No bigshot, we don't! There's two points here:
1. There was a fair amount written in the C19th, particularly in the form of newspaper reviews. However, think about that for a minute ... A newspaper review is typically half a page, rarely more than 1 page and is not intended to be a detailed analysis but the relatively brief personal impression of a critic. What we actually have is therefore a lot of vague, broad descriptions, that often contradict each other. So yes, we might have some information about the timing for movements for example but relatively little about how that timing was achieved, the timing and other performance characteristics of the sections and individual phrases within those movements. We can make some deductions about some of the precise performance details/styles but they're little more than guesses. Also of course, the "pre-recording era" isn't just defined by the era immediately prior to the recording era (the late Romantic Period), the baroque and classical periods are also "pre-recording era" and we have considerably less knowledge of the performance styles of these periods.
2. There has been a gradual reduction of interpretation freedom throughout the history of classical music performance, regardless of the introduction of recording. In the baroque era for example, it was common practice to completely change the orchestration (which instrument played which parts) and the actual notes themselves. In fact, not all the notes were even notated/written down to start with, notation was in the form of "figured bass": Bass notes with numerals/symbols underneath which implied the other notes of the chords but exactly what notes were played, where, when, how and on what instrument, was largely a matter of interpretation. If you go and buy a score for a piece by Bach, you're actually buying a fully notated interpretation by some scholar who probably never even met Bach. By the classical period, this was not the case. All the notes were explicitly notated for specific instruments and the musicians were expected NOT to change them. The exception was "Cadenzas", which were section/s within a concerto that were not notated, where the soloist performed unaccompanied and could play pretty much whatever they wanted. These cadenzas could last anywhere from about 20 secs to about 10 mins and we have reliable anecdotal evidence that soloists sometimes went off on a complete tangent, performing cadenzas that appeared to be completely unrelated to the rest of the concerto. By the romantic era, this freedom was largely a thing of the past, cadenzas were fully notated, although the soloists still had quite a lot of freedom in how they interpreted the cadenzas. Throughout the C17th, C18th, C19th and well into the C20th, notation and musical "markings" gradually became more precise/exacting and the performers' freedom of interpretation thereby reduced. However even in the baroque era, certain musical intents were extremely specific!
From the above two points: We have relatively little evidence of performances styles. Although there are some exceptions, even going back to the baroque era, mostly we just have rather vague generalisations based on educated guesswork, rather than actual specific details. And, although there's no doubt that recording has reduced interpretation freedom, it's not entirely clear how much of the reduced freedom is due to recording and how much would have occurred anyway, as part of the ongoing evolution of music performance (if recording had not been invented).
I'm not sure I can agree with that. It's hard to argue that any late C19th conductor had anywhere near the power, influence or income of say Karajan.
Same with instrumentalists/soloists. It's hard to argue against Farinelli, a C18th star so massive that the other top stars (and even ruling monarchs) begged for audiences. Handel tried for years, Mozart, Casanova and others travelled for days just to spend half an hour with him. He died fabulously wealthy, loaded with honours from different countries, still a household name throughout Europe decades after his retirement and still cited as the greatest operatic singer of all time even a century after his death. Arguably a better example than Liszt, would be Paganini (early C19th), who really defined and invented the blueprint of the "rock star" for all those who followed, including Liszt, who stated that he wanted to be as great a virtuoso on the piano as Paganini was on the violin. Reportedly (though almost certainly one of the numerous myths that sealed his reputation) upon his death a trunk was found amongst his possessions containing the knickers (underwear) of women who had thrown them at him during performances, 3,000 pairs! An outrageously extravagant lifestyle, an alcoholic by the time he was 16, a self-destructive gambler, a serial womanizer, a badly behaved convicted felon and seriously believed by many to have sold his soul to the devil (because his virtuosity was thought to be humanly impossible), Paganini rivalled or exceeded pretty much every actual "rock star" who followed! Not that Liszt didn't contribute to the "rock star" blueprint, he was wildly popular ("Lisztomania" for example) but far more of the "clean cut" rock star than the Faustian Paganini.
G
So well, that I was fooled for years to think solid scientific justification exists!
