Tempo as Form

By Nathan Pell (The Graduate Center, CUNY and Mannes College)

This past Fall, I presented at a nineteenth-century performance practice conference at the Sydney Conservatorium, convened by Neal Peres da Costa, Daniel Yeadon, and Clive Brown.  (Fellow PAIG member Jocelyn Ho also gave a paper, on Debussy’s performance style, that showcased her truly stunning piano playing.)  The theme of this symposium was “Correct, but not Beautiful, Performance,” a phrase borrowed from treatises by Hummel and Spohr.

The idea, these treatises say, is that beginning music students must first learn to execute notes, rhythms, dynamics, articulations, and phrasings with exactitude.  This sort of discipline sits at the core of Western musical training, and lays the groundwork for a student’s solidity in musicianship, technique, and ensemble playing.  Performing the music precisely as written in the score—the principal goal, I and the symposium’s participants would argue, of most modern-day musicians—is only the first step, however.  It is “correct, but not beautiful, performance”:  for composers can only notate so much, and to play a dry rendition of the score, therefore, would be to miss their point.

Beauty in music-making, for musicians of two hundred years ago, required performers to read between the lines, intuiting the unstated implications of the printed page in accordance with a shared awareness of musical practices and styles, so as to tease out what elements the composer might have left out of the score.  To be sure this is an interpretive process, but not really in the modern sense that performers ‘make their mark’ on the music.  Rather, it was understood that composers really wanted to hear such liberties in performance, and as such a certain spirit of fidelity existed even in especially free realizations of the score.  So, the conference organizers were quite right to charge participants with “deciphering the hidden messages in nineteenth-century notation”;  for this is the same activity that sensitive nineteenth-century performers engaged in—albeit with the advantage of living and breathing these performance practices rather than trying to reconstruct them.

The result was a fascinating symposium that brought together musicians of many stripes and included a large number of performances by participants.  Several of these, easily among the best “historically informed” performances I have heard, will stay with me for quite some time:  Ho’s aforementioned Debussy, Peres da Costa’s performance of John Field in the manner of Carl Reineke, Koen van Stade’s 1870s-style rendition of “O Isis und Osiris,” and an open rehearsal/coaching of the Brahms Horn Trio with Anneke Scott, Robin Wilson, and Peres da Costa.  So I was thrilled to learn, a few days ago, that a follow-up conference is planned for this September in Vienna.

My own contribution to the symposium was a paper called “Tempo as Form:  Orchestral Recordings from 1910–1940 in Light of Earlier Sources.”  (I have since delivered versions of this paper at several American theory conferences.)  Especially since the topic touches on questions of analysis, I think it will be of interest to PAIG members.  This medium does not give me space to rehash my full argument, but I will summarize some main points.

 *          *          *

Many musicians believe that tempo fluctuation in orchestral performance began with Wagner.  And, further still, some think that rubato in general is a trapping of “Romantic,” but not “Classical,” performance.  I stressed that I consider these among the major misconceptions in music historyAmple evidence surely exists that eighteenth-century tempos were flexible, and that this was a sine qua non of “beautiful” performance.  The most commonly discussed sort of rubato is that in which the performer speeds up in vigorous, loud passages and slows for tranquil or soft ones:

In compositions whose character is vehemence, anger, rage, fury, and the like, the most forceful passages can be played with a somewhat hastened motion…  For extraordinarily tender, longing, or melancholy passages, in which the emotion, as it were, is concentrated in one point, the effect can be very much intensified by an increasing hesitation.  (Türk, 1789)

But in the eighteenth century such flexibility was mostly confined to solo and chamber music, for reasons of ensemble:

Certain purposeful violations of the beat are often exceptionally beautiful.  However, a distinction in their use must be observed:  In solo performance and in ensembles made up of only a few understanding players, manipulations are permissible which affect the tempo itself; here, the group will be less apt to go astray than to become attentive to and adopt the change; but in large ensembles made up of motley players the manipulations must be addressed to the bar alone without touching on the broader pace.  (CPE Bach, 1753)

(At the end of this last sentence Bach makes an implicit distinction between tempo rubato‘s original meaning—where a melody weaves in and out of tempo against a steady accompaniment—and an actual fluctuation of the main pulse across a musical texture.  As time went on, the term tempo rubato began to refer to the latter practice too, causing a bedeviling confusion in terminology.)

Exceptions were rare enough to earn special mention.  But by the early nineteenth century, when the capabilities of Europe’s orchestras were growing to match the demands composers were placing on them, this was no longer the case:  CPE Bach’s rule underwent reformulation, and the differences between solo and orchestral tempo styles began to fade.  Indeed, important composers are known to have varied the tempo in their orchestral performances, and treatises began to single out specific moments in orchestral works that call for rubato.

For an example, click here to look at the instructions that Anton Schindler added for the Larghetto of Beethoven’s Second Symphony while hearing a performance that follows them.


If you know of Schindler, it is probably for his many forgeries and outlandish claims surrounding the life of Beethoven.  But his testimony here must be reliable, given its full-throated endorsement by Ignaz Moscheles—a far closer friend of Beethoven’s than Schindler was and, moreover, a rather conservative musician:

I agree with M. Schindler in these remarks.  The slight deviations of time recommended must give life and expression, not only to this movement, but also to the imaginative compositions of all the great masters.  Their success, however, can only be assured by intimate acquaintance on the part of the band with the manner of the conductor, and his mode of conveying his intentions, either from long intercourse or careful rehearsals.

Carl Czerny, another prominent pupil of Beethoven’s, provided the fullest account of tempo flexibility in the first part of the nineteenth century—and it is very likely that his precepts reflect the performance practice of his teacher.  Below I have reproduced his advice for the use of the ritardando, interspersing links to a number of early recordings (mostly orchestral performances from before 1940) that, to my ear, exemplify Czerny’s discussion.  I do this not only so we can more easily grasp what Czerny is getting at, but also to make the case that the practices described in early nineteenth-century treatises can be heard to have survived into the twentieth century.  This should come as no surprise, considering that most of these musicians were born and educated in the nineteenth century.  (I have deliberately chosen recordings that display a wide array of approaches, from the subtle to the ridiculous:  there have always been stiff performances and over-the-top performances, and the nineteenth century was no exception.)

§8 The Ritardando, according to the generally established Rule, is much more frequently employed than the Accelerando, because the former is less likely to disfigure the character of the piece, than the too frequent hurrying on in the speed of movement. We may retard the time most advantageously.

a.  In those passages  which  contain the return to the principal subject.
b.  In those passages, which lead to some separate member of a melody.
c.  In those long and sustained notes which are to be struck with particular emphasis, and after which quicker notes are to follow.
d.  At the transition into another species of time, or into another movement, different in speed from that which preceded it.
e.  Immediately  after a pause.
f.  At the Diminuendo of a preceding very lively passage; as also in brilliant passages, when there suddenly occurs a trait of melody to be played piano and with much delicacy.
g.  In embellishments, consisting of very many quick notes, which we are unable to force into the degree of movement first chosen.
h.  Occasionally also, in the chief crescendo of a strongly marked melody, leading to an important passage or to the close.
i.  In very humorous, capricious, and fantastic passages, in order to heighten the character so much the more.
k.  Lastly, almost always where the Composer has indicated an espressivo; as also
l.  At the end of every long shake which forms a pause or Cadenza, and which is marked diminuendo.

Let’s take stock.  Recall the eighteenth-century paradigm described by Türk:  vigorous/fast, tranquil/slow.  The place one is most likely to find tranquil music in a sonata form is the second theme.  We know that performers in the nineteenth century would have slowed here.  Sure enough, so did performers on early recordings, where the second theme is usually one of the slowest passages in a sonata movement.

Listen:  Beethoven, Symphony in E-flat, Op. 55, First Movement—Henry J. Wood, New Queens Hall Orchestra, 1922 (first recording of the work)

Blog Examples

The music just before the second theme, the transition, is normally quite vigorous—as we might expect given its harmonic purpose.  This too is borne out in performance; the transition often contains some of a sonata movement’s fastest music.

Listen:  Mozart, Symphony in E-flat, K. 543, First Movement—Felix Weingartner, Royal Philharmonic Orchestra, 1928

Blog Examples3

Czerny tells us (in Letter A above) that one should slow before thematic restatements.  The biggest of all thematic restatements is the sonata’s recapitulation, and performers almost always heed Czerny’s advice.

Listen:  Beethoven, Symphony in D, Op. 36, First Movement—Clemens Krauss, Vienna Philharmonic Orchestra, 1929

Blog Examples2

From these principles we can begin to construct a generic tempo plan that a typical sonata movement might follow.  Although by no means typical in most other regards, Mozart’s G minor Symphony demonstrates most of the standard tempo conventions of the high classical sonata:

K. 550--Strauss.jpg

Click here to listen along with Strauss’s famous recording while looking at the chart.  Notice how closely Strauss’s tempos match up with our familiar formal units.  This is what I mean by my title, “Tempo as Form.”

How these practices changed across the twentieth century (and change they very certainly did!) is beyond my scope here.  But because the eighteenth and nineteenth centuries had conventional ways of varying tempo in response to specific musical situations, we can surmise that composers wrote their music with these tempo variations in mind.  (In the full paper, I discuss in detail what these specific musical situations were.)  If, indeed, tempo was written into the music in such a way, we would do well to consider tempo a truly form-defining feature, much like harmony and thematic design.

As always, comments are welcome.  Share your thoughts below!


CFP for 2019 Meeting

The SMT Performance and Analysis Interest Group (PAIG) invites proposals for 13-minute lightning talks at its upcoming meeting during the 2019 Society for Music Theory conference in Columbus, OH (November 7–10).

Proposed talks may relate to any aspect of the relation between music performance and theory or analysis. Proposals should reflect that the presentation wil lbe suitable for a lightning-talk format and that it will incorporate recorded or live music in some way. Presentations involving collaborations between scholars and performers are encouraged. Authors may submit only one proposal for consideration.

To submit a proposal, please do the following by July 7:

  • Prepare a PDF file including (1) a title, (2) an proposal of up to 500 words, and (3) up to 2 pages of examples and/or bibliography. The document (including its filename and metadata) should be anonymous, with any self-references in third person.
  • Email the PDF file to edward.klorman (at) mcgill.ca, using subject line “2019 PAIG Proposal.” In the body of the email, please identify the author(s) of the proposal.
  • Our conference room will have a standard AV set-up. If your presentation requires any special equipment beyond this, please describe it in the body of your email. Because of PAIG’s budgetary constraints, you will likely have to make such arrangements yourself.

By submitting a proposal, you agree to the following if your proposal is accepted:

  • To register for and attend the upcoming SMT conference (November 7–10, 2019 in Columbus, OH).
  • To provide PAIG with a short (500–750 word) “teaser” blog post by October 1, to be written in a reasonably non-technical and accessible style. This could simply adapt your proposal or could be a newly composed text. (PAIG blog staff will assist those unfamiliar with the medium.) We will use this blog post to promote our meeting and to allow PAIG members the opportunity to engage with your ideas or listen to recordings in advance of the session.

Program Committee:

  • Ellen Bakulina
  • Wing Lau
  • Edward Klorman (ex-officio)
  • Nathan Pell (ex-officio)

Reflections on the 5th International Performance Studies Network Conference

Lee Blasius recently instructed a program committee:  “One question we don’t need to ask is, ‘Is this music theory?’  The field is an ever-expanding horizon.”  On the surface, many of the presentations and performances given at this summer’s 5th International Performance Studies Network Conference, held at the Norwegian Academy of Music in Oslo, presented analyses that defied or ignored the conventions that many might associate with music theory (as well as composition, musicology, and performance studies).  Yet the multiple perspectives, disciplines, and topics that attendees encountered in Oslo this July should encourage music theorists, in the spirit of Blasius’s instructions, to think more expansively about what music theory and analysis can do, for whom they operate, what tools and methods they might employ, and what new voices to include.

As always, the parallel sessions prevented those of us in attendance from reporting on the entirety and richness of the full program.  There were, however, highlights that we would like to share with our colleagues.


Introduction: David Kopp (Boston University)

The Performance Studies Network (PSN) is a legacy of the Center for Musical Performance as Creative Practice (CMPCP), a five-year initiative based at the University of Cambridge led by John Rink.  CMPCP’s tradition of biennial conferences has continued under the auspices of PSN; this was the fifth in the series, and the first to be held outside the UK.  Darla Crispin, director of research at the Norwegian Academy, was the conference organizer.  The growth and maturation of the field of musical performance studies and the continuation of CMPCP’s mission, framed within the European practice of artistic research , were evident throughout the conference.

The guiding philosophy behind CMPCP/PSN is that musical performance, and by extension music, must be understood as more than the realization of the composer’s vision and its representation on the score.  The performer’s contribution constitutes an essential, active, and creative element of music, thus providing a compelling object of study.  Moreover, because performers are recognized as thoughtful, expert practitioners—not mere channels for a composer’s intent—they are uniquely placed to share in the research endeavor.  Not surprisingly, a majority of conference presenters wore multiple hats:  active equally on the one hand as performers, conductors, composers, educators, and critics, and on the other as scholars (whether in traditional academic fields or as specialists in artistic research).  Many of the performer-presenters were of considerable distinction.

The presentations reflected several topics central to musical performance studies.  These included, among others:

  • Performance practice, from established approaches to historically informed performance to the developing practice of the 21st-century music
  • Determining and articulating performers’ contributions to the realization of music beyond the information on the score, both individually and in collaboration
  • Composer-performer collaborations in which both contribute to the creative product
  • Improvisation
  • Analysis of recordings
  • Anthropological, sociological, and cultural aspects of performance
  • Interdisciplinary and cross-cultural approaches to the development of performance strategies
  • The influence of technology on performance practice
  • Approaches to music production
  • And, last but not least, despite a current of skepticism directed at the privileged position of the score, several talks drawing on the performance and analysis tradition, given by both North American and European presenters.

There were also several concerts given by Norwegian artists, ranging from avant-garde music to music inspired by Norwegian vernacular traditions.  Many of the presentations gravitated toward new music, where the contribution of the performer is often more explicitly acknowledged or readily ascertained.  An equally important goal, though, was evident in other presentations that aimed to demonstrate the unacknowledged and essential contributions of performers to music of the European canon.

The role of music theory in the PSN enterprise was most apparent in, but not limited to, the presentations that involved performance/analysis questions or the analysis of recordings.  If the conference raised any challenges for North American theorists, they might be to better recognize and incorporate those beyond-the-score aspects of performance into analysis, and, through the force of future work, to expand the presence of music theory within the performance studies community.


Daniel Barolsky (Beloit College)

Daniel Leech-Wilkinson, in “Aspects of belief and attachment in the performance of classical music,” equated our deeply rooted reliance on compositional instructions to a kind of religious devotion.  Doing so, he drew attention to how normative processes, especially those of performance style, are not only created but also policed.  Attempts to deviate from these norms have been regulated by critics (and theorists, I would argue) in a manner that not only discourages interpretive creativity but also perpetuates, in euphemistic ways, outdated gendered binaries, homophobia, and misogyny.  Although Leech-Wilkinson focused on music criticism, his larger argument can be extended more broadly to “the heteropatriarchal values that underlie western classical music belief.”

Many of the presentations in Olso offered ways to combat these limits.  The most activist strategies could be found in Challenging Performance, a new project created by Leech-Wilkinson and his colleagues that seeks to challenge traditional practices and gives voice to entirely new ways of analyzing and hearing familiar repertoire.  Analogously, Rosanna Lovell, Brandon Farnsworth, and Darla Crispin led a panel conversation on GRiNM (Gender Relations in New Music), a relatively new organization (2016) that confronts the normative mechanisms that have resulted in such major classical music festivals as Darmstadt programming a disproportionate number of compositions by men.  Readers will find it well worth their while to explore both of these websites and to take note as they develop.

In work that embraces more radical interpretations, Lina Navickaitė-Martinelli and Georgia Volioti introduced new analytical lenses through which we might explore new modes of listening.  Navickaitė-Martinelli, in her “P is for Person, Performance, Pogorelich:  Performer’s Identity as Creative Tool,” used a comparison with a slow cinematic take by Michelangelo Antonioni to help reimagine how we might listen to a hypnotically drawn-out interpretation by the controversial pianist Ivo Pogorelich.  However simple the comparison, many of us in the room, previous critics of Pogorelich, came to recognize how the pianist’s tempo, however unconventional, gives the listener space and time to explore—almost like a temporal X-ray—the inner workings, dynamics, texture, and make-up of a given score, details that would have been lost at a faster tempo.  Similarly, Volioti (in “Narrativity in Grieg’s Ballade Revisited:  The Nineteenth-Century Pianist as Storyteller”) used recorded performances by Eugene D’Albert and and Percy Grainger to conceive of the narrative possibilities of Grieg’s Ballade, a theme and variations that had appeared to resist or fail to convey narrative construction.

Cecilia Oinas and Naomi Woo explored the interrelationship between analysis and embodiment, thus expanding the domain of analysis even further.  In “The Body in Pain at the Piano:  Where Form Meets Failure in Ligeti’s Etudes pour Piano,” Woo presented an “autoethnography” of the author’s process of learning (and at times failing to learn) the Etudes.  She provided us with the critical reminder that any analysis of a score as a “work” will always be “radically incomplete” if we limit our engagement to just reading and listening.  In a discussion of Ligeti’s Étude no. 3:  touches bloquées, Woo drew attention to the cognitive dissonance that occurs not only between what we see and what we hear, but also between what performers hear or feel (including pain) and how they play or practice.  A similar cognitive dissonance was explored by Oinas’s paper, “From four-handed monster to an all-embracing Vishnu:  On sensitivity, intimacy, and corporeal interaction in György Kurtág’s four-handed works.”  Because the negotiation of bodily space is central to any four-hands successful collaboration, such music is more inherently physical than other duos.  Not only do Kurtág’s compositions intentionally force the performers to cross over and invade each other’s space—a critical and rarely analyzed element of this music—but because both musicians are playing the same instrument, what the individual performer plays is often confused with what she hears.  The author will be presenting an expanded version of this paper the upcoming SMT conference in San Antonio.


Michiko Theurer (Stanford University)

Of the presentations I was able to attend, the one I keep coming back to is a multi-modal performance/lecture given by Anna Scott and Valentin Gloor, entitled “Brahmsphantasie:  Performing Historical Fictions.”  Their presentation posed questions about how historical fictions and contradictions can be used as tools for analysis and interpretation of canonical music, using as an example Max Klinger’s Brahmsphantasie Op. 12, a fantastical creation that led Brahms to reimagine the potential of his own music.  When I entered the room, the lights were dimmed and Gloor was singing Brahms as he climbed over and between chairs in the audience, while Scott played the piano at the front of the room.  The presentation unfolded in a series of seamless transitions between musical performance, scholarly lecture, and theater, sometimes layered in such a way that it was difficult to untangle one from the other.  At one point, Scott compared Klinger’s Brahmsphantasie with Baroque emblem books, explaining that the multiple and conflicting layers in the emblem books forced the viewer to interpret or choose between possible readings.  As she spoke, I kept getting distracted by the sound of Gloor crumpling pages of music in the back of the room.  Why on earth was he distracting from what she was saying?, I kept thinking, until I realized this forced hovering-between-ideas (schweben) was precisely Scott’s point.

Toward the end, Scott announced that at a previous conference (Orpheus Institute 2015) they had presented a paper concerning an entirely fabricated invented correspondence between Brahms and Debussy and the (alleged) resulting music.  In exposing this falsehood, she asked:  did she also in some way kill truth?  The tone of her question was extremely complex, and loaded with the intellectually rigorous and musically persuasive power of their presentation.


Victoria Tzotzkova (Massaschusetts Institute of Technology)

When I think about what gives the PSN conferences their distinct flavor, one particular point stands out:  that often there is a palpable sense that the person speaking is a performer.  These are still clearly scholarly presentations, sometimes more historical in approach, sometimes more sociological, psychological, experimental, and so on; but the questions being posed and the ways these questions are treated are, somehow, essentially performerly, to borrow a favorite term from Mine Dogantan-Dack.  In other words, these are questions that come up in the course of practicing, rehearsing, and performing—they are fundamentally relevant to experiences in performance.  Rather than asking performers to do scholarly work in the ways that academic music disciplines already do their work, these conferences seem to have become platforms for research that is not typically native to either the academic music department or the conservatory.

This is not to imply that this research merely consists of the incidental insights performers glean in the course of practice—the work is typically done in depth and with rigor.  But the questions raised seem to be specifically directed at performers, often aim to identify and analyze a particular aspect of performance experience, and generally tend to be readily subsumable back into performance experience.  I found this element of a performer’s voice both engaging and compelling, and I want to briefly trace it in three presentations.  (As a pianist, I gravitated towards more-or-less piano-centric talks.)

Mark Ferraguto’s presentation (“Interior Virtuosity in Beethoven’s Fourth Piano Concerto”) drew on letters, criticism, and philosophical writings from Beethoven’s time and current analytic work to highlight a then-important conversation about distinguishing between real and fake artistry in virtuoso performance.  While I cannot claim to have retained all details of the discussion, the points that stood out for me include the following:  in the context of that distinction, expressive markings can become a way to signal the depth of inner experience; that also within this context emerges the polarity between brilliant vs. expressive virtuosity, both of which can be part of real artistry.  These two styles can become a particular lens through which to analyze and experience specific passages of Beethoven’s keyboard writing, especially in a work like a concerto, where the artistic personality of the composer-soloist is explicitly on display.  Ferraguto drew our attention to passages that exemplify each category, and suggested that knowing about these categories might affect how such passages are treated in performance.  It was exactly that last point that bespoke a performer’s outlook to me:  this understanding of virtuosity—carefully researched and laid out—was offered not only as a piece of historical research but also as an attitude one might adopt in playing the piece.  If, through the many variants of the solo part to the Fourth Concerto, Beethoven was negotiating the full range of his artistic persona, from the brilliant to the expressive, then that same attitude of showcasing a range of inner experience and pianistic prowess may also be productive today.  Ferraguto did not explicitly make this point, but in offering these ways of understanding virtuosity, he certainly opened the possibility.

In “Beautiful Piano Tone–A Matthay Legacy?,” Julian Hellaby presented the findings of an experimental study he undertook which aimed to establish whether pianists in the Tobias Matthay legacy were any more likely to be perceived as playing with “beautiful tone” than pianists who were never in contact with Matthay.  (Being a strong proponent and life-long theorist of beautiful tone in piano playing, Matthay is an obvious choice for questions about this elusive concept in piano performance.)  While the results were personally fascinating to me, the main point I would again like to make is that the question the study asks is fundamentally pianistic and performance-centered.  In the Q&A session, Hellaby used a short Romantic excerpt to demonstrate at the piano an aspect of how he understood beautiful tone.  The very understanding of the concept is rooted in experience at the piano, and sharing that understanding seems to almost require a demonstration, again at the piano.  With a strongly expressed scientific outlook and allegiance to proper experimental design, the study also remains clearly pianistic and steeped in experience:  it treats a question that a performer would ask.

Tor Espen Aspaas’s presentation (“Unfolding Beethoven Extempore”) focused on passages in Beethoven’s piano writing which bear the mark of extemporizations.  Again motivated by performance experience, Aspaas’s research aims to create an improvisational framework in which a performer could experience portions of Beethoven’s oeuvre.  The difference this switch could make is something Aspaas discovered first-hand, having recorded the full Beethoven sonatas cycle twice:  first working from traditional methods for interpretation, then more recently from this improvisational perspective.  There was not sufficient time to spend with his recorded examples, but in his own judgment the extempore framework not only offers a sense of greater freedom in performing the works, but also sounds fresher, more spontaneous, and more compelling than the earlier interpretations.  Again, this is a long-term research project, hosted at the Norwegian Academy and spanning several years, which springs from performance experience and, once developed, aims to directly impact performance experience.


2018 PAIG Meeting at AMS/SMT

“Since Schnabel: Pondering Hypermeter in Beethoven’s Piano Sonatas”

Presented by William Rothstein (The Graduate Center and Queens College, CUNY)

Friday, November 2, 12:30–2:00

** New Location: Lone Star B Crocket CD

Our upcoming meeting will take the form of a 45-minute presentation (see abstract below) followed by an extended analytical discussion.

For those who wish to do a bit of “pondering hypermeter” in advance of the meeting, Prof. Rothstein has generously prepared some study materials about the first movement of Beethoven’s Sonata in E-Flat, op. 33, no. 3. These materials may stimulate your thinking and serve as food for thought for the discussion. Click here to download them in PDF format

We wish to emphasize that this preparatory material is completely optional. Everyone is welcome at the meeting and in the discussion, regardless of whether they have perused the preparatory materials. A handout will be available for the talk, and it will be possible to follow without having done any preparation.

We hope to see you there!


Abstract: According to his pupil Konrad Wolff, Schnabel said that when he began to play a passage, he needed to know how far away the end was. Hence the “metrical periods,” as he called them, that he marked in his edition of Beethoven’s piano sonatas, first published in the 1920s. Unfortunately, he never defined very clearly what a “metrical period” is.

Many performing musicians have felt a need similar to Schnabel’s: how to feel, or count, Beethoven’s rhythms of medium size (3–16 measures). The terms “meter,” “metrical period,” and “hypermeter” have been used by many, but the same term often conceals different meanings, as John Paul Ito has rightly pointed out. In this talk, I consider the views of several writers since Schnabel, from Tovey to Temperley to Ito. Excerpts from most or all of the following Beethoven movements will be discussed: op. 28, i; op. 31/3, i; and op. 90, i.

Glenn Gould’s Uncommon Approach to Alban Berg’s Piano Sonata

By Michael Thibodeau (University of Toronto)

Author’s note:  The following adapts material from the second chapter of my dissertation, which can be found here.

Glenn Gould’s relationship with Alban Berg’s Piano Sonata is often overlooked.  Indeed, he recorded the work eight times between 1952 and 1974 in a variety of circumstances:  live and studio recordings, radio and television broadcasts, and a documentary.


Glenn Gould’s recordings of Alban Berg’s Piano Sonata

As I will show, these recordings bear witness to significant shifts in the pianist’s aesthetic.  This post focuses on Gould’s uncommon approach to manual asynchrony and tempo in these recordings, with the hope of encouraging further discussion about performance practice.

Most pianists are familiar with the stigma surrounding what is colloquially known as ‘breaking’:  the vertical separation of harmonic and melodic material in performance.  I can recall one instance in graduate school where I was cautioned for using it.  My teacher reminded me that it was not 1940 and told me a story about an old pedagogue who warned against its occurrence more than once during a lifetime.  This position may seem extreme (and it is), but it reflects current fashion in performance.  In the introduction to Neal Peres da Costa’s monograph Off the Record, musicologist Clive Brown argues that the device’s scarcity in contemporary practice is symptomatic of a modern misunderstanding:  “an extreme example of the ways in which the implicit meaning of the notation fell victim to an unhistorical conviction that fidelity to the composer’s intentions required the most scrupulous literal observance of the notated text.”

Many instances of manual asynchrony in Gould’s recordings flout common practice.  A particularly noteworthy example occurs in his 1974 Chemins de la musique performance during the development section.  In measure 101, Gould sounds the melody’s opening pitch twice:  first with the bass and then as the final note of the (notated) arpeggio.


Berg, Piano Sonata, mm. 100–102 annotated to show Gould’s manual asynchrony (the caesura and arpeggio are marked in Berg’s score)

Gould’s motivation likely stems from the dominant-function harmony in the preceding measure.  Despite the notated caesura, Gould is compelled to continue the melody upwards so that it sounds at the same time as the left hand’s tonic.  Gould is consistent with this decision later in measures 103, 105, and 106 where variations of the same material occur.

In measure 102 (see score above), Gould again chooses an unconventional asynchrony.  Of the three numbered pitches (C/A/D#), the middle note is sounded last, resulting in the performance order:  C/D#/A. The purpose of this choice is to highlight the inner voice which begins on A, repeats, and then continues upwards to F# and F♮.  Alfred Brendel’s more conventional performance allows us to put Gould’s creative approach into context.

Gould’s use of manual asynchrony allows him to articulate contrapuntal material with great flexibility.  His recordings demonstrate, in practice, how the work is underdetermined by its score.  Pianists could learn from Gould’s novel applications of manual asynchrony, which represent compelling solutions for highlighting contrapuntal lines.

Gould had a penchant for reconsidering tempos, as is well known from his 1955 and 1981 recordings of the Goldberg Variations.  But a remarkable increase in length also occurs across Gould’s recordings of Berg’s sonata, with the opening theme witnessing the greatest change.  On CBC radio in 1952 Gould plays the phrase in 12.36 seconds.  A notable increase occurs through his next four recordings, until 1958 in Stockholm where it is completed in 25.10 seconds.  Gould’s final recording, from 1974, is the longest at 28.57 seconds.  The total increase between the first and last recording is a dramatic 131.15 percent—more than double the initial length.  To situate Gould’s performances alongside common practice, here is Murray Perahia’s 1987 recording.  The aggressive nature of the first phrase’s growth displays Gould’s search for the limit of possibility.  Indeed, the expansion is non-linear and displays a logarithmic trend, meaning that it is inversely exponential:


The increasing length of Gould’s first phrase displayed as a logarithmic trend.

To spur discussion on these topics, I advance the following questions:

First, what should we think about the disappearance of manual asynchrony from common practice?  Have we lost the full expressive potency of revered works?  While the application of the device in Gould’s performance is undoubtedly brilliant, could we ever tolerate again a performance style that involves the dislocation of harmony and melody?

Second, Gould’s continuous reassessment of the work’s tempo demonstrates a restless interpretative search and a dissatisfaction with the habitual.  Should students be encouraged to reconsider their approach when revisiting a work, and can these interpretations reside safely inside of common practice?

Improvisation Experience Predicts How Musicians Categorize Musical Structures

by Andrew Goldman (Presidential Scholar in Society and Neuroscience at Columbia University)


This video explains a recent research project on musical improvisation published in Psychology of Music. It shows that more experienced improvisers, compared with less experienced improvisers but otherwise skilled musicians, categorize harmonies with similar functions as being more similar than those with different functions. We argue that this aids the ability to make appropriate substitutions, and shows a difference in the way improvisers structure their musical knowledge that facilitates their abilities to improvise.

For more information about this research project, please view this article in the journal Psychology of Music.

Beginning in Fall 2018, Andrew Goldman will be based at the University of Western Ontario, where he will work with Jonathan De Souza on the Music, Cognition, and the Brain Initiative.

Form-Functional Ambiguity: The Issue of Closure in Performance

By Ellen Bakulina (University of North Texas)

Last fall, I taught an undergraduate form analysis course at UNT.  I did not initially plan to focus on performance issues, but, as the semester went on, I got caught up in them more and more, partly thanks to the large number of performance majors in the class.

Among the various aspects of form, I find questions of closure the most obviously relevant to performance.  In the classroom, I particularly value two aspects of relating cadence analysis to performance choices:  (1) it is an effective way to show connections between a theory course and the students’ daily performance practice; and (2) it develops their musical thinking.  Indeed, cadence identification carries crucial performance implications, inviting such questions as:  Where is this passage going, and why?  How can one decide what is and what isn’t a phrase ending—and, therefore, how long is the phrase?  How do these endings (punctuation, to use H.C. Koch’s term) partition the larger whole?  And perhaps more difficult: should one project that a certain point is a goal, and if so, how?  Directly showing the result of one’s analysis in performance may be a naïve idea; as William Rothstein suggests, sometimes underplaying an important structural event has greater value than “forcing” it upon the listener.  (His remark, in fact, refers to melodic material, but it can be applied to many other musical dimensions.)

None of the thoughts here are original or new in the field.  The concept of cadence has been, and continues to be, thoroughly explored by theorists, one of the most recent contributions being the book What Is a Cadence, edited by Neuwirth and Bergé.  My goal is to use existing concepts to offer performance-related suggestions about a specific passage that could be demonstrated in class.  (Of course, the opposite process is often useful as well:  existing performances influence one’s analysis, whether consciously or not.)  I also provide my own performance of the passage in three different versions.  The piece is the first movement of Beethoven’s Piano Sonata in E, Op. 14, No. 1.

Beethoven s 9 mt

In my class, which used the theory of William Caplin, I gave this movement for the final analysis project; the wealth of cadence-related interpretations the students came up with enriched my own understanding of the piece.  Here, I focus on the main theme, which offers a truly ambiguous cadential situation with interesting performance implications.  The issue involves both the location of the cadence and its type.  I analyze with the presumption that m. 13 begins the transition, with the main theme ending at this point or shortly before.  The theme is written in a (somewhat loose-knit) sentence form.

Here, I should give some conceptual background.  According to Caplin’s theory, a sentence consists of three phrase functions:  presentation, continuation, and cadential phrases.  The latter two functions are the most likely to be destabilized, or loosened, which is exactly what happens in our theme.  “Problems” here begin in mm. 7–8, whose six-four chord suggests a cadential function.  By m. 9, it becomes clear that the cadence has been delayed, causing an expansion of a supposed underlying 8-measure structure.  The expansion lasts until m. 13 and consists of a cadential progression (the bass line G#–A–A#–B–E) stated twice, finally reaching a true cadence in m. 13 that elides with the transition.  My performance (Recording 1) attempts to represent this analysis.  What kind of cadence is it?  An IAC, one would probably presume at first glance, since ^1 is avoided in the upper part.  But a closer look reveals that this initial analysis misses many musical nuances.

One of the students in my class identified a cadence—a PAC!—in m. 11.  Although it makes harmonic sense (there is a root-position tonic), I don’t like this choice, mainly because the next two measures repeat mm. 9–10 an octave lower and thus suggest that the theme is not yet completed.  (In Caplin’s terms, mm. 11–12 represent an extension of the cadential phrase, and repetition is a defining element of phrase extension technique.)  However, this student’s answer drew my attention to the possibility of hearing the inner voice (G#–G♮–F#–E) as the true soprano line, thus rendering the B on top a non-structural cover tone both times—leading up to m. 11 and m. 13!  This connection, of course, opens up the possibility of hearing a PAC in m. 13 with an implied E3 during the ostensible rest on the downbeat of m. 13.  The descent to E would be the structural descent in a Schenkerian analysis of the theme.

How can one project such an analysis in performance, given that the melodic arrival on E in 13 is so demonstrably withheld?  The key, I think, is to emphasize the inner-voice line G#–G♮–F#–E the first time, making sure the arrival on E at m. 11 is clearly heard, but at the same time avoiding a cadential effect—avoiding making m. 11 too conclusive; and then to emphasize the same line the second time, affording the listener the greatest chance to hear the implied arrival on E at m. 13 in the middle voice.  This is not easy to do, and I’m not sure my performance (Recording 2) does full justice to my analysis.  In this alternative version, I strive to de-emphasize the top-voice B, which I brought out in the first example, illustrating an IAC.

Can one consider another possibility—a half cadence in m. 12 instead of an elided authentic one in 13?  This would mean a linear interruption of the inner voice on the F#, instead of its resolution to E.  This is the issue that Poundie Burstein engages in his recent article on half cadences.  He examines progressions where V is followed by I at phrase structural boundaries; deciding between a half and authentic cadence usually involves determining which harmony is the goal:  the V or the I?

Indeed, several of my students chose to read a half cadence in m. 12, possibly to avoid the phrase elision—of which students are sometimes inexplicably afraid—that an authentic cadence would require.  The dominant of this half cadence is a V7, what Janet Schmalfeldt calls a “nineteenth-century half cadence.”  (Burstein has recently argued that this cadence type can sometimes occur in the eighteenth century as well.)  This reading, supported by the absence of a tonic downbeat in m. 13 (though it is obviously implied), changes the phrase structure and phrase rhythm.  All phrases in the main theme are now two and four measures long, the overlap is gone, and the cadence in m. 12 is hypermetrically weak (an odd-strong hypermetrical pattern has been established from the opening).  This is an important consideration; a hypermetrically strong cadence is likely to sound more conclusive.  Although this is not my personally preferred reading, I have attempted a performance of it (Recording 3).  This version may exaggerate the break between the two sections a little; if it does, the reader will hopefully find their own way to project a half-cadential reading in performance.

I should add that Caplin himself gives the theme up to m. 13, thus implying a cadence there, rather than in 12.  Ultimately, I agree with this interpretation, whether one chooses the IAC or PAC option.  I should also add that both Caplin and Burstein are very much aware of the performance implications of the concepts I have discussed; references to performance choice appear in their work multiple times.  What I have tried to do here is to offer, in a specific situation, some specific performance suggestions for solving the problems one encounters in analysis, and to show its value in a teaching situation.