées", where he would carefully notate fluctuations in measure or phrase length by adding note values to measures. There'd be no time signature, and only the vaguest of tempo indications. While this seemingly leaves a lot of freedom to the pianist, it does prescribe perceived changes in tempo very precisely, and mixing in traditional ideas of rubato clashes with the composer's intentions.

(excerpt from Le Merle Noir, 1952)

Answer 3

It's both composer and interpretation determined. For example, there's more room for rubato in Chopin than in Mozart, but there is room in both to allow some fluidity in the pulse.

In the case of the Messiaen, it's understood (by tradition, by being familiar with the composer) to use a great deal of rubato in the Prelude. However, while the Prelude does include some tempo changes (indicated by the composer), the changes in meter — like between 3/4 and 6/8 — are not necessarily changes in tempo.

Messiaen uses the meter changes to indicate changes in emphasis, not speed. For example, in mm. 7–8 the meter changes from 4/4 to 6/8. But the eighth notes remain at the same pace. What changes is that in 4/4 time, every other eighth note receives a metric emphasis; whereas, in 6/8 time, every third eighth note receives metric emphasis. In the case of this piece, no tempo change should be detected, just a shift in meter.