fugue-writing, Shostakovich, and contrapuntal mistakes

As I have immersed myself in contrapuntal music over my life I have made many enjoyable discoveries. I have also made discoveries that were not so enjoyable. I have enjoyed the contrapuntal works of Bach, Haydn, Beethoven, Hindemith, and I have been fond of works by Igor Rekhin and Durufle. I have even enjoyed contrapuntal music written by Messiaen and Bartok, all great composers.

I have been disappointed, I admit, to discovery that virtually no practitioners of counterpoint since Bach seem to have any use for countersubjects. It would seem that the whole art of counterpoint can be summed up in one's capacity to not merely create subjects but countersubjects that allow for the melodic lines to be exchanged and sometimes even inverted. If you invert your subject and then invert the countersubject then if you've really mastered contrapuntal writing having the subject and countersubject in the SA format (soprano subject, alto countersubject) can yield a pleasing music idea when you invert the subject and put it in the alto voice while inverting the countersubject and putting that in the soprano voice.

But some of my disappointments are more specific than that. My specific recent disappointment has been the result of checking out the 24 preludes and fugues of Dmitri Shostakovich. NOw I love his string quartets, every last one of them, and I enjoy many of the symphonies. His contrapuntal music, however, is frequently very disappointing! I will attempt to explain why as far as is possible on a mere blog where you can't hear any of what I'm talking about.

As I have been composing my own preludes and fugues for solo guitar I have been forming some general rules, rules of thumb, about how to approach the art. One guideline I have discovered is Hindemith's simple observation that the human mind cannot keep track of more than three separate musical lines at once and that by and large any fugue that has four or five voices will invariably have two voices that are basically yoked together, sometimes three. So the application of this observation for guitarist is simple, even in many three-voiced fugues complete linear independence rarely happens. This means that it is wise to imply rather than insist on spelling out all melodic and harmonic developments.

Now straightaway I can say that another application of this observation by Hindemith is to have it piggyback on observations by George Oldroyd. Oldroyd said that many students of counterpoint mistake hustle and bustle for true contrapuntal art. If you create a subject and your countersubject or your counterpoint in response to the answer keeps the same level of melodic or rhythmic activity then you have taken an easy path. It is wise to have your counterpoint embellishing your answer or, better yet, your countersubject written in a way that both complements and contrasts with your initial ideas in the subject. Shostakovich often does not do this.

Shostakovich also too often displays another weakness in people attempting counterpoint, he often goes for the largest number of voices in an exposition he thinks he can pull off with his pianistic skill. Too many people see counterpoint as the art of doing as many things as possible. That is not what counterpoint really is. Counterpoint is the artof doing several musical things simultaneously AS CLEARLY AS POSSIBLE.

Shostakovich has several fugues where he would have done well to limit himself to three voices but he went for four. He frequently does this. That's not clarity for its own sake but technical achievement for its own sake. Now I'll be quick to admit that there's a lot of pride one can take in pulling off even bad counterpoint but the art involves not merely a flurry of melodic and rhythmic activity it also involves penetrating harmonic thought. Bach, Beethoven, Haydn, Bartok, Hindemith, all these composers found ways to more tightly balance the harmonic and linear aspects of their counterpoint. As George Oldroyd might have put it, they had a harmonic PLAN that guided the way melodic activity was developed in the exposition.

Now it is true that in music you learn the rules so that eventually later you know when and how to break them. That said I believe that the loss of the countersubject as a cultivated art in fugue-writing may on its own account for a proliferation of voices beyond what is pleasing. I say this because as I have composed fugues for solo guitar even for my own instrument's limits as I write for it I have settled into a broad observation: the point at which adding a new voice to my contrapuntal texture makes it impossible for me to preserve invertible counterpoint is probably the point at which I have added too many voices.

How do I get the idea I have added too many voices to a contrapuntal texture? Well, let's assume for the sake of discussion that I have a suitable subject for a fugue. It has a couple of unique and memorable ideas that allow for great musical development. Let us assume that I also have a countersubject that does not break any rules of counterpoint with its subject in either the answer or the subject in its original form (keep in mind that countersubjects can be transposed so as to avoid breaking rules about linear movement).

Let us assume that I have a countersubject that works perfectly well in counterpoint with both the subject and the first countersubject when I look at the melodic lines as pairs. That is to say that the subject (S) and (countersubject 1 (CS1) work well. Let us also assume that countersubject 2 (CS2) and the subject work beautifully together. Well, suppose that I go through every combination of S, CS1 and CS2. All the counterpoint is looking wonderful right up to the point where I test out the counterpoint of the formulation CS1, S, and CS2 in descending order. Then I discover, perhaps, that there are perfect parallel fifths between CS1 and CS2 on strong beats.

Now if the parallel motion happens on a weak beat or the weak part of even a stronger beat all is not lost. Depending on the prevailing pulse of the subject and its rhythmic activity this parallelism may be resolved so quickly there is no problem. But if parallelism exists on strong rhythmic accents or parallels not only the linear part of a melody but even its rhythmic profile then if I have parallel perfect fifths between S and CS 1 then my counterpoint is not fully invertible. I have a few options at this point. The blunt advice I often give to myself is that if I'm writing at the keyboard this would actually be the point where I need to keep revising CS2 until I avoid this problem.

If after multiple revisions I can't avoid parallelism in octaves of fifths then I have very probably attempted to add a voice to a contrapuntal texture that goes beyond the boundaries of good sense. In other words, the perfect fifths that result from adding even one more voice than the contrapuntal texture can sustain in all contrapuntal derivations is literally the math that shows I can't afford to add another voice.

Now speaking as a guitarist I am acutely aware that in many cases the guitar is so limited in what may be played as counterpoint that this is a thoroughly academic issue. CS1, S, and CS2 together may be completely beyond the pale of being played on a single guitar. Even so that would be a reason to drop CS2 if the problem of parallel fifths can't be avoided. CS2 would have to add some exceptional musical value to be worth keeping.

Now there is a fugue for guitar where I have a CS2 that ends up creating parallel fifths against the S. The reason I'm keeping it is because the parallelism happens on the weak part of a weak beat in the prevailing meter. I'm also deliberately playing with polymetric elements in a subject that could best be described as in 11/4. So I'm not ignoring that certain things are less than ideal counterpoint but are so fun to play they are worth the trouble. I'm just proposing that the oft-ignored art of testing out countersubjects against all possible assignments of melodies to the voices planned for a fugue allows you to winnow out ideas that don't really help you.

Writing a fugue exposition with more voices in the texture than you need is a really big deal. It disappoints me that I feel obliged to say that Shostakovich seems to habitually ignore even the possibility of adding one too many voices to a texture. Let's face it, most of us won't even get to Haydn's level of writing fugues, let aone J. S. Bach. Even fewer of us will begin to approach the contrapuntal mastery of Tallis or Palestrina or Josquin. Then again, their entire conception of harmony is radically different than those who have composed since the advent of major and minor keys but that's a separate subject for people who actually specialize in music from that period. I would defer to any Tallis scholar on that issue, whether the singers or actual scholars of Thomas Tallis' wonderful music.

I admit that, right now, it seems positively absurd for me to be going on at such length about counterpoint. I'm a guitarist. Most guitarists don't even tackle transcriptions of Bach's violin fugues. I have often wondered how many guitarists actually understand counterpoint well enough to recognize the difference between musical busy-work and actual contrapuntal control of a musical texture. Brouwer can obviously handle contrapuntal writing. Sor was brilliant at it when he did do it. I wish he had been a little more ambitious in tackling counterpoint myself but I'm grateful we have what we have. Mark Delpriora took care of transcribing the closing fugue from Haydn's creation inspired by some prodding from Ophee and the comment Sor made about how it was possible to play Haydn's double fugue on the guitar even in B flat major.

As I wrote on this blog a while back too many guitarists mistakenly equate homophonic music (a melodic line with some kind of independent accompaniment) with contrapuntal music. Guitarists who wish to study counterpoint had probably best devote a great deal of time not to guitar literature but to keyboard lierature and choral literature.

Bach's fugues for violin, often referred to by guitarists, are not particularly accurate depictions of how Bach USUALLY handled fugue writing. They are emblematic and utterly worthy of performance but it is easy for guitarists to have wildly inaccurate misconceptions about how Bach really handled counterpoint. Bach's violin sonatas usually at most suggest or imply rather than state. In the ears of many guitarists these fugues may be the touchstone for what guitar fugues should sound like. No offense to violinists but the guitar is capable of more truly independent linear development than the violin will ever be. We guitarists are more limited than pianists but we are not as limited as violinists are in terms of what we can do to have three independent lines on one instrument.

Chris Kachian correctly pointed out that sustained three-voice counterpoint where all three lines are independent is not possible on the guitar. Having studied quite a few fugues in my life I can cheerfully report that even Bach rarely assigned complete independence to all voices even in his three-voiced fugues. Just look at how much CS 1 and CS2 lean on each other in his famous C minor fugue.

With this observation in mind a three-voiced contrapuntal texture is practical on the guitar even for sustained periods once we realize that even in an ostensibly contrapuntal texture there will be room for homphonic elements (i.e. a harmonic plan and some liberties in using two voices so form rhythmic aaccompaniment or punctuation to the voice of primary melodic interest). If in a three-voiced episode we have an antiphonal relationship betwen the soprano and alto then the tenor can provide something as simple as rhythmic or harmonic punctuation marks to the alto voice.

As George Oldroyd so elegantly if stuffily put it, we must remember that there is no need for all the voices in a fugue to be active at the same time. In fact there's no need for even a three-voiced fugue to have constant activity. Even a simpler texture needs room to breathe. This is why I believe most pianists can utterly fail to grasp counterpoint as well as they could because they do not see that counterpoint is an art that rose out of choral rather than instrumental literature.

Now that I am at 19 of 24 preludes and fugues I can say that I have only written one fugue that even has four voices in its exposition. I shall probably write a couple of fugues that have only two voices (my F sharp major fugue has just two voices, for instance).

I don't particularly enjoy writing at such length and in such a general way about why I believe the fugues of Shostakovich have so far been rightly neglected amongst concert-goers and performers alike. The demands Shostakovich makes on my attention through his works are out of proportion to the rewards of considering how he approaches musical textures. I have listened to Ludus Tonalis dozens of times at least and have pored over the scores with great contentment. I have often found myself, despite repeated efforts to give Shostakovich careful attention mentally checking out through much of what he wrote in his fugues. He gets my interest with about half of his subjects, often losing my interest as soon as his exposition is under way because the material he selects for his answers seems to be busy-work.

What little I can recall of biographies of Shostakovich suggest that he disliked fugue writing in general and was not even particularly thrilled about Bach's works. He generally seemed to view contrapuntal writing as drudgery in older styles. Even if he changed his mind and my memory of this field of study is rusty I'm afraid I have to say that listening to Shostakovich's actual fugues has re-convinced me that this works are decidedly lesser entries in his work as a whole. They sound unusually dashed off and what works in sonata allegro form or other Classic era forms can be dangerously staid and lazy in a compositional type with roots in the Renaissance that reached its apotheosis in the Baroque era.

If the fugue were to be thought of as not merely a contrapuntal form or even a contrapntal style as such but a way of musical thinking that guides musical feeling then I'm sad to say that one of my favorite composers just does not think much in the way a fugue requires a person to think. Now a composer like Joan Tower, what I've heard of her, I think could think through a fugue and write a nice one. The aforementioned Bach, Haydn, Hindemith, Bartok, Messiaen, and Durufle all wrote great fugues. Stravinsky was capable of writing some fine fugal passages even though full-blown fugues were not really his thing.

Now this is probably one of my hang-ups or strengths depending on who's talking with me but I do not see music as merely emotional communication. Music is capable of communicating not merely emotional content but ideas. If you only listen to a fugue with an ear for how it makes you feel then the very nature of the musical art will be likely to befuddle and frustrate you. Bach's works fell into temporary disrepute because his works were considered stuffy and academic. The problem was not so much with Bach's work itself as that as the styles and times changed people stopped being able to think with Bach in his musical thinking. Both serious study of biblical literature and serious study of counterpoint require what I would sloppily call exegesis. Much as I love other works by Shostakovich he has not convinced me that he got very far in understanding counterpoint.

Oh well, the fact that his cycle is hardly a staple of the concert circuit means I don't have to be that wound up in my disappointment. His string quartets are still, arguably, the most important cycle of quartets to have been composed in the last sixty years. Sure, I love Bartok's quartets and half of Hindemith's quartets are great. Music critics may say that Ferneyhough's quartets and Carter's quartets are important but, well, they're important to music critics. There IS a place to say "show me the money". People are willing to part with their own money to hear Bartok and Shostakovich more than Carter or Ferneyhough or even the quartets of Britten or Tippett (which, mind you, aren't bad). But now I've digressed into quartet literature.