Phrasing in classical music
Phrasing comes after rhythm, intonation, tone quality (timbre), dynamics, articulation and accentuation, in describing how a piece is performed. Phrasing means how all of these qualities are coordinated to realize musical meaning, on a timescale ranging from a few notes to a few bars. It is second only to the sense of the overall form of the piece, which one might call 'long-range phrasing'.
All classical music is made up of phrases, which have a similar function to sentences in prose. The fundamental principle of phrasing is this:
In good phrasing, there is one main accent per phrase.
(This principle comes from
Hans Keller and is borne out by my own experience. Curiously,
prose has a similar rule: one main verb per sentence.)
Several questions remain: where do phrases begin and end? Where should the main accent be? How should the main accent be performed? What if there are other accents in the phrase: how are they to be treated? Finally, how can the form of individual phrases fit into the overall structure of the piece?
Where's that phrase?
If you are lucky, the composer will have put phrasing marks on top of the music, which look like big slurs or ties. But, it is not unknown for composers to abuse these markings. More likely, there will be no explicit phrase markings, only slurs which indicate legato articulation, etc., dynamics (hairpins) and accent marks, which may look like this: > or this: sf or something else. In the case of 18th century music, there may be nothing at all to help you.
One mistake some people (especially string players) make is to treat slurs as phrasing marks: they diminuendo at the end of a slur and cut off the last note, as if it were the end of a phrase, when the musical sense actually continues (or tries to). The result, chopped-up phrasing, is highly frustrating, since in order to make sense of the phrase the listener has to contradict his or her own ears. A slur is an articulation marking, not a phrase!
Usually, a good indication of where a phrase ends is that you should be able to stop at the end of it and feel that something has been said: at the end of a phrase there is a momentary sense of release of tension. However this doesn't mean that all phrases have silence at the end - sometimes the end of one dovetails with the beginning of the next. If you can't decide whether some passage consists of one or two phrases, don't worry: smaller sub-phrases can coexist within large super-phrases. More on this later.
Phrases can be as short as two notes (e.g. the start of Brahms's 4th Symphony) or extend over several bars (e.g. the slow movement tune of Beethoven's 4th Symphony). So, putting an accent on the first beat of the bar isn't always a good idea - unless you want to sound like a dance band.
Put your finger on the button...
Having marked off the phrases, locate the main accent. Sometimes the composer makes it easy for you with a sf or a pair of hairpins < >; sometimes there are lots of accents and dynamics within the same phrase, sometimes none. The general rule is that within the phrase, tension increases up until the main accent, and decreases afterwards.
Again, the possibilities of placing the accent are manifold. It may fall on the first note, with everything afterwards being a kind of echo: for example, the start of the Hallelujah chorus in Messiah, or the first two bars of Eine Kleine Nachtmusik. This is called a downbeat phrase. Or, the accent may fall towards the end of the phrase, for example, "For the Lord God omnipotent reigneth" from the same chorus, or the first two bars of Beethoven's Fifth. This is called an upbeat phrase.
Beethoven's Fifth
is a great example of the danger of disregarding the principles of phrasing. Too many performances put two accents on the phrase (DA da da DAAA), making it sound like a triplet, or four (DA! DA! DA! DAAA!), making it sound like a deranged, ranting dictator.
These examples bring us to the point about sub-phrases: each main phrase may contain more than one group of notes, each of which functions as a miniature phrase carrying some of the energy of the whole. Like clauses marked off with commas inside a sentence. The principle applies both within the sub-phrase and within the main phrase. So let's look again at Eine Kleine Nachtmusik:
DA! - dat Da - dat Da! da da da Da -
There are three subphrases: each has its own 'main' accent, the second one being an upbeat phrase, and the third one having a subordinate accent on the last note; within the whole phrase there is a clear hierarchy of accents falling away from the first note. Note the release of tension you feel after the end of the whole thing, and also the feeling that each of the subordinate accents echoes the main accent.
Now, if you know Beethoven's Fifth, think past the loud introduction to where the main Allegro movement starts. There are lots of little four-note phrases (da da da Daaa) passed between the different instruments, but they fall into groups of three that fall away from the accent of the first sub-phrase: a downbeat super-phrase built out of upbeat sub-phrases! Further on, after the second fortissimo outburst, Beethoven puts groups of four sub-phrases together, where within each group the tension peaks at the third: an upbeat super-phrase. In this way he achieves great variety of expression, although the basic cell is the same.
What does sforzando mean?
In later Romantic and 20th century music,
sf means a sudden, forceful accent: something that sticks out. However, this can be very misleading in Classical and early Romantic music, because the marking had different meanings depending on the context.
Sometimes, if used on a syncopation or unexpectedly dissonant chord, it does indeed mean a 'surprise' accent, maybe even preceded by a slight pause, as taught in all the best conservatories. Much of the time it means "The main accent of the phrase is here, not there!" (This point also due to Hans Keller.) In other words it contradicts where you might have thought to put the accent. The main use in this context is for upbeat phrases that start loud and get louder - for which it is essential not to put an accent on the first note. So it means you have to look ahead to see where the accent is.
If the phrase is legato or at all lyrical, then it's probably a bad idea to put a conservatoire-style sore thumb sforzando on it. There is a good example in Beethoven's violin sonata Op. 30 no. 3 where there is a sf near the end of the lyrical second subject: it doesn't mean "Put a big, sudden accent here", rather "This phrase is an upbeater and builds up to this point, in contrast to the previous downbeat phrases." At the Aspen Music Festival this year, I heard this theme executed to death by an otherwise excellent violin/piano team who applied the standard 'minute pause and sharp accent' formula. So much practice for such an unmusical result.
So that's the inside scoop on pre-Romantic sforzando. But the principle applies to every type of music: there are many different types of accent, from a minute dynamic gradation to a musical hammer-blow that rearranges the audience's internal organs, and it's up to the performer to find out which will best suit the phrase in question.
The Beethovenian anti-accent
Beethoven often took the course of writing a crescendo up to the main accent, but then marking the note itself piano. Then the sudden decrease in volume, rather than the expected increase, has the psychological effect of an accent. The performer has a duty not to spoil the surprise by cutting off the crescendo too early - just enough to prevent the piano being drowned by resonance. Too many chicken out...