In Praise of Performance Parts

By David C. Birchler

In preparing a new edition of a musical work, the first step is of course to locate and study the surviving source materials. For works of the Renaissance and even for much of the baroque era, one can often pinpoint a single source of the entire work in print or manuscript form. However, as orchestral composition becomes the norm throughout the eighteenth and nineteenth centuries, multiple manuscript or printed scores may be available (autograph scores, copyists’ scores, published scores, piano reductions, etc.), and these scores are increasingly supplemented by performance parts.

Performance parts are often overlooked or placed last in a list of sources. It is understandable that an extant autograph score is considered the primary source for a symphony, concerto, mass, or opera; or that a copyist’s score or published score, particularly one prepared under the composer’s supervision, would be chosen as primary source if the autograph is lost or presents an earlier or indeed superseded version of the work. But as Rey M. Longyear notes in his edition of Stefano Pavesi’s Dies irae concertato (1818; published in 1998 as Recent Researches in the Music of the Nineteenth and Early Twentieth Centuries, vol. 25), surviving parts may be used to supplement the readings of the score:

I have used the score as my principal guide in preparing this edition. The parts have been drawn upon as necessary to clarify readings that are unclear in the score or to supply music not written out in the score. The full score is hastily prepared, with much recourse to abbreviations. . . . When dynamic marks are present, one dynamic often suffices for the entire score, but the indications are usually (not always) transferred to the parts.[1]

Longyear’s statement about the relationship between the source score and parts could be applied to the source situation of many works of Pavesi’s time and later into the nineteenth century. Even in cases where a composer was more detailed in writing out his or her score, the parts still may provide details not present in the score, including—as Longyear references, and as is often the case in music of this era—dynamics. But the way Pavesi applied his dynamics made it possible for Longyear to adopt a policy of replicating them “without comment for all instruments and voices wherever the score indicates an overall dynamic change.”[2]

As the nineteenth century progressed, composers applied more (and more varied) dynamic markings, along with more (and more varied) performance and expressive indications, to their works. Because of this, in editing these later works, it is less viable to apply generalized policies for the editorial addition of marks judged to be missing. This was the case for John C. Schmidt in his edition of John Knowles Paine’s Symphony No. 2 in A Major (Spring), Op. 34 (1879; published in 2010 as Recent Researches in American Music, vol. 70). The source materials for this work consist of the composer’s autograph full score, a published full orchestral score (Boston: Arthur P. Schmidt, 1880), and published orchestral parts (1880). For the premiere of this work in March 1880, the autograph score and a set of manuscript parts were used, with these materials sent off to be published soon afterward. Schmidt comments:

In the main, the notation of the music as presented in the published full score (S) has been retained in this edition. However, the autograph manuscript (A), which is carefully and clearly copied, as well as the printed parts (P), have been drawn upon as needed in cases of error, inconsistency, or missing notation; such changes are reported in the critical notes.[3]

Thus in the first movement, measure 49, violin 2 lacks an appassionato marking in both scores, but this marking is provided in the part. In measure 134 of the same movement, the part provides an accent mark for clarinet 2; this is a case where the two clarinet parts are distinguished in the score with opposing stems, and the composer wrote a single accent only above the upstemmed eighth note of clarinet 1, perhaps with the implication that this accent would also apply to the downstemmed quarter note of clarinet 2. Nonetheless, the part’s inclusion of the accent for clarinet 2 demonstrates how a partbook, being written out for the use of a specific instrument, can provide a detail of performance that is only implied in a score (see example 1).

Example 1. Paine, Symphony No. 2 in A Major, movement I, m. 134, clarinet 1 and 2, with accent in clarinet 2 taken from partbook.

Two parallel passages in the second movement of Paine’s symphony also demonstrate the efficacy of consulting performance parts for editing. In the autograph score in measures 134, 136, 138, 140, and 142, sf markings appear throughout the orchestra but are lacking in the three trombones; in the published score, the markings appear for the two tenor trombones but not for the bass trombone; only the parts provide the markings for all three trombones. When this passage returns later in the movement, in measures 497, 499, 501, 503, and 505, both scores lack the sf markings for all three trombones, while the parts have them for the two tenor trombones, leaving the bass trombone to receive editorial markings. Having access to the parts in these and many other cases allows the edition to be less cluttered with editorial markings, and more importantly, it allows the edition to fully represent the work as it has survived in its source materials taken as an aggregate.

Another work from this period is also instructive with regard to the importance of performance parts as source materials. As Matthew Phelps notes in his edition of Amy Beach’s Grand Mass in E-flat Major, Op. 5 (published in 2018 as Recent Researches in American Music, vol. 84), the composer completed her work in 1889, and a piano-vocal score (abbreviated PVS) was published in 1890 in Boston by Arthur P. Schmidt. Later, a set of manuscript parts (MP), copied from the manuscript full score (MFS), was prepared for the premiere in February 1892. Phelps describes the treatment of the sources in his edition as follows:

This edition is based on MFS and MP, the primary sources. While PVS is the only published source, and while it is evident that Schmidt publishing had a copy of MFS at some point, it does not take into account revisions made by Beach in the vocal parts. We also know that PVS does not reflect the latest version of the work because it lacks additions that were made for the first performance as shown primarily in MP. Therefore, while PVS might be the only published source, it does not reflect Beach’s final intentions regarding the mass. In contrast, MP was made in consultation with MFS, reflects changes made by Beach, and was used for the first performance. As a result, readings in MP that supply notational elements missing from MFS have been included tacitly in the present edition; where there are discrepancies or inconsistencies between MFS and MP with regard to slurs, ties, articulations, dynamics, and accidentals, the edition follows what is judged to be the better reading based on the specific context, with no need of critical notes.[4]

Phelps also notes that the parts “are written in the same hand as the copyist of MFS, with the exception of the viola and contrabass parts, which are in Beach’s hand.”[5] Beach’s close involvement in preparing for the premiere, and the fact that the manuscript parts flowed directly from the manuscript score, all under her supervision, provide support for the decision to consider the parts as equal in status with the score.

One passage of the Kyrie provides ample evidence for the value of treating the manuscript parts as primary source materials. In measures 37–40, the upper winds and strings enter at a dynamic level of pp, while the lower winds and strings have mf as they double the tenors and basses of the chorus. All parts then crescendo to f in measure 41 in support of the full chorus. Both the paired oboes and paired clarinets have “staggered” entries with opposing stems, with oboe 2 and clarinet 1 entering pp in measure 37, followed by clarinet 2 and oboe 1 in measure 38. The manuscript score provides each oboe with its own pp marking, while a single cresc. below the staff (beat 3 of m. 38) suffices for both parts; in the clarinet staff, only clarinet 1 has pp, and again, a single cresc. below the staff (beat 3 of m. 38) applies to both instruments (see example 2).

Example 2. Beach, Grand Mass in E-Flat Major, Kyrie, oboe 1–2 and clarinet 1–2, mm. 37–42 (manuscript score).

The parts, however, provide more detail, supplying oboe 1 with its own cresc. (beat 1 of m. 39; example 3a), oboe 2 with the tie for measures 39–40 (example 3b), clarinet 1 with its own cresc. (beat 1 of m. 39, moved in the edition to coincide with clarinet 2 on beat 3 in m. 38; example 3c), and clarinet 2 with its own pp (beat 1 of m. 38; example 3d).

Example 3a. Beach, Grand Mass in E-Flat Major, Kyrie, oboe 1, mm. 38–41 (manuscript part).

Example 3b. Beach, Grand Mass in E-Flat Major, Kyrie, oboe 2, mm. 37–42 (manuscript part).

Example 3c. Beach, Grand Mass in E-Flat Major, Kyrie, clarinet 1, mm. 37–42 (manuscript part).

Example 3d. Beach, Grand Mass in E-Flat Major, Kyrie, clarinet 2, mm. 37–42 (manuscript part).

As in Paine’s symphony, with the accent for clarinet 2 only supplied in the performance part, the parts here, being written for each specific instrument, provide details of performance that are simply not written out in the score; example 4 shows how they are implemented in Phelps’s edition of the mass for A-R.[6] These examples help to illustrate how a score is in some respects a more generalized representation of the work, existing at a remove from actual performance, while the parts are right there at the level of performance, where each individual player requires his or her part to be fully and completely notated.

Example 4. Beach, Grand Mass in E-Flat Major, Kyrie, oboe 1–2 and clarinet 1–2, mm. 37–42 (A-R score edited by Matthew Phelps).

Getting back to the staggered entries in the oboes and clarinets in Beach’s Kyrie, one might wonder if perhaps the source score is sufficient in its inclusion of the cresc. markings below each staff on beat 3 of measure 38. Why go to the trouble of adding cresc. to oboe 1 and clarinet 1 at all, even if there are cresc. markings in the source parts? One reason is that the staggered entries with opposing stems set up separate streams, so that each oboe and each clarinet needs its own initial pp dynamic, which in turn means that each should also be given its own cresc. marking to make each stream complete. Then there is the important consideration that the score published by A-R is not the final story, because we will be extracting our own performance parts from the printed score, and each part will need its own cresc. marking (example 5 shows A-R’s performance part for oboe 1 in this work). Having these markings already present in the score simplifies the process of extraction and ensures that each part will be complete in this detail and in myriad other details of the work.

Example 5. Beach, Grand Mass in E-Flat Major, Kyrie, oboe 1, mm. 37–42 (A-R performance part).

So, as you prepare your edited score, look at each staff and follow it along, and read it from the perspective of oboe 1, or oboe 2, and note that if you were playing oboe 1, or oboe 2, with only the oboe 1 or oboe 2 part in front of you, you would need your own cresc. marking too, wouldn’t you? Then look to the source score to see what it provides. And then, look to the source parts to see what they provide—because they, taken together in their details of performance, may just provide the most complete realization of the work you are editing.

David C. Birchler is Senior Editor at A-R Editions.