4.1 Introduction: modes of knowing
We now turn to the practicalities of C19th Romanticist historically informed performance (HIP). To begin with, if the radical culture shift between C19th Romanticism and C20th Modernism, and in the arts in general, extends to music performance practice, how could we tell? One of the great advantages of studying contemporaneous Romanticist performance style is the wealth of sources, particularly when compared to other historico-cultural questions. These resources include the following:
-
Romanticist values reflected in other arts and philosophies;
-
Concert programmes, performance reports, performance etiquette reports (such as clapping, cheering, or jeering during performances), and types of performance spaces;
-
Much critical writing on music, especially by Romanticist composers such as Berlioz, Schumann, Liszt, and Wagner;
-
Many pedagogic tutors and lesson notes, most voluminously by students of Liszt; but also including much material from others, such as students of Chopin, Clara Schumann, and Leschetizky;
-
The characteristics of period instruments, many now sensitively restored to good playing condition;
-
Audio recordings beginning from the last decade of the C19th, in three distinct technologies (cylinders, phonograph disks, and reproducing piano and organ rolls).
The first of these sources has already been discussed in detail, while the final source, unique to C19th HIP studies, will be examined in Chapter 5. With the exception of period instruments, the remaining sources are all literary and require comment.
4.2 Written evidence 1: linguistic problems
HIP studies of most eras are particularly reliant on verbal descriptions and explanations, since apart from scores themselves, music ideas have only been able to be recorded in literary form.1 HIP discussion of pre-Romanticist styles therefore remains based on our performance interpretations of verbal descriptions. Despite obvious problems, the assumption has been that such analysis can result in at least probable, applicable performance data, reassembled into coherent performance style. Certain data can clearly be accurately applied to performance, for example the use of repeats, additions, and cuts to scores, and instructions on individual elements, such as normative reduction of note length in articulation, are also reasonably clear. But much else depends on linguistic context and an assumption that much of this has not changed. Can we test this assumption? Recent studies have compared early C20th recordings with contemporaneous C19th verbal descriptions, and the results have called into question the accuracy of a simple, verbally based reconstruction of past styles. Daniel Leech-Wilkinson writes:
But from around 1900 treatises and teaching books on how to play and sing can be compared with recordings – often recordings of their authors … and what we find is devastating to the whole notion of historically informed performance. It would be impossible to come anywhere near the sounds people actually make by following only what they write. Documentary evidence now seems hopelessly insufficient without sound.2
He goes further, casting doubt on the meanings in score notation prior to the C20th, undermining the Modernist idea of simplified Urtext scores:
And the same evidence has profound implications, too, for studies of scores, since their meanings appear to change far more quickly and radically than we could have supposed.3
Neil Peres Da Costa agrees. In summarising his detailed examination of multiple early recordings compared with written instructions, he concludes:
The comparison between early recordings and contemporaneous written texts has exposed striking contradictions time and time again. … In fact, a style of performance based on the [verbal] advice alone would seldom approach the style of the recordings.4
An important example of this apparent disagreement between descriptive texts and recordings concerns unnotated arpeggiation. Peres Da Costa gives a very detailed account of this important topic, including the limitations and contradictions of pedagogic commentary:
In spite of its widespread use around the turn of the twentieth century, detailed contemporaneous written advice about annotated arpeggiation is somewhat scant. Many pedagogical texts fail to discuss it at all. Some consider it indispensable but describe it only in general terms or very briefly. Others advise its extremely judicious employment or absolute avoidance, branding it as a perfunctory device resulting in over sentimentality. [T]here is a significant gulf between written advice and practice.5
For example, the great pedagogue Leschetizky’s own recording of Chopin’s Nocturne Op.27 No.2 includes significant unnotated arpeggiation, yet his own edition of the work indicates none of this practice.6 There is, therefore, a clear conflict between our expectation of style based on written descriptors and the performances captured from the end of the C19th. Can this be reasonably explained?
There are three central problems in converting verbal descriptions into performance style. First and most problematic, there is a lack of precising definitions7; thus, descriptions such as loud/soft, quick/slow, long/short, ‘not overdone,’ and (especially) ‘in time’ assume a known cultural context and do not include the criteria required to narrow the meaning of the terms for those outside this context, for us a century later. The problem is particularly acute in terms of measurement: how much is too much, or not enough? I write this book in York, England, where a sunny day might be described as ‘hot’ if the temperature reaches the mid-20s Celsius; such a day would not be so described in York, Western Australia (where 40℃ is common in summer). An admonition to arrive ‘on time’ in New York, where social time is applied in minutes, has very different limits in rural African villages. To play ‘in time’ equally depends on a largely absent contextualising degree of rhythmic regularity. The problem is found even in the C20th, where playing jazz ‘in time’ might mean the music should be swung to a notationally unspecified degree. In the same way, Moriz Rosenthal stated that Brahms “arpeggiated all chords.”8 Did Rosenthal really mean all chords, or was he making the kind of verbal generalisation common to discourse (“I went to great party, and everyone was there”)? Can we really take his “all” as a mathematical datum? And even if so, was Brahms’ practice desirable, or the pianistic quirk of a great composer? Further confusion comes from a student of Brahms, Florence May, who reported in her lessons (in 1871) that the composer “particularly disliked chords to be spread unless marked so by the composer for the sake of special effect.”9 Again, does she mean all, or even most unnotated broken chords? Was this perhaps a teaching method to avoid excesses?
Second, character and drama indications obviously suffer from the problem of personalised taste within cultural norms. Without contemporaneous, culturally normalised expectations, it is difficult to apply accurat...
