Three Essays on the Fundamentals of Piano-Playing


While timing and dynamics are completely independent elements in principle, in practice they combine intimately, in particular, in controlling the sense of forward movement. This applies both to language and to music and it is perhaps illuminating to discuss them together.

In general the sounds of speech and music come to us as a sequence of more-or-less separated pulses, - syllables or notes. Some pulses are emphasized more than others. The emphasis is often dynamic; English and German have stressed and unstressed syllables. But Latin languages have very little stress; the distinction is made between long and short syllables, a matter of timing. All these languages have short and long vowels and consonants. In music the barline marks the strong pulse, and the downbeat immediately following is normally the strongest. Rhythm is the pattern of pulses of varying strength. It has three components: tempo, meter and phrasing.



Tempo refers to the speed or frequency with which the pulses occur. In metric verse and in Western music, with some exceptions, strong pulses occur with strict regularity, but even when they do not, as in prose, one has a sense of an average speed. A constant, unvarying tempo is uninteresting: only changes of tempo attract the listener's attention. An increase in speed is stimulating, a decrease relaxing. Actors and professional speakers constantly vary the speed of their words. In print, a sequence of short vowels and consonants seems to move faster than a sequence of longs. In music, a basic tempo is usually established for a given section of a work, but may change abruptly from one section to the next. Accelerandi and ritardandi maintain interest within sections as does, at a finer scale, romantic rubato.

The frequency of strong pulses establishes the primary or basic tempo, but other tempi can coexist with it. In music, two notes per beat seems to move faster than one note per beat, and three or more even more so. The rate of harmonic change also contributes to one's sense of speed. Charles Rosen, in "The Classical Style" (Norton, 1972, pp.229-233) gives an excellent, detailed example of how Mozart exploits various subordinate tempi to give the effect of increasing speed while maintaining the basic pulse unchanged.



In poetry, other than free verse, and in most Western music, each strong pulse is accompanied by a fixed number of weaker pulses. In music, 2/4, 3/4, 4/4, 5/4 meters have, respectively, one, two, three, four weak beats following each strong. (The weak beats are not necessarily of equal strength, but normally vary somewhat.)  In poetry, it is usual to have one or two weak beats with each strong, and it is customary to distinguish the possible ways in which they are grouped. A foot of one weak and one strong may be either iambic, where the weak is attached to the following strong, or trochaic, where it is attached to the preceding. The dactyl has a strong followed by two weak beats, the anapest a strong preceded by two weak beats, while the rarer cretic has a strong between two weak beats. Similar distinctions are present in music, but it is more appropriate to discuss them under "phrase rhythm."

The essential element of meter is its regularity; strong pulses occur at a constant rate. In the usual analogy, it can be considered mechanical like the clickety-clack of a moving train or the tick-tock of a clock, as opposed to the vital like a heart-beat, which is always varying slightly and very much when under physical or emotional stress. Meter is predictable and, once established, sets up an expectation in the listener, against which variations of tempo and phrasing make their maximum effect. Meter is not always present.  Prose has none, and in a musical fantasy, toccata or recitative, there may be only free rhythm without an established meter.

Like a constant tempo, an unvaried meter can soon become monotonous, but once a particular meter is fixed in the listener's ear, variety can be introduced by accents, syncopation and other means. It is thus important, but not always easy, to establish the meter of a given section right at the beginning. This listener is soon irritated if he has to wait for several bars before he knows whether a piece is in duple or triple time, or where the barlines fall.

Common time and alle breve (which could be written 2/2, but never is) both have four quarter notes to the bar. They differ in the relative strengths of the weak beats. In alle breve the second and fourth beats are weakened relative to the third, to the point that they are no longer heard as beats at all, but as filler notes between beats. The distinction can be a delicate one. The opening of Beethoven's Moonlight sonata, marked alle breve, frequently sounds like 4/4, with right-hand thumbs coming down equally on the successive G-sharps. In 4/4, the weak beats are rarely strictly equal, but 3 must never completely dominate 2 and 4. Considerations of this kind makes one realize that time signatures, conventional and somewhat inconsistent as they are, represent only a very coarse grouping of possible meters. In 6/8 the ear simultaneously hears two to the bar and six to the bar. In any meter, one hears the pulse of the downbeat, the pulses of the beats within the bar, and possibly sub-beats within the beat, if present. Two or more notes on each beat, in fact, establish a fast, additional meter with each beat acting as a sub-bar, so to speak, and individual notes as sub-beats.

Filler notes between beats, in general, have to remain subordinate to beats or they could impose a different meter on the listener. As an example, consider repeated eighth-note chords in an accompaniment in a 4/4 meter. If the chords are played with identical strength, one will hear eight beats to the bar instead of four. Again, a very delicate distinction; the off-beat chords must be weaker, but ever so slightly so, because one does not want to attract attention to it. In such a situation one is usually trying to give an impression of continuous, constant sound. Counting the beats aloud while practising is a device sometimes used to evoke the proper shading.

The feet of verse and the bars of music represent what may be called the basic meter. But repeated notes on a beat can establish a simultaneous subordinate meter with each beat acting as a sub-bar. In the other direction, the feet of verse and the bars of music are usually combined into metric groups. The primary grouping of verse is the line. All lines may be the same length, as in pentameter, or if they vary in length, they tend to follow a fixed, repetitive pattern. Lines are grouped into verses, usually of the same number of lines. Bars of music are also grouped. The classical period used groups of different length - 3,4,5,6-bars - but by the romantic period, the four-bar group was dominant. In Chopin, it is hard to find anything else. Just as repeated notes on each beat can establish a fast meter subordinate to the main one, a regular grouping of bars superimposes a slow meter superior to the basic one. Again these groups are combined in a nested hierarchy of groupings: the section, the movement, the whole work. This metric hierarchy provides the temporal framework of poem or musical work, in other words, the primary element of its form.

Once a meter has established expectations in the listener, any pulse, whether on- or offbeat, can be emphasized in some way for a surprise effect. Any such departure from the established pattern is an accent. In Shakespearian pentameter "to b, or nt to b / tht is the qustion," the word "that" receives a strong stress where a weak one would normally occur. This puts great emphasis on the word. An accent is usually louder than expected, but all that is needed is that it be different. An unexpectedly soft note is also an accent. Not only dynamic but temporal accents are possible. A note can be unduly prolonged - a "drift point" - or delayed - "placed." A note of music can be ornamented; in speech, the tone or pitch of the voice may be altered. In music, if off-beats are regularly accented, we have syncopation, but the basic meter must be preserved, or the listener will begin to hear the accented off-beat as a downbeat and the effect is lost. Accents are useful for clarifying meaning and for sustaining interest.


Phrase Rhythm

Although prose has no meter, it does have its own hierarchy of rhythmic groups, which are basically determined by meaning and the organization of subject matter. The basic unit is the phrase, e.g., a noun and its qualifiers, a prepositional phrase, a compound verb group. A phrase consists of just a few syllables and may have as few as one, as in an interjection. Phrases are grouped into clauses, these into sentences, then we have successively paragraphs, chapters, volumes. In print, groupings are indicated by punctuation and white spaces on the page. In speech, however, the end of a group is normally marked by a pause, with longer pauses for the larger groups. The shortest recognizable pause, occasionally necessary to separate syllables, is the glottal stop, a momentary interruption in the flow of breath. Less common in English than some other languages, it does occur. It distinguishes "meat-rack" from "me-track."

The same rhythmic groupings, determined by grammar and meaning, occur in poetry as well as in prose, and are quite independent of the metric structure. Both within a metric group and within a rhythmic phrase, we have a sense of continuing forward motion that comes to a stop at the end. Because phrasing and meter do not necessarily coincide, there is tension between the two. One special case of the conflict has been honoured with a special name. If the end of a sentence coincides with the end of a line of poetry, we have a definite stoppage of flow through mutual reinforcement. But if the sentence continues into the next line - "enjambement" - the movement within the sentence carries us past the line end without stopping. Thus, whenever a rhythmic phrase and a metric group end together, the flow stops; otherwise the movement in one carries us over the break in the other. Northrop Frye in "Anatomy of Criticism" (Princeton 1957, p.251) points out another form of the tension between "recurring rhythm" (meter) and "the semantic rhythm of sense." He remarks that iambic pentameter in English usually has only four stressed syllables in a five-beat line. His example is the first nine lines of Hamlet's "To be or not to be" speech.

In music, a melody is normally broken up into a sequence of rhythmic phrases which are largely innate in the melody itself. (In song, some phrasing is imposed by the need to breathe.) Like the semantic phrases of words, the rhythmic phrases of music are independent of, and interact with the meter. While a performer may have some leeway in how he phrases a melody, the choice is not arbitrary, but must be based on his insight into the melodic structure. In song, the words also have a major influence. Unfortunately the use of slurs by composers and editors is both multipurpose and inconsistent, so that the printed score is of little help in this respect.

The average length of the basic melodic phrase is one or two bars but may, on occasion, be much shorter or longer. Like metric groups, rhythmic phrases may themselves be combined into larger groups so that a parallel hierarchical structure exists. As in poetry, rhythmic and metric groupings do not necessarily coincide, but the larger ones tend to do so, as at the end of a section in a sonata, or the end of a stanza in a song. (A minor deviation. When a new section begins with an anacrusis the metric group, by definition, begins on the barline, the rhythmic group just before.) Just as in verse, whenever a rhythmic phrase and a metric group end together, we get a stoppage of movement; otherwise, the movement within one carries us past the end of the other. This is particularly important in music as a means of controlling forward motion in time.

How do we recognize the end of one phrase and the beginning of the next? The usual sign is a pause or break in the flow of sound, but there may be other indications. In speech it is usual to drop the voice, i.e. lower the pitch and perhaps the dynamic level, at the end of a sentence, and raise it after a question. Pauses always separate larger sections where metric and rhythmic groups coincide. At the end of a movement a pause of several seconds is appropriate; long enough for performer and audience to relax but not so long that conversation starts. The ends of major sections are often marked by slowing down, by ritardandi. Smaller groupings are separated by shorter pauses. Although the form of a work is determined in language by the organization of subject matter and, in music, by harmonic and melodic structure, it is the pauses between the larger sections that make the form clear to the listener.

In spoken language, even in metric poetry, it is still possible to mark off the smaller phrases by very brief pauses; slight variations in pitch or dynamics may contribute but are not primary. In music, a pause at the end of every rhythmic phrase, no matter how brief, would soon destroy the sense of meter. Fortunately, there are both temporal and dynamic ways of separating rhythmic phrases.

One can still use a very slight pause, but only occasionally. It is particularly useful when one wants to "finish the sentence," as it were, and the slight delay acts as a period. But one can also use a break in the flow of sound without any distortion of meter. The last note of a legato phrase can be detach; the last note of a detach or staccato phrase can be shorter than the preceding ones. A brief lift of the pedal can establish a separation between phrases. A sharp change in dynamic level can also mark the beginning of a new phrase. Usually the change would be a sudden slight increase in level, introducing "new energy," but a sudden drop in level also works. Whatever devices are used, singly or in combination, they need not be pronounced and would normally be so slight that they do not draw attention to themselves, like accents do.

Every melodic phrase moves to a peak or apogee and then falls back. Of course the apogee could fall on the first or last note of the phrase. Normally this is achieved by subtle dynamic shading; the sound level increases to the apogee and then decreases. The reverse dynamic in which the apogee is the quietest note is also conceivable, but difficult to pull off. The apogee can also be accented temporally or dynamically, but not so much that it breaks the continuity of the phrase. A very long phrase can have a secondary apogee, not as pronounced as the primary one. This is perhaps equivalent to saying that a long phrase can have sub-phrases but, again, the continuity of the long phrase must be preserved.

While one usually thinks of phrasing in connection with melody, it also applies to passage-work and even to repeated rhythmic patterns. One principle to be followed throughout is that of always maintaining motion across the barline. Hence, one always needs a rhythmic phrase that carries over the barline, except at the end of a section where one wants the effect of a full stop.

At the simplest level consider one-bar phrases in a 3/4 meter. If they coincide with the bar - 123, 123, 1- where the comma indicates a phrase ending, it will reinforce the metric stop at the end of a bar; one gets oom-pah-pah music that goes round and round with no forward motion. On the other hand - , 231, 231, 2- will consistently carry the ear over the barlines, as will , 312, 312, 3- the Viennese waltz. In 4/4 - 2341, 2341, 2- and 4123, 4123, 4- both work well, as does the less common 3412, 3412, 3-. This is analogous to the classification of 'feet' in poetry.

How does one separate short phrases at this level? Primarily dynamically, although a slight break in the continuity of sound can also be used. For example, if there is a sharp drop in dynamic level between 1 and 2, followed by increases in subsequent weak beats back to one, we get a rhythmic phrase beginning on 2. The principle applies at an even finer level. If there are three notes to each beat, as in a triplet accompaniment, or four, as in an Alberti bass, one can think of each beat as a sub-bar with three or four beats to the sub-bar and phrase over the sub-barline accordingly.

The various rhythmic components can interact in extremely complex ways. The tension between meter and phrase rhythm is almost always present, even in a single melodic line. But most music presents one, two or more vocal lines, with or without accompaniment. It is thus possible for different vocal lines and the accompaniment each to have its own meter and its own phrasing. In addition, both meter and phrase rhythm form a hierarchy of levels. Meter is based on the bar, broken up into sub-bars or beats, and combined into groups. Rhythmic phrasing too, occurs simultaneously at a number of levels.

The interpretation of a piece of music is not determined by the above principles, but by the taste and knowledge of the performer. Nevertheless, an understanding of those principles has been of enormous help to me not only for the acquisition of technique, but for interpretation, and may be so for others.



Published by Sandstone ePress