Documenting Folklore in Digital Structures — The Fuzzy Genre Boundary Conundrum

October 24, 2022

How do we make sense of conflicting data to arrive at genre designations that resonate with historical practice? How do we account for overlapping or confusing genre classifications, especially when there may be more than one “correct” answer? How do we organize classifications and typologies in such a way that they are useful for search? These questions were posed in a letter from me to librarian Beth Dwoskin on 06/20/2021 and discussed in a team session at the October 2021 retreat. 

Dear Beth,

…

Someone who put it into the heft/mak spreadsheet described it as a waltz (indicated in the original ms as tempo di vals), to describe the accompaniment texture (we know from research this is plausible for tunes that fall into the dobranotch category). But, it is important to distinguish this potential description of the texture/performance rhythm from the waltz that is used in the cosmopolitan repertoire, and which is the ballroom dance form with which you are familiar.

But. the musical texture could also be zhok — which is found in the title, and which is another plausible texture for tunes of dobranotch category (see brandwein, firn di mekhetonim aheym, forgive bad spelling). The texture could also be the “dobriden/dobranotsch” texture, which is like 1, 2, 3& (down, up, push &). I can see this as a plausible interpretation based on what the tune looks like. 

Meanwhile, the ostensible title is gasn shtikl (zhokul). Gasn nign is a broad category of tunes that are used “on the street.” tunes with this label can have a variety of different rhythmic textures, ranging from waltz, zhok, terkisher, march, khosidl to freylekhs. 

So from this one little tune we have two kinds of dance/non-dance genre classifications:

• Gasn nign
• Dobranotsch

and at least three performance style/texture genres

• waltz
• zhok
• dobriden/dobranotch

Social function genre distinction

• Gasn nign - as a public facing tune
• Dobriden - greeting tunes
• Dobranotch - escort the guests home

with the performance style genres, you can only really perform in one of these at a time, but we need to be able to indicate that there are at least three potential options that are grounded in different kinds of observations about the tune either as title, text annotations, or in the musical structure.

Fig. 1 Mak 1-32-117  Gasn Shtikl (zhokl), found on page 32 of the Makonovetsky manuscript.

Before we begin it should be noted that the reason this piece was identified as a “conundrum” is that it presents a particularly challenging array of information to interpret in order to arrive at a sensible classification. Not every tune in the KMDMP corpus or in the klezmer canon will require this level of deep analysis in order to generate useful genre classification, but it is very clear that robust tools to describe genre and typology are essential for making the klezmer corpus searchable via social function, texture, and musical characteristics.

The questions at the heart of this conundrum are: 

1. If evidence exists for all of these different kinds of classifications, how do we organize this information in a way that is meaningful to a potential user? 
2. Are there “correct” classifications vs. “incorrect” classifications when it comes to assigning genre and typology? 
3. If so, who gets to make those decisions? 
4. How can genre classifications accurately represent a culturally informed perspective grounded in knowledge that is not directly indicated in the source manuscript?  

Klezmer Archive Team Discusssion #

The October retreat session addressed all of these questions, beginning with a discussion of the nature of interpreting data and representing both historical (cultural) consensus and disagreement. 

Figure 2. Historical Fact vs. Historical False. 

Based on the cultural knowledge in the room at the retreat (all but one person present is a klezmer musician), we were able to identify that  while there is a lot of information on the manuscript page, not all of it should be given equal weight in making editorial decisions about how to assign genre classification to the piece. This kind of nuance doesn’t fit well in a technical setting where true/false assertions are the norm, and particularly when a specific assertion doesn’t have associated context such as who made the assertion (and what is their level of expertise) and based on what criteria. As with any data set, a false assertion (whether made by mistake, or because of a lack of knowledge about the topic) can take on an independent “life,” itself becoming the basis for further false assertions.     

With that in mind, how do we arrive at historical/cultural consensus, and how do we use historical/cultural knowledge to assign genre metadata to melodies in our corpus? The group identified several methods:

• Assembling a body of evidence (e.g. academic and other writing, tune books, other corpora, cultural knowledge from practitioners)
• Corroborating evidence derived from the source or other sources
• Relational inferences (e.g. potentially from machine analysis)
• Direct expert commentary on a melody
• Interpretive layers (e.g. a link to an expert-generated description of a genre type)
• Performance as evidence, weighted toward informed interpretations (e.g. a Youtube link to a performance by an acclaimed/respected band)  

Along with some caveats: 

• The presence of information does not equal judgment 
• Context is itself information
• interpretations change/develop overtime, e.g. historical vs contemporary interpretation
• We need an editorial process with different roles for experts and editors in the data ingestion and maintenance pipeline even as we plan to invite the public to create and add classifications and typologies
• Clear definitions for genres and typologies are essential, including information about what is *not* being described. 

The group further discussed the idea from the letter to Beth Dwoskin that there are at least three varieties of genre classifications in klezmer music: 

• By social / ritual function
• By dance (or non-dance) rhythm / texture
• By musicological characteristics. 

Not every tune will have genre distinctions in all three, but some will have multiple genre classifications across in all three of them. Developing a robust set of genre options is essential to making sense of a repertoire whose historical performers were not overly concerned with consistency in naming conventions or “crisp” definition of form and function. Beyond academic and educational concerns this has very practical applications for musicians searching for specific kinds of repertoire. Among thousands of tunes with the title “freylekhs,” for example, it is extremely useful to be able to distinguish those used for “common” circle dancing as opposed to those that would have been used for solo dancing—or to escort the bridal couple from the khupe (fun der khupe)

The group investigated the “conundrum” piece through the lens of our own experience as klezmer musicians. In essence, we worked together to evaluate the melody to determine valid/valuable genre classifications by looking at:

• the music notation on the manuscript
• textual information on the manuscript 
• cultural knowledge about genre, performance style, and social context  

Figure 3: Analyzing Mak1-33-117 Gasn Shtikl (zhokl)

Figure 3 illustrates our discussion of “performance texture”¹ from looking at various pieces of evidence found on the score manuscript. The following paragraphs will illustrate the thought process captured in the sticky notes. The discussion from the meeting has been augmented with references to some of the leading writers on this topic (Moshe Beregovski and Walter Zev Feldman), and condensed for clarity. 

Is it a Zhok?  #

Starting from the bottom, the title includes the word “Zhokl.” Since the time signature is 3/4, and since the word “Zhokl” could be a Yiddish diminutive of the Jewish dance form “Zhok,” it is plausible that the title is indicating that the piece should be performed using the “zhok” rhythm. Alternatively, since the Romanian term “joc” refers to any tune played for folk dancing, the term could simply be a Yiddishized version of that term that became common in the region where Makonovetsky lived. The Jewish zhok rhythm has a strong beat on one (“one”) and a weak beat on three (“and”) and is usually (though not always) represented as either  3/8 or 6/8 time signature. In this rhythm, the second beat does not generally figure in the rhythmic accompaniment and is not usually prominent in the melody, (e.g. One + and, One + and, etc.)².  The group came to the conclusion that title and time signature notwithstanding, the melodic structure of the piece does not fit with the zhok accompaniment, and that the melody doesn’t naturally fit with the phrasing implied by that texture. Furthermore, additional text notes with the melody (see below) make it clear that this is not a tune for group dancing.

Conclusion: Not a Zhok

Is it a Waltz?  #

On the middle thread we gathered evidence that pointed to the concept of “Waltz.” On the manuscript we find two tempo notes: “tempo waltz,” and “allegro.” Because the time signature is 3/4 the question naturally arises: Is this melody a waltz?  There are other clues about the social function of this melody that complicate the seemingly natural waltz categorization. The first is the title Gasn Shtikl which indicates that it is a piece (shtikl) to be played on the street (gas). The text notes at the top of the melody give further information about how the piece was used:  

“A ‘Gasn Shtikl’ (street tune, zhokl).  With this nign, customarily/often, escorting the inlaw[s] back home after the night, in the day [dawn],  [filled (tipsy) with drink  as they dance joyously to the beat of the musicians playing the nign] which they call one-two-three ….”³

This text places the melody squarely into the realm of a specific, well-known Jewish wedding custom: escorting the in-laws home at the conclusion of the wedding feast (Firn Di Mekhutonim Aheym). Thus, the tune has a different social function than entertaining guests at the wedding either as a waltz for dancing or for listening.

That said, since Makonovetsky includes tempo markings of allegro and tempo waltz. If we know that the social function is different from a typical setting for either a listening or danced waltz, what are we to make of these instructions? Elsewhere in his manuscript Makonovetsky writes: 

“This waltz is to be played according to its character: everybody understands where to go out of strict time, where to speed up, where to play softly and where to become stronger. Because I had no printed music… And in the same way the mazur is also played naturally as one understands it.”⁴

In this quotation, Makonovetsky seems to refer to waltz melodies that derive primarily from western ballroom/social dance that were used for both listening and dancing. Indeed, the melody this note refers to is a multi-page, multi-section waltz that is clearly grounded in a “western” social dance compositional style and that was gifted to Makonovetsky in 1880 by a klezmer from Warsaw (composer unknown). One possible interpretation is that Makonovetsky is indicating that the tempo should be lively, to match the “joyous dancing” of the inlaws, but that as with other waltzes, the melody can be phrased to speed up, slow down, and otherwise accommodate the natural phrasing in the piece. In other words, the piece is not a waltz in social function (i.e. to be performed by couples on a dance floor), but should be played with a waltz-like tempo and phrasing.  

Conclusion: Not a Waltz 

Is it a Dobranotsh?  #

The top line of sticky notes speaks to the process of situating this melody into several overlapping categories of ritual melodies as defined by important scholars of klezmer music: “Street tunes,” “Greeting/escorting tunes” (muzyka dlia slushania/music for listening)⁵,  and moralishe nigunim (tunes of high moral character)⁶.  Street tunes (gasn nign) refers to melodies that are performed in public spaces, e.g. “on the street,” and can include melodies used for the wedding (e.g. greeting and escorting) or for other public purposes. Greeting/escorting tunes, or muzyka dlia slushania (music for listening) are described by Beregovski as

 “Compositions that were performed at the wedding in greeting and accompanying the guests, during the ceremony of the kale bazetsn (seating the bride), under the wedding canopy (khupe) and the procession leaving it, at the table, etc.”⁷

Walter Feldman refers to the non-dance Jewish wedding repertoire as Moralishe Niggunim, a term used by Israel Rabinovitch in his 1940 book, Muzik Bay Yidn (Monreal: Eagle, 1940) and he further reports that the term is still used by Hasidim in Israel⁸. 

There are no specific shared musicological characteristics that define tunes across these broad, overlapping categories (though some musical characteristics can be distinguished within the subcategories)⁹. However, according to Walter Feldman’s definition of moralishe nigunim and his four-fold genre classification system, the most important distinction between a waltz, a zhok, and any tune in the category of Gasn Nign, Greeting Tune, or Moralishe Nign is that these melodies are primarily comprised of Jewish musical thinking and are part of the “core” Jewish repertoire rather than deriving from cosmopolitan (waltz) or Moldavian (zhok) sources. As Feldman concludes his chapter on Moralishe Niggunim: 

“In these wedding genres, the slower triple time seems to have conveyed the appropriate dignity for the seriousness of the wedding day, and elicited another more inward-looking response, what Rabinovitch described as batrakhten, the “meditation” into which the early morning onlookers were plunged as they heard the sounds of the dobriden resounding outdoors at sunrise. All of these uses seem to have come together in the klezmer concept of the moralishe niggunim, the melodies of high moral character, which were not purely liturgical but which adapted the musical bases of prayer to the worldly reality of life, marriage, and community.” ¹⁰

Returning finally to the Gasn Shtikl, an examination of the harmonic movement in the first section of the Gasn Shtikl reveals what is known as the “Sim Shalom Maneuver,” which is found in the musical nusach of Jewish daily prayers ¹¹. And, though the melody uses more chromaticism than is typical, it is grounded in the Freygish (Ahava Rabo) mode rather than any obvious Bessarabian (zhok) or Western (waltz) compositional technique. 

Conclusion: It is a Dobranotch 
Conclusion: Dobranotch is a subcategory of Gasn Nign
Conclusion: Dobranotch is a subcategory of Greeting/Escorting tunes (muzyka dlia slushania/music for listening)
Conclusion: Dobranotch is a subcategory of Moralishe Niggunim

Conclusions  #

What are we to take away from this discussion? The process of making distinctions and evaluating tunes requires a great deal of cultural and musical knowledge, and while not every tune in the KMDMP corpus (or klezmer music writ large) is this challenging to interpret, many are. Archivists with culturally informed domain expertise have the ability to assign authoritative genre classifications at the same time that it is  also possible for other experts to contest or complicate a given classification. The proposed klezmer archive structure will allow for both of these instances. Thus, the exercise of attempting to classify tunes that have potentially conflicting evidence in the source is a way to add to our overall understanding of klezmer music and the Ashkenazic culture that it served in the past and that continues in practice today. 

Some conclusions:

• genre classifications that incorporate a culturally informed perspective grounded in knowledge not directly indicated in the source manuscript is a way to make that cultural knowledge visible in the data. 
• Trained archival specialists with deep domain expertise will be necessary to create complex entries and to edit/verify more simple entries. 
• Trained archivists can make authoritative decisions about classification.  
• Commentary and discussion space associated with entries will allow for interrogation and contestation of editorial classification choices. 
• Genre and typology definitions will need to be crisp where possible, but acknowledge that not every melody will fit all characteristics of a defined genre. 
• A comprehensible set of genre classifications and typologies will allow users to search for melodies that have social, dance, or ritual functions that may not be revealed by the tune title, time signature, or modality. 

Categorizing the Gasn Shtikl, some preliminary metadata. #

Title: Gasn Shtikl
Alternate Title: Zhokl
Tempo: Allegro, Tempo Waltz
Time Signature: 3/4
Tonic: A
Key Signature: none
Primary Modality: Freygish
Melodic Range: A4-B5
Number of Measures: 32
Number of sections: 2
Section length: 16
Section repeat: yes
Pickup measure: yes
Part Structure: AA BB AA
Beregovski Category: muzyka dlia slushania
Feldman Category: Core Repertoire
Social function: Dobranotch, Gasn Nign, Moralishe Nign, greeting/escorting tunes 
Texture/Rhythm: Dobriden, dobranotch
KMDMP Number: Mak1-32-117
KMDMP Notator: Hannah Ochner
KA Number: ???

\\\///////////\\\\\///////////\\\\///////////\\\