### Cardinality (music)

Six-element set of rhythmic values used in Variazioni canoniche by Luigi Nono

A set (pitch set, pitch-class set, set class, set form, set genus, pitch collection) in music theory, as in mathematics and general parlance, is a collection of objects. In musical contexts the term is traditionally applied most often to collections of pitches or pitch-classes, but theorists have extended its use to other types of musical entities, so that one may speak of sets of durations or timbres, for example.

Prime form of five pitch class set from Igor Stravinsky's In memoriam Dylan Thomas

A set by itself does not necessarily possess any additional structure, such as an ordering. Nevertheless, it is often musically important to consider sets that are equipped with an order relation (called segments); in such contexts, bare sets are often referred to as "unordered", for the sake of emphasis.

A time-point set is a duration set where the distance in time units between attack points, or time-points, is the distance in semitones between pitch classes.

## Serial

In the theory of serial music, however, some authors (notably Milton Babbitt) use the term "set" where others would use "row" or "series", namely to denote an ordered collection (such as a twelve-tone row) used to structure a work. These authors speak of "twelve tone sets", "time-point sets", "derived sets", etc. (See below.) This is a different usage of the term "set" from that described above (and referred to in the term "set theory").

For these authors, a set form (or row form) is a particular arrangement of such an ordered set: the prime form (original order), inverse (upside down), retrograde (backwards), and retrograde inverse (backwards and upside down).

A derived set is one which is generated or derived from consistent operations on a subset, for example Webern's Concerto, Op.24, in which the last three subsets are derived from the first:

```B B♭ D E♭ G F♯ G♯ E F C C♯ A
```

Represented numerically as the integers 0 to 11:

```0 11 3 4 8 7 9 5 6 1 2 10
```
Webern's Concerto Op. 24 tone row, composed of four trichords: P RI R I

The first subset (B B D) being:

```0 11 3 prime-form, interval-string = <-1 +4>
```

The second subset (E G F) being the retrograde-inverse of the first, transposed up one semitone:

```  3 11 0 retrograde, interval-string = <-4 +1> mod 12

3  7 6 inverse, interval-string = <+4 -1> mod 12
+ 1  1 1
------
= 4  8 7
```

The third subset (G E F) being the retrograde of the first, transposed up (or down) six semitones:

```  3 11 0 retrograde
+ 6  6 6
------
9  5 6
```

And the fourth subset (C C A) being the inverse of the first, transposed up one semitone:

```  0 11  3 prime form, interval-vector = <-1 +4> mod 12

0  1  9 inverse, interval-string = <+1 -4> mod 12
+ 1  1  1
-------
1  2 10
```

Each of the four trichords (3-note sets) thus displays a relationship which can be made obvious by any of the four serial row operations, and thus creates certain invariances. These invariances in serial music are analogous to the use of common-tones and common-chords in tonal music.

## Non-serial

Major second on C    .
Minor seventh on C    .
Inverted minor seventh on C (major second on B)    .

The fundamental concept of a non-serial set is that it is an unordered collection of pitch classes.

The normal form of a set is the most compact ordering of the pitches in a set. Tomlin defines the "most compact" ordering as the one where, "the largest of the intervals between any two consecutive pitches is between the first and last pitch listed". For example, the set (0,2) (a major second) is in normal form while the set (0,10) (a minor seventh, the inversion of a major second) is not, its normal form being (10,0).

Rather than the "original" (untransposed, uninverted) form of the set the prime form may be considered either the normal form of the set or the normal form of its inversion, whichever is more tightly packed. Forte (1973) and Rahn (1980) both list the prime forms of a set as the most left-packed possible version of the set. Forte packs from the left and Rahn packs from the right ("making the small numbers smaller," versus making, "the larger numbers ... smaller"). However, these only differ in five instances and are the result of different algorithms (Rahn's being preferred by programmers).