One of the most grown-up review sites around

54,416 reviews
and more.. and still writing ...

Search MusicWeb Here



International mailing

Founder: Len Mullenger                                     Editor in Chief: John Quinn              

PURCHASE Through MusicWeb International
£60 post-free

Purchase via PayPal secure online shopping. All major credit cards accepted. For further information on purchasing via the internet, please visit the PayPal website.


Bach Performance Practice, 1945 - 1975: A Comprehensive Review of Sound Recordings and Literature
by Dorottya Fabian

University of New South Wales, Australia.
October 2003.
42 illustrations and 17 music examples.
Hardback, 328 pages, 234 x 156 mm.
Including a CD consisting of brief excerpts from various performances of the Goldberg Variations, Brandenburg Concertos, and Passions from published recordings discussed in the text.
ISBN 0 7546 0549 3.

In this book the author explores that most fascinating revolution in the world of music. The rediscovery and restoration of authentic Baroque performance circumstances occurred mostly during the third quarter of the twentieth century, particularly as evidenced by sound recordings of music by Bach. Naturally there is much more to this story, and the book has considerable value as a bibliography of essays and books on historical performance practice and historical instruments in general. She actually begins her narration in the earliest part of the century well before 1945, and includes some comments on what happened afterwards. A remarkable facet of her exploration is that from the very earliest times all responsible persons involved have publicly admitted the utter impossibility of ever reproducing historical conditions with any exactitude, the best that could be achieved being a gesture, an approximation, a suggestion, an exploration; but an immensely worthwhile one.

The author has compiled a comprehensive list of recordings of Bach’s Brandenburg Concertos. Missing from her list is the 1949 Fritz Reiner recording with soloists from the New York Philharmonic Orchestra on CBS/Columbia label LP which was one of the first to include a high trumpet brilliantly played by William Vacchiano, and Sylvia Marlowe as harpsichord soloist; it was one of the best selling in the United states in the 1950s and the hit of my college days and afterwards as it was reissued successively on various bargain labels and is available today on historical CD issue.

Also she fails to include either of two versions by Hermann Scherchen, both the 1954 recording with the Cento Soli Orchestra of Paris, and his 1960 recording with members of the Vienna State Opera Orchestra, "Willy Boskovsky concertmaster," both of which are in print. The Cento Soli recording has perhaps the slowest final movement for concerto #4 ever recorded (6’05"), the overall length of the concerto coming in at 18’40"; she lists the Karajan/BPO recording, timed at 18’10", as the slowest she is aware of.

Fabian refers to the second disk only of the 1960 recording in her end-of-book summary as "MCA 80121" giving the recording date as 1959, listing Willi Boskovsky as soloist, and stating "no other soloists listed." Clearly she has a copy with defective packaging since my two-disk complete set [MCAD2 9831] lists all the soloists and gives the recording date as March 1960, as does the Scherchen discography by René Trémine (pub. TAHRA Productions). Clearly, Brandenburg #6 is played by soloists only, while #3 sounds like it has extra strings to double parts as well as a solo chest. The #4 in this recording has the longest second movement, 6’13", and is slower over all, at 20’30," than any she has listed. Since in her table she performs statistical analysis on the timings of the movements, inclusion of these data points could change all her numbers.

Omission of a recording or two could hardly be considered a serious flaw, except that Scherchen is not an eccentric who can be ignored, but was one of the central figures in the historical performance movement, pioneering not only in Bach recordings but in using small orchestras and choruses for Handel, Beethoven, Mozart, and Haydn as well. Although Fabian has many and extensive comments on Sherchen’s vocal recordings throughout the book, unfamiliarity with his total oeuvre would be as serious a flaw in any survey of historical performance practice as would be, say, unfamiliarity with Harnoncourt.

The Reiner Brandenburg recording was popular precisely because it was one of the first to use harpsichord and high trumpet and went far to create an audience for future recordings which ventured further into the area of authenticity. Interestingly it was produced by the same musicians and at about the same time as the famous Louis Kaufman Vivaldi Four Seasons recording, in its time a daring example of original performance practice, and shows that New York musicians were important in the historical performance movement.

It is of course a trap to consider the overall timing of a movement to be an expression of exactly how fast the movement is played, because added or omitted repeats or prolonged or truncated fermate, for instance, can skew the correlation between timing and pulse. She rightfully avoids this in many instances by referring instead to the measured metronome pulse.

In listing the Haas recording with the English Chamber Orchestra, she does not mention that the solo harpsichord in #5 was played by Robert Veyron-Lacroix, although she includes similar information for other recordings in her list. Again, I don’t mean to suggest that a few omissions invalidate her work, however it is disconcerting that, in an expensive work that claims to be comprehensive and authoritative, a person so casually interested as I can at once detect lacunae.

Fabian refers to a harpsichord with the capability to change registers by means of foot pedals as a "pedal harpsichord" whereas one would assume she would mean by that a harpsichord with 16’ rank and separate pedal keyboard. She deduces that Zusana Ruzickova played a Neupert harpsichord, whereas I am sure I recall a listing to that effect on one of Ruzickova’s recordings, making conjecture unnecessary.

In her list of performances of the Goldberg Variations stopping at 1975 means she does not include the Fernando Valenti recording. Valenti was the most popular harpsichordist of the 1950s, his recording being all the more remarkable since he was using a different instrument, a smaller instrument than the one he used to record his legendary series of Scarlatti harpsichord sonatas. The recent burgeoning number of recordings of the "Goldbergs" on pianoforte by the likes of Perahia and Schiff, attest to the works’ durability in the medium of piano literature, and the author later admits this point. A similar point might be made for the six keyboard Partitas.

I used to be a harpsichord snob myself, so I am sympathetic to that attitude. The assumption that only the harpsichord is appropriate for Bach’s keyboard music can no longer be accepted uncritically, and to her credit Fabian is flexible about this point. Certainly by 1748 it was clear to any musician with a grain of sense, and Bach is generally accredited with at least that, that the pianoforte was the instrument of the future. A nearly airtight case can be made that the second book of the Well Tempered Clavier as well as Art of Fugue and the Musical Offering can be taken to be works conceived for pianoforte. The Goldberg Variations from 1742 are clearly a watershed consideration, but one could argue that the continuing success of pianoforte performances of the work in the past as well as the present time constitutes a Heuristic demonstration of what Bach had in mind. For this reason, use of a harpsichord for late solo keyboard works cannot be taken ipso facto as evidence, or the use of pianoforte as lack of evidence, for respect for original performance practices.

On page 72 we find the following paragraph: "It is important to add that the early examples of a homogeneous, mostly simple registration (especially the performances of Leonhardt 1953, Walcha and Kirkpatrick) do not otherwise represent an historical style of playing as conventionally understood today (2002). Rather, they (as well as Newman much later) exhibit a literalistic approach, which could reflect the belief that playing the correct notes on the ‘right’ instrument is all the music needs to ‘speak for itself’—and in this case any harpsichord might be acceptable, as long as it is not a piano. In an attempt to second-guess the motivation of those using varied registrations (e.g., Richter, Malcolm, Pelleg, Payne) I would venture to suggest that their approach—which I labelled 'post-romantic' earlier—might represent the solution to a conflict of personalities within the performer; the ‘interpreter’ pitted against the ‘scholar’. The latter is convinced that the piano is not an appropriate instrument, but the former is frustrated by the ‘limitations’ of the harpsichord (see the earlier citation from Hubbard). To me it seems plausible that while searching for expressive means and a compensation for the blunt tone of their ‘surrogate instruments,’ artists like Richter, Marlowe, Malcolm, Ruzickova, Galling, and others turned to a Regeresque sound ideal, rich in sound colours and dramatic changes of registrations. Not having enough information at their disposal about historical harpsichord technique and baroque means of expression, registration might have remained their only interpretative vehicle."

This paragraph shows the author to have a badly flawed understanding of artistic impulse. In the first sentence she lumps Leonhardt, Walcha, and Kirkpatrick together in one parentheses and sends my jaw to the floor. Three more diverse artists could hardly be imagined — the schoolteacher, the lofty mystic, and the Italian street player. Does Fabian mean to tell me she cannot tell the difference between them? Or that the difference, at least for the sake of this argument, means nothing? Assuming the latter, she then goes on to psychoanalyse artists who pursue organ technique — changes of register — on the harpsichord; unwilling to grant them artistic taste, she must categorise them not merely as ignorant, but desperate to find compensation for inadequate personalities. The thought that they may have, with intelligence, knowledge, and genuine artistic sensitivity and impulse, come to different conclusions than the majority of academics of a later time seems not to occur to the author.

I must at this point summarise the arguments against her and her selected sources. If a majority of instruments in museums are of a certain type, I suggest that instruments which end up in museums were not played. Instruments which were played a great deal eventually collapsed to kindling and were burnt for firewood. Large harpsichords, one such as was owned by Bach, did not survive because they were rebuilt as pianofortes, and we would expect very few historical examples. It may be true that there were few instruments in Baroque times having three or more registers, which had knee levers and hand pulls instead of the much more expensive foot pedals to change registers and engage couplers. In this sense, in Victorian times most keyboard instruments were parlour spinets and uprights. Must we then be prohibited from playing Mendelssohn’s Songs Without Words on a concert grand piano because such instruments were relatively uncommon in 1847? It must be pointed out that the reason pianos were invented is because harpsichordists were unhappy with the limitations of the harpsichord sonority. The people who bought pianos were harpsichordists who wanted more expressiveness and greater variety of tone colour. Some of these same people bought large harpsichords with multiple easily changed ranks and couplers. They were not neurotics or defectives, they were people who basically liked harpsichord sound and wanted to explore the lengths to which it could be taken. An artist in the modern age who does the same is engaging in historical performance practice, showing us something that was not merely possible, but was actually done, and might even by enjoyed by modern audiences. This is not ‘post-Romantic’ but rather ‘late Baroque’.

Most lovers of Baroque music can remember a period some decades past when nothing over a single 8’ rank was allowed, and the possibilities of expression were all but nil. Bereft of anything but a single weak tinkly sound, the players’ only remaining means of expression was all kinds of distortions of tempo and phrase, and endless series of maiden blushes, retards, slipped and anticipated beats, complicated patterns of staccato. Compounding the problem, artists sometimes recorded on museum instruments with which they were not familiar, leased at high cost only for the recording session with no extra time available for practice. Most of these recordings never sold well and have disappeared. What worries me is that if Ms. Fabian and her cohorts have their way, the magnificent recorded legacy of Marlowe, Malcolm, Ruzickova, and Valenti may also disappear. A tape master recording has a practical life of about fifty years. Acetate disks can turn to powder in thirty years, vinyl pressings stored under perfect conditions may last eighty years. Recordings made in the 1950s are today in serious jeopardy of literally falling to bits. In the face of scholarly hostility, I have made it a personal mission to restore these recordings to the digital medium while these disks can still be played, and I am making good progress. If I don’t do this, by the time people catch on to what they are missing, the recordings may be — literally — gone and we will have nothing but scholarly sneering to describe them.

One such sneering critic is quoted as describing "the constant dancing upon the pedals". In fairness, the number of recordings that could reasonably be described this way can be counted on the fingers of one hand, and some of them may have been considered experimental, e.g., Sylvia Marlowe’s Handel Harpsichord Suite #7, first movement; the other movements of this suite and the other pieces on the same disk were played much more conservatively. She was having a romp, not establishing a norm for all performances. Neither Fabian nor her sources describe the use of swell shutters on harpsichords, and I would like to know more about that, just who and when. I can hear them operating in recordings by Valenti, Malcolm, and Dart, but who else used them? What are the historical precedents?

On page 84 she takes another swipe at what she calls "pseudo harpsichords." Please understand, I am not shooting the messenger. Ms. Fabian is expressing not exclusively the opinion of others but her own opinion, and with considerable force. However, lest you think I intend to trash her entire book, let me praise her discussion in the following section of the use of various registers of voices in passion performances. It is informative, clearly reasoned, and utterly fascinating. When discussing the relative merits of boys versus female sopranos she does summarise the arguments in favour of the former. What she doesn’t comment on is the ability of some young women to imitate the sound of a boy soprano; perhaps this wasn’t encountered during the time slot of her study, 1945 to 1975. However, a few of these young women’s voices in the midst of a boys’ choir greatly stabilises the tone and improves pitch and tonal accuracy, so much so that I think this practice is now almost the norm. Emma Kirkby, for instance, has trained herself to sound either like a boy or even like a falsettist counter-tenor! Whether this is authenticity or not, the question is authentic what?

On page 85 she translates Bach as "[spring]" whereas my German dictionary gives "brook, stream, rivulet," all clear references to flowing courses, not sources, of water. Spring in German is Quelle.

There is no discussion of the observation of repeats as part of original performance practice. In 1945 the general approach was to ignore them. Today, omitting a repeat will get a musician thrown in the dock behind the butcher with his thumb on the scale for trying to sell his audience short weighted music. An important difference between the 1955 Glenn Gould recording of the Goldberg Variations, weighing in at 35 minutes, and the recent András Schiff performance, weighing in at over an hour, is who repeats what and why. Maybe in another book.

Her discussion of ornamentation begins promisingly with words from many authorities pointing out that ornamentation might be different at different times, and going on with a good discussion of various ways of discovering the correct ornamentation of a given phrase of Bach. But the case is ludicrously understated, and the fundamental point missed. Baroque composers did not write out ornamentation, not because there was a secret code which "everyone" knew (à la Tureck), but because even with the same instruments ornamentation would be different on different occasions. The tables of ornaments we have are all alike in one important respect: they are all intended for students who would not be expected to have experience or mature judgement and thus needed a bare-bones guide. A given prelude or fugue from WTKI might be played on the organ, clavichord, lute, or harpsichord and would be ornamented very differently on each. Furthermore, the ornamentation would be different depending on which of the popular unequal tuning systems was in use — a trill or mordent on a "wolf" note would make it sound less dissonant. Ornamentation could be used to improve audibility of musical lines by helping to match the instrument to the acoustics of the music room, which rooms as well as the instruments in them were widely different from each other. Even in WTKII, where the piano is likely the intended instrument, pianofortes of the time were so different from each other that appropriate ornamentation would be radically different from one instrument to another. And, since instruments were evolving rapidly, Bach must certainly have understood that pianofortes in 1758 would be different from pianofortes in 1748. A scheme of ornamentation would affect fingering, and vice versa, and hence the very lilt of the whole phrase, even on unornamented notes. In the Baroque period, music rooms were also used for many other purposes. For the past 150 years there has been much discussion about "good" and "bad" halls, that is, halls which deviate from a consensus standard. Only when instruments and halls were standardised could composers such as Liszt write out every single note to be played, and write out acceptable alternatives as well. A good book on this is needed, obviously much beyond the scope of Fabian’s study, but she’s heading in the right direction.

The last half of the book is dedicated to a fascinating discussion of such questions as rhythm ("notes inégale") dotting and double-dotting, tempo, phrasing, and so on, and how these qualities may be discovered for music we’ve never heard played. Fabian and her sources mention the experiments in swing beats by jazz musicians. The curious statement is made that an andante (which is generally taken to mean "walking" or "at a walking pace" in Italian) cannot be syncopated, whereas anybody who enjoys walking will often sing or whistle a syncopated tune in rhythm with his steps. The point that lies just under all this verbiage is that a musical phrase from any age of music expresses the sounds and motions of the human body. If a person wants to know how the phrase should go, try dancing to it, or singing it. Mention is also made of the rhetorical structure of music, that is liking musical phrases to poetry or prose expression. If Fabian’s book weren’t restricted to discussion of Bach performance practice, she might have mentioned Leonard Bernstein’s Norton Lectures series where he discusses this very point using Mozart’s music as an example.

The accompanying disk is not intended to be entertaining, but is fascinating, 76(!) tracks, little snippets of performances, many of them less than a minute long, first of the Goldberg Variations, then Brandenburg Concertos, then the John and Matthew Passions. The variety is enormous and they illuminate various points raised in the book. Perhaps the most important point proved here is that all the versions are musical, the variety of styles a matter of taste or exploration rather than any sense of right or wrong, god or bad. Sound quality in digitisations from analogue sources ranges from good to abominable, but I suspect many of these recordings were available to the author only on ancient worn LP’s or tape copies. In the Goldberg Variations section, neither the Landowska nor the Glenn Gould recordings are included, among the most popular recordings of the century, but also among the least exemplary, in our current sense, of historical performance practice. Of course a hundred such disks could be put together depending on one’s view of who is important and what is most interesting.

In summation, I should say that the author is perceptive and widely informed and her analyses and conclusions are incisive and relevant. Where I disagree or perceive a lapse I have been unsparing of the author, and I hope that she as well as my readers perceive the implicit flattery intended by my having paid good attention and intently discussed and evaluated her research and her points of view.

Another point is simply to large to be ignored. The printing press created a revolution because it made printed books so much cheaper to own than hand-copied manuscripts. This book is priced to sell for £55. Were I a friend and professional colleague of Ms. Fabian, perhaps she would allow me to make a photocopy of her thesis; I could probably accomplish this for £10, including the CD. Or, she could make for me a software copy on two CD-Rs for less than £0.70. Printed books of this type are in danger of becoming an expensive anachronism, just like hand-copied manuscripts became in the Sixteenth Century. Just what the future holds none of us knows, but considering the economic pressures involved, we sit on the verge of an explosion.

My feeling is that the omission from consideration of the Scherchen Brandenburg Concerto recordings is sufficiently material that Fabian should publish a journal paper revising any of her conclusions as made necessary by the inclusion of the Scherchen recordings, and should make the appropriate revision to future editions of this book.

Paul Shoemaker

A detailed and fascinating look at one of the most interesting periods in the history of music which answers as well as raises a number of interesting questions. ... see Full Review

Advertising on

Donate and keep us afloat



Recordings of the Month


Conner Riddle Songs

Rodzinski Sibelius

Of Innocence and Experience


Symphonies 1, 2, 3



Aho Symphony 5

Dowland - A Fancy


Rachmaninov_ Babayan


Return to Index

Untitled Document

Reviews from previous months
Join the mailing list and receive a hyperlinked weekly update on the discs reviewed. details
We welcome feedback on our reviews. Please use the Bulletin Board
Please paste in the first line of your comments the URL of the review to which you refer.