Questioning Schenkerism

by Bengt Edlund (Author)
©2015 Monographs 497 Pages


During the past fifty years Schenkerian theory has been adopted as the main method for analysing tonal music. This book questions the value of Schenker’s «tonal analysis» for musical description and interpretation, and discusses its relations to «generative» theory and «implicational» analysis – taking into account its links with linguistic syntax and the perception of tonal closure. It is observed how auxiliary theoretical concepts transform the music so as to pave the way for preordained tonal structures. Alternative readings of the music examples are provided.

Table Of Contents

  • Cover
  • Title
  • Copyright
  • About the author
  • About the book
  • This eBook can be cited
  • Table of Contents
  • Preface
  • Chapter 1: Schenkerian theory and better comparison: An out-of-the-way perspective
  • Introduction
  • Schubert’s Das Wandern and the question of consecutive fifths
  • What went wrong with this “good comparison”?
  • A further Schenkerian reading of Das Wandern
  • A free reduction of a free composition
  • The initial theme of Beethoven’s Op. 90: Schenker’s background
  • Salzer’s reduction
  • Beethoven’s theme: some preliminary observations
  • An alternative reduction to be discarded
  • Final attempts at a reduction
  • Beethoven’s theme: some conclusions
  • Schenker and Schumann’s Aus meinen Thränen sprießen
  • A critical scrutiny of Schenker’s reduction
  • Schenker’s analysis of the middle section: the foreground
  • Schenker’s analysis of the middle section: the middleground
  • Harmonic observations
  • Conclusions and a bottom/up attempt at reduction
  • Bringing in the context of the song
  • General conclusions
  • Chapter 2: Disciplining reduction and tonalizing interpretation
  • Introduction
  • Reduction and interpretation
  • Disciplining reduction
  • Schenker’s analysis: the first phrase of the antecedent
  • The second phrase of the antecedent
  • The first part of the middle section
  • The second part of the middle section
  • The entire theme
  • von Cube’s reduction
  • Correcting the disciple
  • von Cube’s defence
  • Enforcing discipline
  • Maintaining discipline: Beach teaching teachers
  • The A-sections according to Beach
  • The B-section according to Beach
  • Tonalizing interpretation
  • Schenker’s analysis
  • von Cube’s dual descent reading
  • Beach’s analysis
  • Searching for the theme’s “tonal content”
  • Rising fourths and falling seconds; a network of implications
  • A model and its expanded, inverted-counterpoint replica
  • Tonal structure in terms of “drones”
  • Tonal structure in terms of “focal” events
  • “Focal” reduction and interpretation
  • Conclusions
  • Chapter 3: Is tonal music hierarchic? An impenitent sermon
  • Introduction
  • Two ways of presenting tonal reductions
  • Schenker’s reduction: the first half of the chorale
  • The second half of the chorale
  • Cook’s discussion of Schenker’s analysis
  • Cook’s alternative reduction
  • Lerdahl and Jackendoff’s “Generative Theory”
  • The first phrase: metric accents and rhythmic groups
  • The first phrase: time-span reduction
  • The first phrase: prolongational reduction
  • Issues of “final-state” understanding
  • Cadences and branching
  • Issues of grouping and form
  • Assigning hierarchical structure to music
  • Summary of the reductive accounts
  • Towards a non-hierarchical analysis
  • Finding the “focus” of the chorale
  • Chapter 4: Prolongation vs. implication
  • Introduction
  • Implications and prolongations in the Les Adieux introduction
  • Comparing prolongations and implications
  • A selection of observations
  • Chapter 5: A hitch-hiker’s guide to the repeat
  • Introduction
  • Main objection
  • The first section
  • The third section
  • The middle section
  • Schachter’s middleground and Schubert’s sketch
  • Motivic relationships and the element of dialogue
  • Schachter’s “phrase rhythm”: some critical observations
  • Metre and rhythm in the middle section
  • The Trio
  • A question of maps
  • An alternative structural account
  • The outer sections
  • The middle section
  • Conclusions
  • Some issues of interpretation
  • Chapter 6: Schubert, Schumann, and Schenkerism. Tonal vs. focal Reduction
  • Introduction
  • Salzer’s reading of a Schubert waltz
  • A bottom/up reduction of the waltz
  • Schumann’s Albumblatt: a preliminary musical description
  • Some attempts to make reductive sense of the Albumblatt
  • “Outer” form and “tonal” form; the problem of repeats
  • Revisiting Schubert’s waltz
  • Conclusion
  • Chapter 7: Syntactic vs. rhetoric structure. Language, music, and tonal reduction
  • Syntax and closure in language and music
  • Clauses and sentences in language and music
  • Closure, unity, and coherence
  • Tonal structure vs. tonal content
  • God Save the King: two syntactically independent units
  • A Schenkerian reduction
  • Alternative readings
  • Conclusions
  • Gott erhalte Franz den Kaiser: three relatively independent units
  • Schenker’s reduction in Der freie Satz
  • Schenker’s reduction in Der Tonwille
  • Alternative readings
  • Conclusions
  • General discussion
  • Chapter 8: Tonics and returns. A modest investigation
  • Introduction
  • Salzer’s readings of a Schubert waltz and a Mozart Adagio
  • Modulating variants of the waltz and the Adagio
  • Lerdahl and Jackendoff’s prolongational trees
  • Experimental design
  • Results
  • Conclusions and discussion
  • Chapter 9: Shaving Schenker
  • Introduction
  • Schenker’s and Bursteins’ readings: a comparison
  • Criticism of Schenker’s analysis
  • Starting from scratch
  • Three deep-layer structures
  • Summary and discussion
  • Music Examples
  • Chapter 1
  • Chapter 2
  • Chapter 3
  • Chapter 4
  • Chapter 5
  • Chapter 6
  • Chapter 7
  • Chapter 8
  • Chapter 9
  • References

← 10 | 11 → Preface

Ever since I got to know Schenkerian analysis during my student years, I have been sceptical of it, and the more I learnt about it, the more negative I turned. More often than not, the music under analysis fared so badly. And yet, to my amazement, this method was embraced so enthusiastically by so many analysts, had so many adherents.

The fact that, when listening to music, we pay less attention to some events in favour of others that emerge as more important, is most productive when it comes to analysis. It is therefore a pity that this idea has virtually always been used for one and the same thing in Schenkerian analysis: to force Ursätze onto tonal music in order to demonstrate that the music exhibits tonal unity. This busyness is superfluous, however, since Schenker’ theory has established beforehand that, given the analytical devices warranted by its success story, any non-deficient piece of tonal music is bound to exhibit an Ursatz.

But there are so many other and more worthwhile things to say about music, so much else to discover if you cease to treat it as a quasi-visual, through-and-through hierarchical thing, if you try to describe it as a process. Maybe it is time to proclaim a fifty-year moratorium of Schenkerian analysis, or at least to ask for a less orthodox approach to reduction. Meanwhile, it is necessary to disturb a tradition of panegyrics and routine analysis with some criticism of Schenkerian theory as it emerges in its practice, to ask questions rather than provide answers.

The nine chapters in this volume – they are connected in various ways but must not be read in succession – are written during a period of some thirty years. Chapter 1 discusses the value of Schenkerian analysis while chapter 2 is mainly devoted to a study of how it has been taught to its disciples. In chapters 3 and 4 the methods advanced by Lerdahl and Jackendoff, and Leonard B. Meyer are brought in for comparison. Problems of “tonal” form make up the theme in chapters 5 and 6. Using language as a point of departure, chapter 7 deals with the relationship between syntactic structure and rhetoric content. Chapter 8 accounts for an experiment on ← 11 | 12 → tonal closure. The subject of chapter 9, finally, is a number of auxiliary concepts frequently resorted to in Schenkerian analysis.

The choice of the works to be discussed entirely depends on the texts or occasions that once sparked off my spirit of contradiction. To make up for all criticism – and to set things right in the maltreated music – alternative readings are proposed, readings that bring in other ideas and adopt a non-Schenkerian approach to reduction.

For five further studies mainly or partly devoted to the shortcomings of Schenkerian analysis, the reader is directed to Chopin. The Preludes and Beyond (Frankfurt 2013, Peter Lang Verlag).

This book and the previous volume on Chopin’s Preludes have been generously supported by Sten K Johnssons stiftelse.

Lund, 12 September 2014

Bengt Edlund


← 12 | 13 → Chapter 1 Schenkerian theory and better comparison: An out-of-the-way perspective

“as long as we remain silent about the questions, we may keep the illusion that we might one day be able to find the answers”. From The Year when Ricardo Reis died by José Saramago


Taking part in a discussion devoted to the problem of whether or not music analysis can or should raise claims to be a scientific activity, Nicholas Cook has argued that analyses are not be measured by scientific standards.1

One of his reasons for denying scientific status to music analysis is that the scientific stance has turned out to be untenable in the light of negative empirical evidence, brought up not least by Cook himself.2 People tend to notice very few of the things that analysts are keen on observing, and speaking specifically about Schenkerian analysis, he concludes: “If the principle of tonal closure has little or no perceptual significance at the larger time scales found in most tonal compositions, is there not something radically wrong with a theory that ascribes fundamental aesthetic importance to it?” But this is, Cook maintains, too easy a way to get out of the dilemma: “To ask this is to assume that a theory of musical structure has to be also a theory of perception. But there is no intrinsic need for the theorist to conceive of musical structures in the same manner that the listener perceives them.”3

← 13 | 14 → “What we need is a rationale for adopting Schenker’s analytical methods while rejecting his epistemology. And a clue to how such a rationale might be formulated can be found in the concept of Darstellung as developed by other writers in Vienna during the early decades of the present century”. (p. 124)4 According to Cook, then, analyses – and Schenkerian tonal reductions in particular – are not meant to account for how and what people actually hear; tonal analyses make up representations guiding the musical experience by drawing attention to certain traits of the music. Subjecting analyses to empirical tests is therefore simply beside the point.5

Dismissing empirical verification, Cook proposes another way to establish the merits of analytical descriptions. They should be evaluated with respect to whether they seem enlightening or useful; the assessment becomes a matter of the individual reader’s appreciation. “A Schenkerian explanation is validated when its readers accept it as a satisfying account of the music in question. […] But a Schenkerian analysis does not simply present an interpretation; it provides reasons for the interpretation, implicitly if not explicitly.” (p. 128)

The value of an analysis, in turn, depends on the extent to which the reading does not just reproduce the musical surface, but goes beyond it, transforming it and making possible a “good comparison” with the actual ← 14 | 15 → music.6 “The function of an analysis, then, is not to reduplicate the composition in question; it is to focus the readers’ attention on its individual qualities. And this means that it is wrong to judge an analysis according to how directly it mirrors the surface of the music, with its tunes and silences and abrupt changes of texture. What matters is the extent to which it illuminates the surface.” (p. 132)

All this may seem acceptable in as far as there is no doubt more to be said about a piece music than what can be heard immediately by just anybody. But it seems that when defending Schenkerian analysis in this way, Cook lets it come off the hook too easily.

Firstly, Schenkerian readings are in fact quite often vindicated by recourse to how the music is heard or, in an explicitly normative vein, by appealing to how it should be heard.

Secondly, since (as Cook himself has shown) listeners tend to lack ears for tonal closure even in fairly short and simple pieces, one might suspect that thousands upon thousands of middleground and background graphs found in Schenkerian analyses do not make up heard structures but are products of close and quite deliberate visual inspection of the music as printed in the score. If listeners cannot even deal with short pieces in the way required or predicted by the theory, they are most likely to completely lose track of the long-range harmonic and voice-leading connections supposed to lend tonal unity to extended works, and making up the uniquely valuable essence of tonal reductions. Using the score as the main (or only) source of analytic discovery, and presumably corroboration as well, means that it is possible to rely quite heavily on a top/down, end-towards-beginning, perspective of the musical process, a vantage position agreeing all too well with the unfortunate normative character of Schenkerian theory.

Thirdly, since Schenkerian theory provides analysts with various criteria of reduction, telling them which events and parameters that are to be taken into primary account, and since it also posits the structures that must ultimately emerge, it may be argued that the results all too closely, ← 15 | 16 → and in a most unscientific way, “mirror” not “the surface” of the music, but the basic assumptions of the analytic undertaking.

Finally, the idea of using “good comparison” as a yardstick when evaluating tonal reductions seems unreliable. Since, if Cook’s view is adopted, a worthwhile reduction involves transforming what is given, one cannot escape the crucial question of evaluation vs. validation, which is more of a one-way relationship than Cook apparently wants to think. Analytic results are not necessarily valid because we find their outcomes “illuminative”.7 Validation should precede evaluation, not the other way around. If the various analytic decisions, upon which a tonal reduction is based, emerge as arbitrary, far-fetched, dogmatic, or incorrect – i.e. as invalid – this is bound to affect the assessment of the value and usefulness of the result. And yet it turns out that such analyses, being more or less irrelevant because they miss their objects, are accepted as “good comparisons” by indiscriminate readers as well as by (properly attuned) analysts.8 If the “illuminations” are delusions, one cannot just go ahead as if nothing had happened.

Besides, since Schenkerian Darstellungen are predicated on the idea that all tonal pieces of music are (or should be) modelled on authentic cadences prolonged by means of strict counterpoint, successful tonal reductions invariably produce Ursätze, i.e. authentic cadences prolonged by strict counterpoint, as the ultimate representations to be compared with the actual music. But are comparisons based on such standardized schemes always very “illuminating”, haven’t they lost the charm of novelty and the excitement of boldness after almost a century of persistent reductive efforts. And, considering the element of dogmatism involved in these undertakings, do they really capture the “individual qualities” of the works? Isn’t there a possibility of alternative accounts, of “comparisons” issuing from other theoretic agendas, comparisons that may be just as good – or even better? Indeed, if analyses are “representations”, that are not devised ← 16 | 17 → to (primarily) record one’s own listening but to “guide the musical experience” of others, it would be a pity if a very influential analytic theory has fobbed off on us readings that are pedestrian as to approach and results, readings that suppress other “representations” that might be more rewarding?

According to Cook, “a Schenkerian explanation is validated when its readers accept it as a satisfying account of the music in question”. But this is too simple to be acceptable: turning to the consumers of Schenkerian analyses, a quite select group of Kenner und Liebhaber, from where do they draw their standards of satisfaction? If Schenkerian reductions are to be spared the risk of empirical testing in current, scientific sense, they must at least be subjected to thorough and unbiased analytic validation.

In what follows, the three specimens of tonal reduction chosen by Cook to substantiate his “good-comparison” idea will be critically examined, laying bare the roots of his analyses and proposing alternative accounts. To the extent that these Schenkerian “representations” lack validity, as assessed by the musical ear and by analytic common sense – a kind of testing that does not equal scientific corroboration, but at least amounts to a sceptical attitude as becomes any scholarly effort – they cannot very well provide “good comparisons”.

Schubert’s Das Wandern and the question of consecutive fifths

Asking rhetorically “how would you analyze the second half of Schubert’s song ‘Das Wandern?’”, Cook offers the reduction shown in Exs. 1 a–c. As he points out in his comments, this reading (“your” reading) exhibits “glaring consecutive fifths between the outer parts” at level 1c. (pp. 126–128) At the previous level 1b they are less glaring due to the fact that the notes making up these objectionable intervals turn out to be non-simultaneous.

Cook also presents two further background layers, Exs. 1d and 1e, that do not feature consecutive fifths, but calls them in question since they do not match what we are likely to hear. Thus, a listener might complain that in 1d the V harmony is “given precedence over” the previous VI chord, forming ← 17 | 18 → the first member of the sequence, whereas in 1e it is hard to hear “a structural dominant” during the concluding “rocking alternation of F and Bimages”.

While one is bent to agree with this listener, it must be recalled that the argument is inconsistent considering Cook’s dismissal of listening when it comes to validating tonal analyses: “But there is no intrinsic need for the theorist to conceive of musical structures in the same manner that the listener perceives them.”

Moreover, it must be pointed out that these alternative backgrounds are just as problematical as the one shown in 1c, as long as layer 1b remains intact. If there are any undesirable middleground consecutive fifths buried in the music, and this is what 1b suggests (no matter the lack of exact co-ordination), they are not wiped out of existence just because they are not allowed to contaminate the background 1c. The first of the forbidden fifths is simply omitted when moving from 1b to 1d or 1e.

It should also be observed that while Cook appeals to listening when 1d and 1e are rejected, he does not subject 1c to such a test. He just says that it “looks uncouth”. But the fact of the matter is that neither the background 1c, nor its middleground 1b are adequate if you venture to check them against Schubert’s music – Das Wandern does not sound, nor does it look, “uncouth”.

Finally we must ask ourselves whether there are not alternative, and better, reductive accounts than 1b and 1c.

Returning to Cook, he maintains that “(c) is not in itself a less accurate formulation of the tonal structure” of the song, “than (d) or (e), but that, due to the consecutive fifths, it is less satisfactory as an expression of that structure in terms of the metaphor of Fuxian counterpoint. It makes the music look ungrammatical and, therefore, incoherent.” And he continues: “But this is not because the middleground consecutives contravene any natural law of musical organization. It is because they run counter to the representational means adopted in Schenkerian analysis. They spoil the comparison between Schubert’s song and Fuxian counterpoint.”

First of all, it must be pointed out that within Schenkerian theory Fuxian counterpoint amounts to more than a vehicle for the Darstellung of musical structures. Strict counterpoint is also adopted as a main heuristic device in ← 18 | 19 → the reductive process and as a decisive Satzprobe norm when validating the results – that is why Cook can say that 1c looks “uncouth”.9

Turning specifically to undesirable consecutive intervals, this is how Cook summarizes Schenker’s views: “But he accepted the traditional prohibition of real consecutives. Now Schenker did not see his theory as a metaphor; he believed that there are natural laws, which operate equally at the level of large-scale and that of note-to-note structure. So one might expect to find an equally strict prohibition of consecutive fifths when these are generated at a single structural level in the middleground. In the event Schenker is more pragmatic, saying that forbidden intervallic successions may be found in the middleground, but that ‘it is then the task of the foreground to eliminate them’. In practice, however, both Schenker and present-day Schenkerians tend to avoid middleground consecutives.”

As regards the second part of Das Wandern, graph 1c is not a middleground, but rather a background, and this fact cannot but turn the “glaring” consecutive fifths even more objectionable, indeed unacceptable, from a Schenkerian point of view. And as anybody familiar with the practice of tonal reduction can testify, the analysts take it to be their task to let layers beyond the foreground eliminate prohibited consecutives whenever they impend. It is just a matter of selecting suitable notes, of avoiding notes that would produce undesirable intervals if they were allowed to show up at deeper levels, and this is (among other things) what using strict counterpoint as a heuristic device in tonal analysis amounts to.

The avoiding job is facilitated – and the analysts’ scholarly doubts, if any, are considerably alleviated – if one follows Schenker’s lead and pursues tonal reduction as an exercise in top/down prolongation, i.e. if one does not begin by selecting middleground notes from the foreground but starts the analysis by producing them out of the background. However dubious it may seem to outsiders, this path of analytic discovery is considered quite acceptable to those knowing that they possess the truth when it comes to structure in tonal music.

← 19 | 20 → But it is possible to conceive of other, true kinds of reduction, procedures where it is not “the task of the foreground to eliminate forbidden intervallic successions”, but the task of the analyst, conscientiously pursuing reduction as a bottom/up process, to start from the foreground and derive the following layers without preconceived notions as to what should or should not be there.10 In a genuine reduction, the foreground/the musical surface is the inviolable point of departure. If the foreground/surface, or any conscientiously derived deeper layer, features prohibited consecutives, they must be accepted, and if emerging as structural at that layer, they have to be retained at the next layer of the reduction, and possibly beyond.

Cook is not altogether satisfied with his reduction, however.11 Due to the wretched consecutive fifths, it “spoils the comparison between Schubert’s song and Fuxian counterpoint”.

But before trying to find out what went wrong, one thing should be made clear. Far from spoiling the “comparison” with Fuxian counterpoint, the background 1c – in virtue of being patently un-Fuxian – makes up a quite interesting comparison between Schubert’s song and strict counterpoint. This “uncouth” background is in a way a “good comparison”, suggesting no less than three conclusions.

Thanks to Schenker’s supreme analytic method, Schubert’s little song, agreeable as it is, has been exposed as a flawed Machwerk. What else can it be since its background gravely fails to meet the standards of strict counterpoint? But it is also possible that the idea of using Fuxian counterpoint as a “metaphor” when producing “illuminating” Darstellungen is not as universally applicable as the adherents of Schenkerian tonal analysis are prone to think. When dealing with this simple but arguably very fine song in the theoretically sanctioned way, an illegitimate background emerged. ← 20 | 21 → Finally, the analysis offered by Cook in Exs. 1 a/c might after all be an incorrect and misleading one. If Das Wandern itself seems tonally quite healthy and coherent while its background 1c looks “uncouth”, the Schenkerian manner of dealing with the song has in fact produced an invalid representation of its tonal structure.


ISBN (Hardcover)
Publication date
2015 (June)
Schenkerian theory analysis of tonal music tonal closure musical interpretation
Frankfurt am Main, Berlin, Bern, Bruxelles, New York, Oxford, Wien, 2015. 497 pp., 122 b/w fig.

Biographical notes

Bengt Edlund (Author)

Bengt Edlund, trained as a pianist, has been active as a music critic, and as lecturer at the Department of Musicology, University of Lund, where he was appointed Professor in 2000. His main fields of interest are music theory and analysis, music cognition and aesthetics, and musical interpretation.


Title: Questioning Schenkerism