Erratum

The Evolution of a ‘Memeplex’ in Late Mozart: Replicated Structures

in Pamina’s ‘Ach ich fühl’s’

Journal of the Royal Musical Association, 128 (2003), 30–70

Steven Jan

This article appeared in the above issue but the uncorrected version
of the music was published in error. Therefore the complete article
follows in full.
Journal of the Royal Musical Association, 128 (2003) © Royal Musical Association

The Evolution of a ‘Memeplex’ in Late

Mozart: Replicated Structures in Pamina’s
‘Ach ich fühl’s’
STEVEN JAN

1. INTRODUCTION: INTERTEXTUALITY THROUGH REPLICATION

Any text is constructed as a mosaic of quotations; any text is the absorption

and transformation of another . . . In the space of a given text, several
utterances, taken from other texts, intersect and neutralize one another.1

ONE of the most conspicuous manifestations of postmodernism in

musicology is the weakening of the structuralist conception of the
closed and unitary work of art. Its destabilization as a consequence of
the development of the notion of intertextuality – influence, leading
to explicit and implicit connections between works – has been striking
in the past 40 years, and it is now widely accepted that works of art are
sometimes best understood as fluid ‘networks or relational events’,2
rather than ‘closed and static entities’.3 Whether one subscribes to the
Derridean notion of ‘dissemination’, the Bakhtinian concept of
‘dialogic’ relationships,4 or the Bloomian idea of the ‘anxiety of influ-
ence’,5 one is essentially accepting that a work of art can be understood
in terms of what it shares – or does not share – with its antecedent texts
as part of a historical dialectic.
It is perhaps true to say that intertextual connections in music
became more highly charged in the nineteenth century, when the
notion of the work concept – and its attendant call to originality and
progress – was becoming strongly established in Western art culture.6
In the eighteenth century, by contrast, when this notion was rather
more vestigial, musical material was more freely assimilated by

The author is grateful to Julian Rushton for stimulating the development of some of the ideas in
this article, to Douglas Jarman and two anonymous reviewers for perceptive comments on earlier
drafts, and to Graham Cummings for the translation of Metastasio’s ‘Ecco quel fiero istante’ in
Section 6.
1 Julia Kristeva, in Kevin Korsyn, ‘Beyond Privileged Contexts: Intertextuality, Influence, and

Dialogue’, Rethinking Music, ed. Nicholas Cook and Mark Everist (Oxford, 1999), 55–72 (p. 56).
2 Ibid.
3 Kevin Korsyn, ‘Towards a New Poetics of Musical Influence’, Music Analysis, 10 (1991), 3–72

(p. 15).
4 Korsyn, ‘Beyond Privileged Contexts’, 56.
5 Korsyn, ‘Towards a New Poetics’, 6–7.
6 Lydia Goehr, The Imaginary Museum of Musical Works: An Essay in the Philosophy of Music

(Oxford, 1992).
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 331

composers from their own and each other’s works, in a manner

untinged by ‘anxiety’ – one thinks of Bach’s and Handel’s numerous
borrowings,7 and the tradition of Figurenlehre, as only the most
conspicuous aspects of this more generalized phenomenon.
Mozart is one composer in whose works such intertextual connec-
tions are of central importance. As is richly evident from the music, he
was capable of assimilating elements of other composers’ styles and
integrating them into his own – from the echoes of J. C. Bach in such
early symphonies as the D major, K.19 (1765), and paraphrases of
Haydn’s Sturm und Drang manner in the G minor Symphony,
K.183/173dB (1773), to the Handelian Qui tollis of the Mass in C
minor, K.427/417a (1783). More generally, Mozart and his contem-
poraries, living at a time when the universality of style matched the
cosmopolitan society and politics of Europe, partook of a rich lexicon
of distinctive musical figures – including the loci topici discussed by
Ratner, Allanbrook and Agawu8 – which they assimilated from each
other and from their own earlier works. This question of recurrence
in Mozart – made transparent by the crisply articulated nature of the
Classical style, the clarity of his textures, and his frequent highlighting
of the melodic element – is one which has motivated a rich tradition
of research by Mozart scholars for many years, perhaps receiving its
greatest single impetus with the work of Chantavoine.9
While it is relatively straightforward to observe examples of inter-
textual connection, certainly in Mozart, it is perhaps less easy to deter-
mine why such relationships should exist. Even the invocation of
pragmatic considerations and appeals to the weight of traditional
procedures – the pressures to compose imposed upon the musician-as-
servant, the acceptance of the conventions of modelling and imitation
– have the flavour of assertions rather than explanations. To attempt
to offer a new perspective on this question, I shall outline here what
some may regard as a radical reinterpretation of intertextuality, seeing
it not as a guiding principle in itself, but rather as a function and conse-
quence of the ‘meme’. The memetic paradigm, formulated by the evol-
utionary biologist Richard Dawkins, maintains that human culture is
dominated by the replication of units analogous to genes which are
ultimately ‘selfish’ – in Dawkins’s famous phrase – in that they tran-
scend the intentionality of the composer and appear to act to favour
their own survival and evolution over time.
To illustrate this view, I shall examine in memetic terms a particu-
larly clear and focused case of intertextual connections between four
late pieces by Mozart. This investigation will begin from a largely idiom-
orientated standpoint (Section 2), to use a term from Meyer,10 but

7 Norman Carrell, Bach the Borrower (London, 1967; repr. Westport, CT, 1980).
8 Leonard G. Ratner, Classic Music: Expression, Form, and Style (New York, 1980); Wye J. Allan-
brook, Rhythmic Gesture in Mozart: ‘Le nozze di Figaro’ and ‘Don Giovanni’ (Chicago, 1983); V. Kofi
Agawu, Playing with Signs: A Semiotic Interpretation of Classic Music (Princeton, 1991).
9 Jean Chantavoine, Mozart dans Mozart (Paris, 1948).
10 Leonard B. Meyer, Style and Music: Theory, History, and Ideology (Philadelphia, 1989), 24.
332 STEVEN JAN

then will move on, after examination of memetic theory and its appli-
cation to music (Sections 3 and 4), to a dialect-orientated perspective
(Section 5), suggesting how some of Mozart’s material may have come
from particular antecedent sources and how it may have been adapted
in other contexts by later composers. Finally, I shall consider how
certain verbal-conceptual (textual) memes exist in association with the
purely musical memes in these four pieces (Section 6), before sum-
marizing the principal issues raised here (Section 7).

2. MELODIC COMPLEXES IN LATE MOZART

First, I consider the nature of the intertextual connections between the

four Mozart pieces. In an insightful article Eckelmeyer identifies ‘two
complexes of recurrent melodies related to Die Zauberflöte’.11 The first
complex links Pamina’s aria ‘Ach ich fühl’s’ (no. 17), the Lied der
Trennung ‘Die Engel Gottes weinen’, K.519 (1787), the Adagio intro-
duction to the finale of the G minor String Quintet, K.516 (also 1787),
and melodies from the D minor Piano Concerto, K.466 (1785), and is
illustrated in her Example 1, ‘Melodies related to Pamina’s aria’.12
While there are indisputable similarities between the melodies from
the concerto and the other works (particularly in relation to the
opening phrases of the quintet’s Adagio, bars 1–9), they are arguably
of less significance than the connections between the aria, lied and
quintet, and are, it will be seen, not directly relevant to the present
discussion.
Eckelmeyer’s Example 1 shows, by artful alignment, the remarkable
correlation between the melodies of ‘Die Engel’, the Adagio and ‘Ach
ich fühl’s’, but although the considerable harmonic similarities
between the passages are acknowledged,13 they are not explicitly incor-
porated into the diagram, which is essentially linear. My Example 1
represents both melodic and harmonic affinities, and includes a fourth
piece not discussed in Eckelmeyer’s article, Madame Herz’s arietta ‘Da
schlägt die Abschiedsstunde’ (no. 1) from Der Schauspieldirektor, K.486
(1786), to which, it will be seen, the other three pieces are very closely
related.14

11 Judith A. Eckelmeyer, ‘Two Complexes of Recurrent Melodies Related to Die Zauberflöte’,

Music Review, 41 (1980), 11–25.

12 Ibid., 13–15. The second complex, illustrated in her Examples 3(a) and (b), shows ‘Melodies

related to the final chorus, Heil sei euch Geweihten!’ (no. 21, bars 828ff.), and ‘Melodies related to
the Adagio before the chorale’ (no. 21, bars 190ff.), respectively (ibid., 20–1). The melodies consti-
tuting this complex might best be regarded as based upon a (1̂–7̂ . . . 2̂–1̂) ‘changing-note’
archetype/schema in the Meyer/Gjerdingen sense; see Robert O. Gjerdingen, A Classic Turn of
Phrase: Music and the Psychology of Convention (Philadelphia, 1988), 55–9.
13 Eckelmeyer, ‘Two Complexes’, 16.
14 The example is a particella showing the principal melodic line, the bass line and an occa-

sional inner part of the arietta, Adagio, lied (here transposed from F minor to G minor) and aria.
(To facilitate comparison with the other three pieces, all analytical discussion of the lied in this
article will refer to this transposed version, speaking of it as if it were in G minor. The relevance
of the difference between the key of the lied and that of the other three pieces will be considered
in Section 6.) Figured bass numerals are added at times to clarify the harmonic content. The
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 333

As a result of taking harmonic factors into consideration, incorpo-

rating the arietta, and examining the relationship between the
opening of the final Allegro of the quintet, bars 39–42, and the coda
of the aria, bars 38–41, some of my alignments – i.e. my interpretations
of the correlations – differ from those of Eckelmeyer. In interpreting
the major connections, Example 1 proposes the existence of a
sequence of configurations which recurs largely intact in the four
pieces. This sequence consists of seven discrete units which I term
memes A–G: a detailed explanation of this terminology will be
provided in Section 3 below. These seven units are identified and
defined by virtue of their melodic and harmonic analogy in all four
pieces.
The following are the salient characteristics of the seven memes.

Meme A: The opening tonic area

This meme is characterized by a prolongation of tonic harmony by



means of the progression i 3– 3–iv6–4–i 3, the raised third of the tonic
chord preparing the central emphasis on subdominant harmony. In
the Adagio and aria this basic sequence is further intensified by follow-
ing the iv6–4 by a grating  7–  7–6–4–2 on G. Melodically, the meme is
defined by its primary focus on the scale-degree sequence 5̂–1̂–6̂–5̂, the
pitches varying to some extent in their register and disposition in each
copy of the meme.

Meme B: The subsequent dominant prolongation

Meme B is defined by the prolongation of dominant harmony by
means of an augmented-sixth chord (in the arietta and aria) or a
diminished-seventh chord ( vii7/V) (in the Adagio and lied).
Melodically, the meme gravitates around the scale-degree collection 1̂,
 4̂ and 5̂, the pitches again varying in their register and disposition
between pieces (1̂ is not present in the lied, except in the lower voice,
sounding against the  4̂ of the upper voice on the first quavers of bars
9 and 11).

————
music is disposed so as to align vertically major structural correspondences – the seven memes
A–G – between groups of bars. The symbols used in the particella are as follows. (1) The seven
memes are shown enclosed by solid vertical continuity of the barlines. Where the differences of
metre permit, these continuous lines also occur within the groups. (2) Continuous dotted
(bar)lines indicate functional parallelisms, e.g. bar 19 of the arietta and bar 8 of the aria, where
the passages linked by the dotted line are analogous contrasting themes in the relative major. (3)
Without continuity of barlines, even when barlines are aligned vertically, passages are not in direct
correspondence; such coincidences are necessary to allow alignment of the structural corre-
spondences. (4) Brackets and lines (apart from such lines as indicate the voice exchanges within
meme C) draw attention to specific pitch correspondences (as in meme G). Some secondary
correspondences, because of the overriding influence of more important similarities, are not
vertically aligned (e.g. the bass of bars 12–13 of the arietta and the bass of bars 3–4 of the aria).
334 STEVEN JAN

Example 1. Memetic replication in four Mozart pieces: (a) Der Schau-

spieldirektor, K.486 (1786), no. 1, ‘Da schlägt die Abschiedsstunde’, bars
1–42; (b) String Quintet in G minor, K.516 (dated 16 May 1787), IV,
bars 1–42; (c) Lied der Trennung ‘Die Engel Gottes weinen’, K.519
(dated 23 May 1787), bars 1–22; (d) Die Zauberflöte, K.620 (1791), no.
17, ‘Ach ich fühl’s’.
1 2

Arietta
(1786)

5
1 2 3 4

Adagio
(16.v.1787)

Lied
(23.v.1787)

Aria
(1791)

3 4 5

Arietta

6 7 8 9

Adagio

Lied

Aria
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 335

Example 1 continued
A
6 7 8 9 10

Arietta 6 5
4 3

11 12 13 14
10

6 / 7 6
4 6 3
Adagio 4
2

2 3 4 5
1

6 5
Lied 4 3

1 2 3

6 / 7 5
Aria 4 6 3
4
2

B
11 12 13 14

Arietta

15 16

Adagio

6 7 8 9

Lied

4 5

Aria
336 STEVEN JAN

Example 1 continued
15 16 17

Arietta

17

Adagio

10 11 12

Lied

6 7

Aria

18 19 20 21

Arietta

Adagio

Lied

8 9 10

Aria
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 337

Example 1 continued
22 23

Arietta

Adagio

Lied

11 12 13

Aria

24 25

Arietta

Adagio

Lied

14 15

Aria
338 STEVEN JAN

Example 1 continued
26 27 28 29

Arietta

Adagio

Lied

16 17 18

Aria

30 31

Arietta

Adagio

Lied

19 20 21 22

Aria
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 339

Example 1 continued
C
32 33 b.34

Arietta

18 19 20

Adagio

15
12 13 14

Lied

25 26 3
23 24

Aria

Arietta

22 23 24
21 25

Adagio

15

( )
Lied

26

Aria
340 STEVEN JAN

Example 1 continued
D

Arietta

26 [ ]
27 28 29

Adagio

16 17

Lied

27 28 29 30

Aria

34 35 36

Arietta

30 31

Adagio

Lied

31

Aria
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 341

Example 1 continued
E
37 38 39

Arietta

32

Adagio

18

Lied

32

Aria

40 41

Arietta

33 34

Adagio

Lied

33 34 35

Aria
342 STEVEN JAN

Example 1 continued
F

Arietta

35 36

Adagio

20
19

Lied

36 37

Aria

G
Allegro moderato

Arietta

37 38 Allegro

Adagio

22
21

Lied

38

Aria
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 343

Example 1 continued
42

Arietta

39 40

Adagio

Lied

39

Aria

Arietta

41 42

Adagio

Lied

40 41

Aria
344 STEVEN JAN

Meme C: 4̂/6̂– 7̂–1̂/voice-exchange patterns

Here an upper-voice sequence 4̂– 7̂–1̂ (in the arietta) or 6̂– 7̂–1̂ (in
the lied, Adagio and aria) resolves onto a 6–3 on B
. Thereafter, in the
arietta and lied, an upper-voice motion 4̂–3̂ and lower-voice 7̂–1̂ give
rise to a double interlocking voice exchange, as indicated by the inter-
secting diagonal lines in Example 1. The upper ( 7̂–1̂–4̂–3̂) and lower
(4̂–3̂–  7̂–1̂) lines of this structure are inverted in the aria, its initial
melodic gesture 6̂–  7̂ (bar 241) being similarly treated.

Meme D: Quasi-tetrachordal figures

Not present in the arietta, this meme is derived from a tetrachordal
pitch collection which, in its fullest form, in the Adagio and aria, pairs
a descending lower-voice sequence 7̂–  6̂–
6̂–5̂ with an ascending
upper-voice  3̂–  4̂–  4̂–5̂.15 The melodic element is articulated in
these two pieces by distinctive falling diminished sevenths,
2̂–  3̂ and

3̂–  4̂. The lied also presents a falling melodic diminished seventh

2̂–  3̂, but this is associated with only a small subset of the tetrachordal
collection, and effects the conversion of the E
major 5–3 harmony of
bar 161 to the C minor 6–3 of bar 171, an implied 5–6 motion above
bass E
. Example 1 shows that the aria synthesizes the tetrachordal
structure of the Adagio with the V–VI interrupted progression and 5–6
motion over E
of the lied.

Meme E: End-dominant preparation and elaboration

As with meme B, meme E is characterized by the prolongation of
dominant harmony. Here, again, the melodic scale-degree collection
1̂,  4̂ and 5̂ of meme B is in evidence (especially in the arietta and aria),
but the upper neighbour-note lower-voice motion of the earlier meme
(in the arietta and aria) is more pronounced and is combined with the
diatonic and chromatic lower neighbour notes. As in the prolongation
of the dominant in meme B, the augmented-sixth chord, associated
with the upper neighbour note, and the diminished-seventh chord
( vii7/V; but V6–5/V in the Adagio), associated with the chromatic
lower neighbour, are prominent. I interpret bar 18 of the lied as articu-
lating meme E and it is here that my reading diverges most substan-
tially from that of Eckelmeyer. Her Example 116 aligns bar 18 of the

15 In some works of Mozart’s – such as bars 21–5 of the second movement of K.516 – the

rising upper-voice structure
3̂– 3̂–  4̂–  4̂–5̂ is associated with the descending chromatic tetra-
chord 1̂– 7̂–  6̂–
6̂–5̂ in the lower voice. It is this association which prompts my reading of the
structure in the Adagio and aria as ‘quasi-tetrachordal’, for these passages are perhaps best under-
stood as selections from a total resource. For a fuller discussion of these figures, see Steven Jan,
Aspects of Mozart’s Music in G Minor: Toward the Identification of Common Structural and Compositional
Characteristics (London and New York, 1995), Chapter 7. See also Peter Williams, The Chromatic
Fourth during Four Centuries of Music (Oxford, 1998), Chapter 6; the tetrachordal figures in the
Adagio and aria are, however, not examined by Williams.
16 Eckelmeyer, ‘Two Complexes’, 15.
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 345

lied with bars 36–8 and bars 37–8 respectively of the Adagio and aria
(these passages being designated meme F below). This is certainly
tenable, for there is a shared melodic contour; but if the piano coda
of the lied’s strophe (bars 19–22) and not its bar 18 is held to contain
meme F, then the harmonic content of bar 18, essentially a dominant
chord preceded by its lower neighbour note, supports a reading of this
bar as meme E. Example 1 shows that the arietta and aria present
meme E in a very similar manner.

Meme F: Closing melodic descents

This meme, absent from the arietta, is characterized by a melodic
descent by step
/ 6̂ . . .  7̂ over dominant harmony (resolving subse-
quently to 1̂/i in the lied), preceded, in the Adagio and aria, by the
arpeggiations  7̂–2̂–4̂ and 2̂–4̂–/
6̂. It will be seen from Example 1
that a substantial part of the total melodic resource of this meme, the
segment 2̂–4̂–6̂ . . .  7̂, also appears earlier in the aria, in its presen-
tation of meme C at bars 23–4.

Meme G: Continuations
Not present in the arietta and lied, this meme – the basis of the theme
of the 6/8 G major Allegro which forms the main part of the quintet’s
final movement, and of the instrumental coda to the aria – is
characterized by a scalic descent 5̂–  7̂(–1̂). Harmonically, aside
from the common tonic opening and analogous concluding


ii6–3/
II 6–3–I6–4/i6–4–V–(i) cadences, there is an equivalent emphasis
on the subdominant in bar 405 of the quintet’s finale and bar 393 of
the aria. The change of mode and mood at this point of the quintet is
akin to the procedure of the arietta (bars 42ff.), although I do not read
any memetic connections between these passages.

3. AN OVERVIEW OF THE MEMETIC PARADIGM

These intertextual similarities are clearly striking, and call for a

conceptual framework within which they may be understood. As
suggested at the opening, one way of comprehending such analogies
is in terms of the meme concept. A meme, in Dawkins’s celebrated
definition from The Selfish Gene, is
a unit of cultural transmission, or a unit of imitation . . . Examples of memes
are tunes, ideas, catch-phrases, clothes fashions, ways of making pots or of
building arches. Just as genes propagate themselves in the gene pool by
leaping from body to body via sperms or eggs, so memes propagate them-
selves in the meme pool by leaping from brain to brain via a process which,
in the broad sense, can be called imitation.17

17 Richard Dawkins, The Selfish Gene (2nd edn, Oxford, 1989), 192.
346 STEVEN JAN

While constraints of space prevent a full discussion of the nature and

ramifications of the memetic paradigm,18 it is necessary to outline the
following central points in order to contextualize the memetic
interpretation of the Mozart passages which follows in Sections 4–6
below.
Memetics – which some regard as a ‘long-awaited scientific theory
unifying biology, psychology, and cognitive science’19 – is predicated
on the notion that the meme, the unit of cultural information – in the
verbal, visual and, as I assert here, musical realms – may be compared
to the unit of biological information, the gene, in that it is dissemi-
nated by replication between the members of a cultural community,
giving rise to a number of copies or ‘coequals’.20 Moreover, the gene
and the meme are themselves members of a larger class of entities
which Dawkins terms ‘replicators’ – that is, ‘any entity in the universe
of which copies are made’.21 This principle is the foundation of the
concept of ‘universal Darwinism’, whereby
in addition to biological evolution as it is normally understood, Darwinian
evolution is also operating to produce the transformations in time that we
see in certain other spheres, such as immune system function and even the
way science itself [as a component of culture generally] operates.22

18 For book-length treatments offering an overview of memetics, see Susan Blackmore, The

Meme Machine (Oxford, 1999); Richard Brodie, Virus of the Mind: The New Science of the Meme
(Seattle, 1996); and Aaron Lynch, Thought Contagion: How Belief Spreads through Society – The New
Science of Memes (New York, 1996). For more critical voices, some of the essays in Darwinizing
Culture: The Status of Memetics as a Science, ed. Robert Aunger (Oxford, 2000), come from disci-
plinary perspectives (such as anthropology and sociology) which are somewhat sceptical of the
meme concept. A more technical, mathematically orientated account of memetic transmission
and evolution is given in Aaron Lynch, ‘Units, Events and Dynamics in Memetic Evolution’,
Journal of Memetics – Evolutionary Models of Information Transmission, 2/1 (1998) (<http://
jom-emit.cfpm.org/1998/vol2/lynch_a.html>). For more on the application of memetics to
music, see Steven Jan, ‘Replicating Sonorities: Towards a Memetics of Music’, Journal of
Memetics – Evolutionary Models of Information Transmission, 4/1 (2000) (<http://
jom-emit.cfpm.org/2000/vol4/jan_s.html>); ‘The Memetics of Music and its Implications for
Psychology’, Proceedings of the Sixth International Conference on Music Perception and Cognition, ed.
Christopher Woods, Geoffrey B. Luck, Renaud Brochard, Susan A. O’Neill and John A. Sloboda
(CD-ROM, Keele, 2000); ‘The Selfish Meme: Particularity, Replication, and Evolution in Musical
Style’, International Journal of Musicology, 8 (2002), 9–76; and ‘The Illusory Mozart: Selfish Memes
in the Priests’ Marches from Idomeneo and Die Zauberflöte’, International Journal of Musicology, 10
(forthcoming, 2004). A few passages in the present article are adapted from Jan, ‘Replicating
Sonorities’, ‘The Memetics of Music’ and ‘The Selfish Meme’. The recently established online
Journal of Memetics – Evolutionary Models of Information Transmission (<http://
jom-emit.cfpm.org>) offers a forum for articles on various aspects of the subject.
19 Brodie, Virus of the Mind, 13.
20 That is, forms of analogous configuration whose similarity is such that a relationship of

copying – as opposed to random convergence – may, on the basis of statistical probability, be

inferred to exist between them. When testing candidate strings of data (in a variety of media,
including music) for memetic content, the presence of such corresponding segments in two
contexts allows the initial and terminal elements and medial content of the meme to be defined
by reference to that portion of the information stream which is replicated (i.e. the coequal). In
this way, a ‘quasi-digital’ order arises from what is otherwise – from a memetic perspective – undif-
ferentiated, ‘analogue’, data.
21 Richard Dawkins, The Extended Phenotype: The Long Reach of the Gene (London, 1982; repr.

Oxford, 1983), 293.

22 Henry C. Plotkin, Darwin Machines and the Nature of Knowledge: Concerning Adaptations, Instinct

and the Evolution of Intelligence (London, 1995), xvii.

THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 347

Such a particle as the phrase ‘To be, or not to be, that is the question’
(Shakespeare, Hamlet, 3, i, 56) survives intact by replication within
Western culture because it has the property of cultural salience – in
Brodie’s terms, it has the effect of ‘pushing our [psychological-
emotional] buttons’ because of its arresting existential imagery23 –
which promotes its replication. The attributes of ‘longevity’ (the dura-
bility of any single copy of a replicator), ‘fecundity’ (the propensity of
a replicator to engender its duplication), and ‘copying-fidelity’ (the
accuracy of the replication process)24 ultimately decide the fate of all
replicators, be they genes, patterns of pitches and rhythms, or ‘a sheet
of paper that is xeroxed’.25
Just as with the gene, however, there is no hard-and-fast definition of
what constitutes a meme. Indeed, the biologist Seymour Benzer
proposed refining the somewhat loose concept of the gene by focusing
it into the categories of ‘muton’ (‘the minimum unit of mutational
change’), ‘recon’ (‘the minimum unit of recombination’) and ‘cistron’
(‘the unit responsible for synthesizing one polypeptide chain’).26 In
Dawkins’s view, however, the gene is
any portion of chromosomal material that potentially lasts for enough
generations to serve as a unit of natural selection . . . a genetic unit that is
small enough to last for a large number of generations and to be distrib-
uted around in the form of many copies . . . a unit which, to a high degree,
approaches the ideal of indivisible particulateness.27

This definition – Dawkins speaks of the ‘optimon’, ‘the unit of natural

selection’28 – is perhaps the most utilitarian. Moreover, it is readily
applicable to the meme. That which is replicated and survives over
time attains memetic status. In music, this essentially restricts the
meme to small collections of pitches and rhythms, generally of no
more than a few bars’ length – as is the case in the seven memes shown
in Example 1.29
Much writing on memetics imports the biological distinction
between the ‘genotype’ (the total genetic complement of an indi-
vidual) and the ‘phenotype’ (the expression of that complement in the
form of the individual’s morphology and its patterns of behaviour).
Memetics regards the fundamental form of the meme to be patterns
of neuronal interconnection in the brain, an individual’s total memetic
complement being known as the ‘memotype’. The physical manifes-
tations of these memes, the ‘phemotype’, are behavioural patterns and
the resultant artefacts they engender.30 In the category of behavioural

23 Brodie, Virus of the Mind, 91.

24 Dawkins, The Selfish Gene, 18, 194.
25 Dawkins, The Extended Phenotype, 83.
26 Ibid., 81.
27 Dawkins, The Selfish Gene, 28, 32, 33.
28 Dawkins, The Extended Phenotype, 81.
29 See Jan, ‘Replicating Sonorities’, Section 5, for more detailed discussion of the segmenta-

tion of the musical continuum into memes by means of the criteria of particulateness and co-
equality.
30 Blackmore, The Meme Machine, 63.
348 STEVEN JAN

patterns, one would include the activities of composition and perform-

ance; in the category of artefacts, one would include notated musical
scores and the patterns of sound waves for which they code in perform-
ance. While a full discussion of this matter is not possible in the present
context, it is clear that the memotype–phemotype distinction has
profound philosophical implications for the ontology of music.
The memotype–phemotype dualism also underpins what might be
regarded as the epidemiological aspect of memetics. Despite being
strictly part of the body – it is, after all, a collection of nerve cells – in
its memotypic form a meme residing in a brain is analogous to an infec-
tive agent, for it has been acquired, like an infection, by contact with
an infected individual, or rather by contact with the phemotypic
products of an infected individual. It therefore exists in a sometimes
mutualist, sometimes commensal, but often parasitic relationship with
its human host.31 In its phemotypic form, the meme is transmitted to
other hosts, often spreading to infect whole swathes of geographical
and conceptual space; the epidemiological and evolutionary accounts
of language transmission and mutation given by Darwin32 and
expanded upon by Cavalli-Sforza and Feldman33 describe only one
element of this central cultural phenomenon.
The memetic replication underpinning human culture occurs at
several hierarchic levels. Meyer’s formulation of this hierarchy – not,
however, in memetic terms – proposes the levels of ‘laws’ (which are
‘transcultural constraints . . . the principles governing the perception
and cognition of musical patterns’); ‘rules’ (which are ‘intracultural,
not universal’, and which ‘constitute the highest, most encompassing
level of stylistic constraints’); ‘dialect’ (which represents ‘substyles that
are differentiated because a number of composers . . . employ
(choose) the same or similar rules and strategies’); ‘idiom’ (which
refers to constraints ‘a composer repeatedly selects from the larger
repertory of the dialect’); and ‘intraopus style’ (which is ‘concerned
with what is replicated within a single work’).34 As sets, each of which
is included by the set at the hierarchic level above it, one might
imagine these stratified relationships in the form of a Venn diagram
or, as in Nattiez, an inverted pyramid.35 The level of laws represents

31 See Juan D. Delius, ‘Of Mind Memes and Brain Bugs: A Natural History of Culture’, The

Nature of Culture: Proceedings of the International and Interdisciplinary Symposium, October 7–11, 1986
in Bochum, ed. Walter A. Koch (Bochum, 1989), 26–79. Examples of such harmful memes, by
which is meant those which in some way impede the replication of the host’s genes, include, at
their most extreme, suicide cults (Blackmore, The Meme Machine, 47, 51) and, more subtly, memes
for contraception (Lynch, Thought Contagion, 91–2). The complex subject of gene-culture (meme)
‘coevolution’ is examined in William H. Durham, Coevolution: Genes, Culture, and Human Diversity
(Stanford, 1991).
32 Charles Darwin, On the Origin of Species by Means of Natural Selection, or the Preservation of

Favoured Races in the Struggle for Life (2nd edn, London, 1860), ed. Gillian Beer (Oxford, 1996),
342.
33 Luigi L. Cavalli-Sforza and Marcus W. Feldman, Cultural Transmission and Evolution: A Quan-

titative Approach, Monographs in Population Biology, 16 (Princeton, 1981), 19–29.

34 Meyer, Style and Music, 13, 17, 23–4.
35 Jean-Jacques Nattiez, Music and Discourse: Toward a Semiology of Music, trans. Carolyn Abbate

(Princeton, 1990), 136, Figure 6.2.

THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 349

our species’ ‘hard-wired’ perceptual and cognitive attributes, which are

genetically determined and transmitted and which act ultimately as a
‘filter’ for memes, dictating which patterns can be perceived,
processed and stored in memory. The lower levels represent our
species’ ‘soft-wired’ memory profiles, which are memetically deter-
mined and transmitted.
A single-unit meme is just one element in the myriad web of repli-
cated patterns constituting a cultural community. As such it exists in
varying degrees of ‘coadaptation’ with other memes. At times, memes
are intimately connected with others such that, while they are occasion-
ally replicated individually, they are also regularly, perhaps even
preferentially, copied as part of ‘coadapted meme complexes’, or
‘memeplexes’, a notion discussed more fully in Section 4 below. As will
be seen in Section 6 below, coadaptation may at times link musical
memes with memes in the verbal-conceptual realm.
The copying of memes in cultural systems is considerably less
accurate than the copying of genes in biological systems, which is
almost always completely faithful. Many such deviations from exact
imitation – mutations – do not survive to be replicated: they may, for
instance, be overwhelmed by the collective force of the parent form in
the ‘meme pool’ (which can be taken to be any level of the cultural
hierarchy higher than the idiom; for convenience, the dialect is gener-
ally the most appropriate level). Others, however, given the often more
distinctive cultural profile of mutants compared with their antecedent
forms, eventually come to overwhelm their precursors by spreading,
epidemiologically, through the meme pool. Over time, the complex-
ion of the meme pool gradually changes as a result of this incursion of
mutant memes with higher fecundity than their antecedents. This
differential survival of mutant forms is the engine – Darwin’s natural
selection, or, as Cavalli-Sforza and Feldman adapt it, ‘cultural selec-
tion’36 – which drives cultural-stylistic evolution. Dennett concisely
summarizes this process as follows:
(1) variation: there is a continuing abundance of different elements[;] (2)
heredity or replication: the elements have the capacity to create copies or
replicas of themselves[;] (3) differential ‘fitness’: the number of copies of
an element that are created in a given time varies, depending on interac-
tions between the features of that element and features of the environment
in which it persists.37

All three conditions are satisfied by the meme, from its tendency to
miscopying (variation), its capacity for replication within the members
of a cultural community (heredity), and the contrasting fecundity of
different memes (differential fitness). On this reasoning, then, the
transmission and reception of musical culture is, in essence, the
outcome of a process of relentless competition between a near infinity

36 Cavalli-Sforza and Feldman, Cultural Transmission and Evolution, 15.

37 Daniel C. Dennett, Darwin’s Dangerous Idea: Evolution and the Meanings of Life (London, 1995),
343.
350 STEVEN JAN

of replicators, each vying for that most precious, sustaining resource:

memory capacity.38 Section 5 below returns to the issue of memetic
evolution in more detail by examining the process as it applies to
memes A–G in the Mozart pieces.
Central to the striking conception embodied in Dawkins’s phrase
‘selfish gene’ is the notion of the gene’s appearing to act egotistically
in furtherance of its own interest. To paraphrase him, an entity, such
as a baboon or a gene, is said to be selfish if it has the effect (not the
purpose) of promoting its own welfare at the expense of that of
another entity.39 The most significant competitors of the gene in this
sense are its ‘alleles’, its close equivalents or ‘alternative forms’, which
act as rivals for the slots (loci) on a chromosome.40 Devoid of conscious-
ness, the gene cannot be selfish, selfless, or in possession of any other
attribute of higher organisms, yet the anthropomorphic metaphor of
conscious intentionality effectively articulates the truism that those
genes which tend to act – i.e. engender phenotypic effects – in such a
way as to increase the chances of their survival – via that of their host
organism – will, self-evidently, prevail and be replicated, to the
disadvantage of those less phenotypically consequential genes.
Powerful as the selfish gene metaphor is, it is founded upon a
(tauto)logical and mechanistic process entirely devoid of conscious
agency or intentionality. It is in the nature of such processes that they
often appear to be powerfully goal-directed and intentional.
Applying this logic to the realm of culture, a meme, to reiterate
Dawkins’s formulation, can be said to be selfish if it has the effect
(again, not the purpose) of promoting its own welfare – that is, its
survival by replication – at the expense of that of another meme. Again,
its most significant competitors are its memetic alleles – in this case, its
mutational antecedents or consequents, or its configurational homo-
logues. As is noted above, a mutant meme may possess greater cultural
salience than its original form and may thereby have an increased
propensity to imitation. Within the meme pool, such a meme may be
selectively favoured – it may be differentially imitated in cultural
contexts – and may therefore increase its statistical representation,
being subject to a process of neo-Darwinian natural selection exactly
equivalent to that operating upon the gene. As with the phenotypic
manifestation of the gene, the phemotypic incarnation of the meme
has a direct bearing upon its tendency to imitation and therefore ulti-
mately on its representation in the meme pool. It is in this sense that
we may comprehend the reflexive ‘themselves’ in Dawkins’s statement
cited earlier that ‘memes propagate themselves in the meme pool by
leaping from brain to brain via a process which . . . can be called
imitation’;41 and it is in the same sense that all references in this article

38 See Jan, ‘The Memetics of Music’ for an account, couched in terms of Narmour’s impli-

cation-realization model, of how mutational changes to a meme may increase its perceptual and
cognitive salience and therefore its propensity to imitation.
39 Dawkins, The Extended Phenotype, 284.
40 Ibid., 283.
41 Dawkins, The Selfish Gene, 192.
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 351

to the apparent conscious intentionality of memes should be under-

stood.
To conclude this overview of the memetic paradigm, it seems that
the most challenging implication of the ‘meme meme’ is that it
proposes that our mental processes are largely meme-driven. As
suggested in the discussion of stylistic hierarchies above, beyond the
innate, ‘hard-wired’ attributes of our neural hardware, which regulate
automatic and instinctive functions, our higher, most distinctively
human, capacities are dependent upon a ‘soft-wired’ (memetic)
component. Such knowledge and habits are acquired by exposure to
the information other members of our cultural community propagate
– both ‘vertically’, as parents, and ‘horizontally’, as peers. One contro-
versial development of this line of argument, from Dennett, is that
Human consciousness is itself a huge complex of memes (or more exactly,
meme-effects in brains) that can best be understood as the operation of a
[serial] virtual machine implemented in the parallel architecture of a brain that
was not designed for any such activities. The powers of this virtual machine
vastly enhance the underlying powers of the organic hardware on which it
runs, but at the same time many of its most curious features, and especially
its limitations, can be explained as the byproducts of the kludges [i.e. ad hoc
software alterations] that make possible this curious but effective reuse of
an existing organ for novel purposes.42

Of course, such an interpretation fatally corrodes the notion of the

‘great composer’, whose conscious intentionality – so the work-concept
memeplex induces us to believe – gives rise to the monistic, organic
work of art. The memetic paradigm leads us inexorably to the melan-
choly conclusion that the composer is merely a medium for selfish
units of information which use his or her brain as a temporary repos-
itory, and then as the means to engender the phemotypic products
which will facilitate their transmission to other brains. I shall return
briefly to this point in Section 7.

4. MUSICAL MEMES AND MEMEPLEXES: PARAMETRIC REPLICATION AND MEMETIC

HIERARCHIES

From this synopsis of the memetic paradigm, and from the connec-
tions between the four pieces shown in Example 1 and discussed in
Section 2 above, it is reasonable to suggest that each of the patterns
A–G is indeed a meme, because each is a discrete and particulate entity
replicated in at least two of the four pieces – and in the case of four of
the memes, replicated in all four pieces. From their initial union in the
arietta, memes A, B, C and E are replicated in the Adagio, lied and aria
with three additional memes – D, F and G – being added in the Adagio
and subsequently replicated (aside from the absence of meme G in the
lied) in the later pieces.

42 Daniel C. Dennett, Consciousness Explained (London, 1993), 210.

352 STEVEN JAN

Moreover, because the pieces – certainly the Adagio and the aria –
contain essentially the same complement and sequence of memes, the
succession in each piece might be regarded as a memeplex43 consist-
ing of the seven memes A–G. As Dawkins notes,
Memes, like genes, are selected against the background of other memes in
the meme pool. The result is that gangs of mutually compatible memes –
coadapted meme complexes or memeplexes – are found cohabiting in indi-
vidual brains. This is not because selection has chosen them as a group, but
because each separate member of the group tends to be favoured when its
environment happens to be dominated by the others.44

This sequence of memes, as the schematic representation in Figure

1(a) below illustrates,45 may itself be regarded as a superordinate unit
meme, existing at what might be termed a memetic ‘level 1’, and repli-
cated essentially intact, although in distributed form, in the four
different contexts. Conversely, it will be evident that each of the memes
A–G, located at what might be regarded as a memetic ‘level 2’, is itself
a memeplex, consisting of discrete, independent memes in such
parameters as melody, harmony and indeed rhythm, thereby creating
a memetic ‘level 3’.46 As with the memes at levels 1 and 2, these
uniparametric level-3 units may also be considered to be memes, for
they too are replicated in the context of the four pieces.47
Of course, the level-3 memes are part of the common memetic
currency of late eighteenth-century style; and, as is explored in Section
5 below, the level-2 memes derive – spontaneous generation aside –
ultimately from works of Mozart’s predecessors and/or contempo-
raries, and go on to be replicated in works of Mozart’s contemporaries

43 For a memeplex to occur, it seems reasonable to suggest the following two necessary

conditions. First, each of the particles constituting the memeplex must be replicated individually
in at least one other context, in order for each particle to attain independent memetic status.
Secondly, the collection of memes, the memeplex, must be replicated collectively in at least one
other context, in order for this higher-level grouping itself to attain memetic status.
44 In Blackmore, The Meme Machine, xiv.
45 It will be understood that the two types of hierarchy represented here are structural, as

opposed to Meyer’s cultural hierarchies (the concepts of laws, rules, etc.) discussed in Section 3.
The arrangement of this figure into ‘Context 1’ and ‘Context N’ represents the notion of repli-
cation over time.
46 That memes also exist in the rhythmic dimension, sometimes independent of pitch, can

be seen by comparing bar 6 of Example 1(a) with bar 10 of Example 1(b), where the pattern
is replicated. Indeed, a memetic reading of Maury Yeston, The Stratification of Musical
Rhythm (New Haven, 1976), suggests the existence of rhythmic memes at hierarchic levels above
the immediate foreground of local attack points. For present purposes, however, I will focus
largely on memes in the parameter of pitch.
47 Narmour might regard the level-3 memes as ‘style forms’. These, he notes, are ‘those para-

metric entities which achieve enough closure so we can understand their functional coherence
without reference to the specific intraopus context from which they come – all those seemingly
time-independent patterns, large and small, from parameter to parameter, which recur with statisti-
cally significant frequency’ (Eugene Narmour, Beyond Schenkerism: The Need for Alternatives in Music
Analysis, Chicago, 1977, 173). Narmour might consider the level-2 memes to be ‘style structures’.
He notes that ‘the [statistically common] contexts which result from [the syntactic] arrangement
[of style forms] can be called style structures in the sense that they are directly tied to and contribute
to the structure of real pieces, not just to constructed classes of things, as are style forms. Unlike
the description of style forms, the identification of style structures involves ascribing time-
dependent function to patterns . . . in intraopus relationships’ (ibid., 174).

 

 
 

 
 
Level 1    

Level 2              

Level 3

 

 
 
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART

 
 

∧ ∧ ∧ ∧ ∧ ∧
Background " !  " ! 

Middleground ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧
" # $ # ! " !  " # $ # ! " ! 

Foreground ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧ ∧
!  $ " # $ "  !  $ " # $ " 
353

Figure 1. Hierarchical memetic replication and the generation of memeplexes.

354 STEVEN JAN

Example 2. A network representation of meme A.

x

6
Arietta

10

Adagio

1
Lied

1
Aria

and/or successors (replication at the level of dialect). The level-2

memes might also conceivably be found replicated individually, or as
a memeplex, in other works by Mozart (replication at the level of
idiom).
As memeplexes of replicants in different musical parameters, the
level-2 memes may be analysed using a network representation. In
Example 2, level-2 meme A (a style structure) is shown in terms of the
melodic, harmonic and rhythmic level-3 memes (the style forms)
which constitute it in its four copies. Such a representation has
relevance to understanding the perception and cognition of memes –
and therefore their transmission and evolution – for Gjerdingen sees
the style structure as a category of cognitive ‘schema’, its constituent
style forms being akin to generative lower-level ‘features’.48 In terms of
the perception and cognition of meme A, the fundamental element –
the pattern common to all four copies of the meme – is the shallow
middleground-level pitch structure marked ‘x’ in Example 2 and
linked to the largest, all-encompassing box. This is its schematic basis,
the attribute which defines the identity of the meme; the features are
those elements which generate this schema. None of these features is
common to all four pieces, but all features are found in at least two
and sometimes in three of the pieces.
In perception – by the listener or the composer-as-listener – bottom-
up perceptual processes at the opening of meme A in one of the four
pieces will first identify some of its component features, but without at
48 Gjerdingen, A Classic Turn of Phrase, 45–6.
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 355

Example 3. Schematic pitch structure of level-2 memes A–G.

Meme A B C

( )

Meme D E F

( )

Meme G

( )
)
(
( )

this stage necessarily comprehending their broader context. At some

point, the cumulative evidence of the features will elicit the selection,
on best-fit criteria, of the middleground schema, which then regulates,
by top-down perceptual processes, the search for the remaining
features. The meme is deemed closed – instantiated – when all those
expected features generating the schema are registered.
In (re-)generation – by the composer – a middleground-level
schematic meme is engendered by the bottom-up tendency of a set of
foreground-level feature memes to conglomerate in a particular
sequence, to form a memeplex. From a top-down perspective, however,
the middleground meme has a number of loci along its length, anal-
ogous to those on a chromosome, for which allelic feature memes
conforming to the parameters of the class of features for that locus
compete. The first position of the schema underpinning meme A, for
instance, may only be filled by memes expressing tonic harmony and
the scale degree 5̂ (see Example 2, bar 1 and box x). I will return to the
top-down, bottom-up dichotomy in Section 5 below, in connection with
a discussion of the evolution of Mozart’s memeplex.
In terms of pitch configurations, the memetic equivalence between
the particles at level 2 of Figure 1(a) inheres in that which is invariant,
in the underlying schematic (shallow-middleground) configuration of
the memes, for at the foreground level most of the memes in any one
of the pieces are clearly not exact copies of those in the other pieces,
owing to the differences between memes at level 3. Example 3, after
Example 1 and box x of Example 2, isolates the common schematic
structures underpinning the memes at level 2 by reducing them to
their basic voice-leading configurations. In this way, the rhythmic
component of the level-2 memes, the independent level-3 rhythmic
memes, is eliminated, as are surface pitch diminutions. In cases where
there is variability between pieces in an element of the meme, the
alternative components are enclosed in boxes (whereby the whole
content of one box can be substituted for the whole content of another
box); less substantial variants are enclosed in brackets.
356 STEVEN JAN

It should be noted that Figure 1(a) represents replication hierar-

chies, not the voice-leading hierarchies which are generated in
Schenkerian analysis by transformations deriving from the precepts of
Fuxian counterpoint. The three levels of Figure 1(a) are only loosely
analogous to Schenkerian foreground, middleground and background
strata because – as a general principle and also in this particular
context49 – some pitches of the superordinate level-1 meme may not
necessarily exist at a deep structural level according to Schenkerian
voice-leading considerations. Conversely, it may be the case – again as
a general principle and also in this particular context – that some
pitches of the level-3 memes are of background status in a Schenker-
ian reading. Nevertheless, as Figure 1(b) shows, the model sketched in
Figure 1(a) may be adapted to represent, in hypothetical terms,
memetic propagation at the three principal Schenkerian structural
levels, and the generation of memeplexes at foreground and middle-
ground levels.
In Figure 1(b) each shaded box represents a meme; 27 such units
are shown at the foreground level. (For clarity, multiples of three,
generating a 3̂–2̂–1̂ Ursatz, are used in this illustration; and only one
middleground level is represented.) The careted numeral in each fore-
ground-level box represents the pitch upon which each meme is based;
at this level, if the meme is lightly reduced, this pitch will be seen to
be its principal element, the product of the various diminutions
making up the extreme foreground level of the meme. As represented
by the arrows leading up from the foreground level, pitches from these
memes may exist – by virtue of their voice-leading status – at the next
higher level, generating three three-component middleground
memes. Similarly, arrows leading up from the middleground level
represent three pitches from these middleground memes existing at
the next higher level, generating a 3̂–2̂–1̂ Ursatz meme at the back-
ground level.
Given that a collection or complex of foreground-level memes which
generates a given middleground-level structure might theoretically
occur in another context – as the pieces under discussion demonstrate
– then while existing as a middleground-level meme (it is memetic at
this level because the same middleground-level structure occurs in
these two contexts, and possibly others), it also exists as a memeplex at
the foreground level (see again note 43 for a generic definition of a
memeplex). Each individual foreground pattern is a meme (it exists in
this form in these two contexts, and may exist independently in other
contexts), but the complex is also memetic.
Furthermore – applying the same principle recursively to higher
levels – if it is accepted that a complex of middleground-level memes
which generates a given background-level structure might theoretically
occur in another context, then while existing as a background-level

49 Constraints of space prevent a detailed examination of the four pieces in Schenkerian terms.
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 357

meme (it is memetic at this level because the same background-level

structure occurs in these two hypothetical contexts and – according to
the precepts of Schenkerian theory – certainly others), it also exists as
a memeplex at the middleground level. Each individual middleground
pattern is a meme (it exists in this form in these two contexts, and may
exist independently in other contexts), but the complex is also
memetic.
Clearly ‘real’ units at the foreground level generate ‘virtual’
configurations at higher structural levels. It may be the case that the
same pattern is replicated at more than one level in the same work; or,
in different works, the same pattern may be propagated not at their
foregrounds, but at higher levels. In this second case, on a strict defi-
nition, these are not units of direct imitation; but they are units of
consequential replication and should still therefore be regarded as
memetic.50 The structure of such higher-level memes is potentially
instructive for what it can tell us about the ‘conglomerative grammar’
of foreground-orientated memes – how such memes stick together
during the compositional process. This is ultimately a function of their
initial and terminal nodes, their pitch content, and the way these
elements interact with our innate perceptual and cognitive attributes.
It will therefore be understood from these observations that there
are two ways in which memes may participate in structural hierarchies.
First, they are hierarchic in the sense represented by Figure 1(a),
whereby uniparametric memes (at level 3) conglomerate to produce
multiparametric memes (at level 2) and ultimately engender a global
unit meme (at level 1) encompassing each of the pieces. Secondly, they
are hierarchic in the sense represented by Figure 1(b), whereby a given
pitch meme or pitch memeplex at the foreground level may be recur-
sively propagated at higher structural levels, retaining its virtual
identity while disappearing from real perception.51

5. ANCESTORS AND DESCENDANTS: THE MEMEPLEX IN EVOLUTIONARY CONTEXT

Beginning at the level of idiom, and examining the four pieces in

chronological sequence, it is evident that, while the level-2 memes A–G
essentially retain their middleground schematic identity over time
(Example 3 above), they also undergo foreground evolutionary

50 Psychological studies have shown that the perceptual and cognitive reality of such virtual

structures, at least to most listeners untrained in formal/structural analysis, is often fragile. Long
concerned with this dichotomy, Cook speaks of ‘discrepancies between the way in which theo-
reticians and analysts think of compositions . . . and the way in which listeners respond to them’
(Nicholas Cook, Music, Imagination, and Culture, Oxford, 1990, 57). As a further complication, it
will be evident that the conceptual and graphical constructs used to comprehend musical pattern-
ing and structure are themselves memeplexes, being propagated within music-theoretical
communities and subject to variation and selection, especially according to the criterion of
perceived fit with that which they purport to explain. As such, they may not necessarily be suited
clearly to resolve (or even detect) elements of the music the composer situates below the immedi-
ately perceptible surface.
51 See Jan, ‘Replicating Sonorities’, Section 6, for more detailed discussion of the structural-

hierarchic organization of memes and memeplexes.

 358

TABLE 1
PITCH EVOLUTION IN LEVEL-2 MEMES
Adagio lied aria a

meme A Adds a dissonant 7– 7–6–4–2 on G in bar 131. Reverts to the arietta’s treatment of the 6̂–5̂ melodic Restores the dissonant  7–  7–6–4–2 on G of the Adagio in bar 26.
motion (arietta bars 8–9; lied bar 4).

meme B Intensifies the dominant prolongation of bars Retains the  vii7/V prolongation and its associated  4̂. Mutates the  vii7/V of the Adagio and lied into an augmented-sixth

15–17 by means of the  vii7/V of bar 16 and its Replaces the sequence V– vii7/V–V of the Adagio chord ( iv6–3, bar 56; ii 6–4–3, bar 66), perhaps derived from the
associated  4̂. (bars 15–17) with V– vii7/V–V– vii7/V–V (bars 8–12). arietta, bar 162–3. Retains the V–X–V–X–V pattern of the lied,
where X is  vii7/V or an augmented-sixth chord.

meme C Mutates the melodic 4̂–  7̂ motion of bar 32 of Retains the 6̂–  7̂ motion of the Adagio in bar 13, and Mutates the rising motion 6̂–  7̂ into a falling pattern (bar 241).
the arietta to 6̂–  7̂ in bar 18. Eschews the double restores (bars 13–14) the double voice exchange of Inverts the double voice exchanges of the arietta and lied to
voice exchange f "/c'–g"/b
–c"/f –b
'/g of the the arietta. c"/f –b
'/g–f '/c–g'/B
(bars 244–253).
arietta (bars 32–4).

meme D Simplifies (bars 16–17) the descent
2̂ . . . 3̂ (bar 26) Retains (bars 264–273) the interruption V–VI of the lied (bars
of the Adagio. Reverts to the diatonic iv in bar 17. 152–161), and its unfilled melodic descent
2̂–  3̂. Combines this
(bars 27–303) with the harmonic plan of the Adagio (bars 26–9).

meme E Eliminates the cadence pattern of bar 39 of the The descending motion from  7̂ of the arietta (bars The pattern of bar 39 of the arietta is mutated and intensified by
arietta; but the emphases on V prolonged by 39–401) is simplified and taken further, to  7̂ (bar
2̂ in bar 32. Bars 33–5 intensify the prolongation of V of bars 40–1
harmonies containing  4̂ and 6̂ of the arietta 182), the bass figure of the arietta being retained. The of the arietta, introducing the chromatic  6̂ (bar 346).
(bars 40–1) are recreated in bars 33–5. prolongation of V of the arietta and Adagio is
eschewed.

meme F The descent  6̂ . . .  7̂ of the Adagio (bars 37–8) is Bars 362–38 of the Adagio are recreated in the tonic minor in bar
recreated in bars 20–2, but retains the tonic minor 374–6.
and cadences on 1̂.
meme G Recreates, in bars 384–406, the contour of bars 384–423 of the
Adagio/finale, retaining the tonic minor and adding chromatic
emphases on iv in bar 393 and the melodic
2̂ in bar 404.

a Looking vertically at some of the memes represented in Example 1 (or horizontally in Table 1), it might be said that, at times, the aria represents an evolutionary synthesis or amalgamation of features found in the other

pieces. For instance, in meme A the aria adopts the pitch sequence d'''–g" of bars 10–11 of the Adagio but in the register d"–g', first opened up in bars 1–2 of the lied. Similarly, the dissonant  7–7–6–4–2 on G of bar 13 of the
Adagio is restored in bar 4 of the aria, after its absence in the lied, but with the prefatory melodic octave rise g'–g" in bars 2–3, first introduced in the analogous bars 2–3 of the lied. Comparable syntheses may be found in
other memes in the aria, such as that concerning meme D noted in Section 2.
STEVEN JAN
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 359

change, driven by mutation and emigration/immigration at level 3.52

Such mutations, in Mozart’s replication of the memeplex and in its
broader evolutionary history to be considered below, often involve
increasing complexity, a process of expansion and filling-in of the
parameters of pitch, rhythm and texture, although there are certainly
exceptions to this general principle. Taking Example 1 above as its
frame of reference, Table 1 charts the principal evolutionary changes
occurring, in the parameter of pitch, at level 2 of Mozart’s memeplex
as a consequence of these lower-level changes.
From this, and from observations made in Section 4 above, it seems
that memetic evolution is contingent upon hierarchic location. The
various forms of the seven level-2 memes have a common middle-
ground schematic structure despite the surface variation engendered
by memes at level 3; and the meme at level 1, certainly in the Adagio
and aria, is largely invariant, despite the differences at lower levels.
There appears, in other words, to be greater resistance to mutation the
higher up the memetic-replication (Figure 1(a)) or Schenkerian-
structural hierarchy (Figure 1(b)) a meme is situated. Conversely,
there is a greater degree of what might be termed memetic entropy at
the musical surface than at higher replication/structural levels.
Another way of conceptualizing this phenomenon is in terms of the
top-down and bottom-up perceptual/cognitive dichotomy discussed in
connection with Example 2 above. The top-down pressure of the level-
2/middleground schema (style structure) is a force for conservation,
acting to preserve the meme unchanged over time, whereas the
bottom-up pressure of the level-3/foreground features (style forms) is
a force for mutation, driving the evolutionary dynamic. Ultimately, as
noted in Section 3 above, the fact that our perceptual and cognitive
systems quickly become weary of the familiar – and that the verbal-
conceptual ‘novelty meme’ (‘that which is new and different is to be
valued’) has been a powerful element of Western culture since the
Renaissance – means that the balance is tipped slightly towards the
bottom-up pressure, in favour of those memes which maximize their
propensity to imitation by means of perceptually and cognitively arrest-
ing mutations. Again, to reiterate an observation made in Section 3,
there is no conscious intentionality here. ‘ “Good” [memes]’, to para-
phrase Dawkins, ‘are blindly selected as those that survive in the
[meme] pool. This is not a theory; it is not even an observed fact: it is
a tautology.’53
Moving on to the level of dialect, it is the case that memes A–G are
not confined to the four Mozart pieces under investigation, although
52 Strictly speaking, each level-3 meme, once it has been mutated, ceases to be a meme until a

copy is made of the new form. When Mozart mutated the memes of the arietta in the Adagio in
May 1787 the resultant level-3 particles may not have been memetic, for they may, for a time, have
been unique, not yet imitated by other composers. Lynch speaks of a ‘mnemon’ – ‘an item of
brain-stored memory. When copied from one brain to another, it becomes a meme’ – in such
cases (Aaron Lynch, Mnemon 1998a: Y2K Memes (Issue 1), <http://www.mcs.net/~aaron/
Mnemon1998a.html>). See Jan, ‘Replicating Sonorities’, Sections 7 and 8, for more detailed
discussion of the mechanism of memetic mutation and stylistic evolution.
53 Dawkins, The Selfish Gene, 86.
360 STEVEN JAN

it may be that their conglomeration into a memeplex first occurred in

Mozart.54 These level-2 memes may perhaps be found separately, or
indeed as a memeplex, elsewhere within Mozart’s idiom, but they are
not restricted to this level of the cultural-stylistic hierarchy, being part
of the wider dialect of eighteenth- and nineteenth-century musical
style. They therefore have an evolutionary history, beyond the limited
snapshot shown in Table 1 above, of forms antecedent and consequent
to those of Mozart, of incarnations before and after Mozart’s repli-
cation of them as a memeplex in the four pieces considered here.
Table 2 indicates possible antecedent and consequent coequals of
memes A–G from the dialect of Viennese classicism and other chrono-
logically and geographically adjacent dialects. ‘Possible’ here means
that while a direct imitative and causal connection between the
antecedent patterns and Mozart’s memeplex, and between Mozart’s
memeplex and the consequent configurations, may be difficult to
verify in the absence of what Nattiez would term strong ‘poietic’
evidence,55 the connections between the particles are certainly sugges-
tive. Nevertheless, it may well be the case that some or all of Mozart’s
memes originated from different antecedents from those proposed
here, or from a synthesis of multiple antecedents (which may or may
not have included those proposed here); and it may also be the case
that Mozart’s memes begat different consequents from those proposed
here, and that the consequents here therefore arose from antecedents
other than Mozart’s.
Despite these caveats, the coequals proposed in Table 2 retain the
underlying middleground schematic (style structure) configuration of
Mozart’s level-2 memes, as represented in Example 3 above, but differ
to varying degrees in their foreground (style form) detail, the pattern-
ing analogous to that at level 3 in the Mozart memeplex. The level-2
memes therefore retain their archetypal identity over considerable
periods of time – over a century, in the case of meme C – despite their
shifting articulation by different surface features. In this sense, each of
Mozart’s memes A–G and their antecedent and consequent analogues
may be regarded as constituting members of a synchronic set consist-
ing of units of sufficient schematic consistency to be read as essentially
the same meme at the shallow middleground level, despite the often
substantial diachronic-evolutionary changes at the foreground level
driven by mutational pressures over time.
Example 4 illustrates a small selection of these coequals, from
which the following issues (among others) arise (see also Table 2).
Example 4(a), the antecedent coequal of meme A, is close morpho-
logically to its first incarnation in the Mozart pieces, bars 6–9 of the
arietta, with its upper-voice scale-degree sequence 5̂–1̂–6̂–5̂ and inflex-
ion to iv via I 3 over a tonic pedal. The registral disposition of the
melodic pitches in the Stamitz is, however, closer to the form of the

54 The difficulty of verifying this point – requiring the examination of all Mozart’s output –

will, however, be readily apparent.

55 Nattiez, Music and Discourse, 10–16.
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 361

TABLE 2

POSSIBLE ANTECEDENT AND CONSEQUENT COEQUALS OF MEMES A–G

antecedent coequal consequent coequal

meme A Karl Stamitz: String Verdi: La Traviata (1853), no. 6,
Quartet in G minor, op. 2 ‘Ah, fors’è lui che l’anima’, bars
no. 2 (c.1774), I, bars 1–4 26–30

meme B Haydn: Keyboard Sonata in Schubert: Quartettsatz in C minor,

G minor, Hob. XVI/44 D.703 (1820), bars 181–91
(c.1769?), I, bars 9–12

meme C Handel: Rinaldo, HWV 7a Beethoven: Fidelio, op. 72 (1814),

(1711), recitative before no. 9, ‘Abscheulicher! wo eilst du
no. 20, ‘Qual incognita hin?–Komm, Hoffnung’, bars 62–4
forza’, bars 17–18

meme D Bach: Goldberg Variations, Beethoven: 32 Variations on an

BWV 988 (1742), Variation Original Theme in C minor, WoO
25, bars 2–4 80 (1806), bars 3–6

meme E Haydn: String Quartet in Beethoven: Missa solemnis in D

F minor, op. 20 no. 5 major, op. 123 (1823), Kyrie, bars
(1772), I, bars 10–12 201–2

meme F Haydn: Keyboard Sonata in Beethoven: Piano Sonata in A

E minor, Hob. XVI/47 major, op. 110 (1822), III, bars
(before 1766), I, bar 3 12–14

meme G Haydn: String Quartet in Beethoven: Grosse Fuge in B
major,

G major, op. 17 no. 5 op. 133 (1825), bars 727–34
(1771), IV, bars 1–8

meme in the Adagio and, especially, the aria; the continuation of

meme A in the aria (bars 3–4) also echoes that of Stamitz (bars 5–6).
The consequent coequal of meme A, from the first phrase of ‘Ah, fors’è
lui’ in La Traviata, is a clear echo of this element of the aria, eliminat-
ing, however, the earlier piece’s dissonant  7–  7–6–4–2 on G – a
change which supports the assertion, made at the start of this section,
that memetic mutation does not invariably involve increasing com-
plexity.
Example 4(b), the antecedent coequal of meme C, prefigures the
melodic shape of the meme in the arietta, although without the voice-
exchange structure of the Mozart pieces. Such gestures are common
in recitativo secco and it may be the case that they are the original
source of meme C. Certainly its forms in the Mozart pieces are often
declamatory and recitative-like in character, as is the consequent
coequal of this meme, from a climactic moment in Leonore’s ‘Komm,
Hoffnung’.
362 STEVEN JAN

Example 4. Some antecedent and consequent coequals of memes

A–G.
(a) Antecedent coequal of meme A: Karl Stamitz, String Quartet in G minor, op. 2 no. 2 (c.1774),
I,
(a)bars 1–8.
Andantino grazioso

Violin

Viola I

Viola II

Violoncello

Vln

Vla I

Vla II

Vc.

(b) Antecedent coequal of meme C: Handel, Rinaldo, HWV 7a (1711), recitative before no. 20,
‘Qual incognita forza’, bars 16–18.
(b)
Rinaldo Eustazio Rinaldo

(l’ar )dir! Non de vo. Pen sa a ca si tuoi. Il cor non pa ve.

4+ 6
2

Example 4(c), the consequent coequal of meme D, while initially

appearing dissimilar in melodic structure to its Mozartean precursors,
has an analogous upper-voice shape (the shallow middleground
pattern e "–f"–f "–g", corresponding to Mozart’s b '–c"–c "–d") above
the tetrachord. Unlike Mozart’s version of this meme, however,
Beethoven’s uses the full chromatic descending tetrachord (i.e. includ-
ing  7̂), as does the antecedent coequal of meme D, from Variation 25
of Bach’s Goldberg Variations. It is clear from these examples that the
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 363

Example 4 continued
(c) Consequent coequal of meme D: Beethoven, 32 Variations on an Original Theme in C minor,
WoO
(c) 80 (1806), bars 1–8.
Allegretto

(d) Antecedent coequal of meme E: Haydn, String Quartet in F minor, op. 20 no. 5 (1772), I,
bars
(d)
9–12.
Allegro moderato
9

11

Baroque descending chromatic tetrachord – with its coadapted verbal-

conceptual memes articulating notions of death, burial and grief – is,
ironically, a meme of considerable vigour and longevity.
Example 4(d), the antecedent coequal of meme E, anticipates the
circling around the dominant note in the bass by its upper and lower
neighbour notes – these supporting augmented-sixth and diminished-
seventh chords respectively – which is a characteristic of the meme
particularly evident in the arietta and aria. The consequent coequal of
this meme, from the Kyrie of Beethoven’s Missa solemnis, shows a
similar, albeit more compressed, level-3 coadaptation between
upper/lower neighbour-note linear pitch sequences, their associated
harmonies and the dominant triad.
364 STEVEN JAN

Example 4 continued
(e) Consequent coequal of meme G: Beethoven, Grosse Fuge in B
major, op. 133 (1825), bars
726–34.
(e)
Allegro molto e con brio

726

cresc.

729

732

al più

It is appropriate at this point to recall that memes B and E in the

Mozart pieces have an essentially analogous configuration, consisting
basically of a prolonged dominant harmony. Their locations are,
however, quite different: meme B is situated towards the beginning of
the pieces, prior to a move from the tonic to another harmonic/tonal
area, whereas meme E is deployed towards the end of the pieces, in
preparation for a firm closure on the tonic. This use of patterns of
equivalent structure in two functionally different contexts is
comparable to Gould and Vrba’s notion of ‘exaptation’ in biology.
They note that an ‘adaptation’ is
any feature that promotes fitness and was built by selection for its current
role (criterion of historical genesis). The operation of an adaptation is its
function. . . . We may also . . . [label] the operation of a useful character not
built by selection for its current role as an effect. . . . But what is the unse-
lected, but useful character itself to be called? . . .
We suggest that such characters, evolved for other usages (or for no
function at all), and later ‘coopted’ for their current role, be called
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 365

exaptations. . . . They are fit for their current role, hence aptus, but they were
not designed for it, and are therefore not ad aptus, or pushed towards
fitness. They owe their fitness to features present for other reasons, and are
therefore fit (aptus) by reason of (ex) their form, or ex aptus.56
While an account of the aetiology and evolution of the dominant
prolongation is beyond the scope of this article, it is not difficult to
imagine, in the context of late-modal/early-tonal music, the notion of
the end-dominant being an exaptation from a historically earlier begin-
ning-dominant – or vice versa.
Example 4(e), the consequent coequal of meme G, is closest to the
aria in its harmonic structure of chords based upon the descending-
third sequence G–E
–C (the antecedent coequal of this meme, from
the finale of Haydn’s String Quartet op. 17 no. 5, is perhaps closer,
certainly in melodic terms, to the quintet’s finale theme). Beethoven’s
progression occurs, however, in the context of B
major, not G minor,
and starts with an initial third drop B
–G. Moreover, Beethoven’s
phrase is not a mutation of the whole of Mozart’s meme but of only a
part of it; were Mozart’s bar 38 to begin in B
, then Beethoven’s
passage would be a copy of Mozart’s bars 38–395. In this process of
cutting and splicing, in which a segment of a meme is taken from an
antecedent source and embedded into a later context, the passage
from the Grosse Fuge may be regarded as a memetic analogue to the
biological process of ‘crossing over’. Here, the random exchange of
components of the chromosome occurring during meiosis (cell
division) results in segments of genes being extracted and repositioned
on other chromosomes.

6. VERBAL-CONCEPTUAL MEMES COADAPTED WITH THE MEMEPLEX

Although the memetic paradigm has been employed here primarily to

interpret purely musical connections between the four pieces under
investigation, it should be acknowledged that three out of the four
involve texts. As suggested in Section 3 above, memes are propagated
in many media, of which the verbal-conceptual realm is one of the most
important. I conclude, therefore, by considering, in memetic terms,
some issues arising from the association of verbal-conceptual memes
with elements of the musical memeplex.
The arietta and the aria form part of a sequence of five operatic solos
in the key of G minor from the last decade of Mozart’s life: ‘Padre,
germani, addio!’ (Illia), no. 1 of Idomeneo, K.366 (1780); ‘Traurigkeit
ward mir zum Lose’ (Konstanze), no. 10 of Die Entführung aus dem
Serail, K.384 (1782); ‘Da schlägt die Abschiedsstunde’; ‘Zum Leiden
bin ich auserkoren’ (Königin der Nacht), no. 4 of Die Zauberflöte; and
‘Ach ich fühl’s’. Unlike Mozart’s previous G minor arias, also five in

56 Stephen J. Gould and Elisabeth S. Vrba, ‘Exaptation – A Missing Term in the Science of

Form’, Paleobiology, 8/1 (1982), 4–15 (p. 6).

366 STEVEN JAN

number,57 these last five are intimately linked by the content of their
texts. To my knowledge, no commentator has identified that, despite
the widely different dramatic context of each aria, the sentiments of
the characters have a common motivation: the experience of parting
or loss.58
From this, it seems reasonable to suggest that G minor is a key Mozart
began to associate, in the 1780s, with loss – not, as Einstein maintains,
with the arguably less focused notion of ‘fate’59 – and that the repli-
cation of the same memes in the four pieces considered here may
imply that the G minor quintet’s Adagio has ‘hidden’ the same verbal-
conceptual memes explicitly presented in the three vocal contexts. Put
in more conventional terminology, the quintet’s Adagio might be seen
as having a ‘subtext’ of parting or loss. This notion is readily explica-
ble from a semiotic perspective, wherein the musical memes fulfil the
function of ‘signifier(s)’, and the verbal-conceptual memes with which
they are coadapted in the arietta, lied and aria act as the ‘signified’. In
the Adagio, the signified, whilst not explicitly presented, is neverthe-
less resonant by virtue of the unequivocal presence of the signifiers.
Memes A, D and F, in particular, are strongly conformant in the
Adagio and aria, and it seems reasonable to suggest that a memetic
perspective may facilitate a hermeneutic analysis of the Adagio,
whereby one may ‘retrospectively’ append the verbal-conceptual
memes associated with certain musical memes in the aria to the corre-
sponding musical memes in the Adagio.
Why does the F minor Trennungslied contain memes replicated in
three G minor pieces? This seems to be the consequence of some
striking similarities between the verbal-conceptual memes of ‘Da
schlägt’ and ‘Die Engel’, which may have acted as a cue for the repli-
cation in the lied of the musical memes associated with the verbal-
conceptual memes of the arietta. The texts of the arietta, by the
librettist of Die Entführung aus dem Serail, Gottlieb Stephanie
(1741–1800), and of the lied, by Klamer Eberhard Karl Schmidt
(1746–1824), contain general similarities in imagery which gravitate
around the remarkably specific replication of certain verbal-concep-
tual memes. In the arietta, ‘the hour of parting strikes’ and the singer
contemplates, in the meme ‘Und du, und du, vielleicht auf ewig
vergißt . . .’ (bars 26–32), the possibility that her departed lover will
forget her. In the lied, this same meme is a persistent refrain at the end
of the vocal element of each strophe (bars 12–19). This verbal-concep-
tual meme occurs in the context of meme C, perhaps the most overtly
rhetorical of the seven level-2 memes in its surface realization; it seems

57 These arias are ‘Betracht dies Herz’ (Der Engel), no. 2 of the Grabmusik (Passionskantate),

K.42/35a (1767); ‘Nel sen mi palpita’ (Aspasia), no. 4 of Mitridate, re di Ponto, K.87/74a (1770);
‘Ma qual virtù’ (Cabri), no. 2 of La Betulia liberata, K.118/74c (1771); ‘Vorrei punirti indegno’
(Arminda), no. 13 of La finta giardiniera, K.196 (1775); and ‘Tiger! wetze nur die Klauen’ (Zaïde),
no. 13 of Zaïde, K.344/336b (1779).
58 Jan, Aspects of Mozart’s Music in G Minor, 56–7.
59 Alfred Einstein, Mozart: His Character, his Work, trans. Arthur Mendel and Nathan Broder (6th

edn, London, 1966), 264.

THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 367

to come from the world of opera seria, as does its possible progenitor
shown in Example 4(b) above.
The memetic connections between these texts are the result of their
descent from Metastasio’s celebrated canzonetta ‘Ecco quel fiero
istante’ (La Partenza) of 1746, as shown by the extracts in Table 3. ‘Die
Engel’ (1779) is a ‘second-generation’ copy, being modelled on ‘Da
schlägt des Abschieds Stunde’ (Die Trennung) of 1773 by Johann
Joachim Eschenburg (1743–1820), an earlier reworking of Metastasio’s
canzonetta.60 Both Eschenburg and Schmidt preserve the metrical
structure of Metastasio’s poem in their reworking of its content into
German. ‘Da schlägt’, as Table 3 shows, is based on the first verse of
the Metastasio and is derived from the Eschenburg version, as
evidenced by the replication, with mutation, of the meme ‘Da schlägt
des Abschieds Stunde’; indeed, Stephanie takes nothing from Schmidt
that is not already in Eschenburg. The memetic connections between
the four poems are italicized in Table 3.
From these observations one might identify three context-specific
conditions governing the relationship in Mozart between the key of G
minor, the memeplex considered here, and its associated verbal-
conceptual memes. First, it may be said that the strongest ‘trigger’ for
this memeplex appears to be verbal-conceptual memes expressing the
sentiment of loss plus the key of G minor. Secondly, the complex may
also appear in other minor keys provided that the verbal-conceptual
memes are present. Thirdly, from the case of the Adagio, it can be seen
that the memeplex may also appear when only the key of G minor is
present, in the absence of the verbal-conceptual memes. Nevertheless,
it will also be evident – given that the memeplex does not appear in
the G minor ‘loss’ arias ‘Padre, germani’, ‘Traurigkeit’ and ‘Zum
Leiden’ – that the first of the three conditions noted above is neces-
sary but not sufficient.

7. CONCLUSION: TAKING THE ‘MEME’S EYE VIEW’

The following closing points should draw together the various themes
considered here. First, I hope the discussion of the four Mozart pieces
in Sections 2, 4 and 5 is seen as offering convincing evidence for the
validity of the memetic paradigm, both as a means of understanding
the nature of human culture in general, and as a specific tool for
comprehending synchronic musical organization and diachronic
musical style change.
Secondly, I hope to have offered a different perspective on Mozart’s
compositional processes, showing how large-scale intertextual struc-
tures may have been an important element of his approach to compo-
sition, and indeed that of other composers.
Thirdly, however, I hope to have suggested – yet not without a tinge

60 Ludwig Ritter von Köchel, Chronologisch-thematisches Verzeichnis sämtlicher Tonwerke Wolfgang

Amadé Mozarts, ed. Franz Giegling, Alexander Weinmann and Gerd Sievers (8th edn, Wiesbaden,
1983), 583.
 TABLE 3
368

VERBAL-CONCEPTUAL MEMES IN FOUR ‘TRENNUNG’ POEMS

Metastasio 1746 Eschenburg 1773 Schmidt 1779 (lied) Stephanie 1786 (arietta)

Verse 1 Verse 1 Verse 1

Ecco quel fiero istante; Da schlägt des Die Engel Gottes weinen, Da schlägt die
Nice, mia Nice, addio. Abschieds Stunde, wo Liebende sich trennen! Abschiedsstunde,
Um grausam uns zu trennen; um grausam uns zu trennen.
Come vivrò, ben mio, Wie werd’ ich leben können, Wie werd’ ich leben können, Wie werd’ ich leben können,
così lontan da te? O Mädchen, ohne dich? o Mädchen, ohne dich? o Damon, ohne dich?
Io vivrò sempre in pene, Ein Fremdling aller Freuden Ein Fremdling allen Freuden, Ich will dich begleiten,
io non avrò più bene; Leb’ ich noch, um zu leiden, leb’ ich fortan dem Leiden! im Geist dir zu Seiten
schweben um dich.
e tu, chi sa se mai Und du, und du Und du? und du? Und du, und du,
vielleicht auf ewig Vielleicht auf ewig vielleicht auf ewig
ti sovverrai di me! Vergißt nun Daphne mich! vergißt Louisa mich! vergißt dafür auf mich!
Doch nein!
wie fällt mir sowas ein?
Du kannst gewiß nicht treulos sein,
ach nein, ach nein.

Here is the fierce moment; The hour of parting strikes, The angels of God weep, The hour of parting strikes,
Nice, my Nice, farewell. that so cruelly sunders us; where lovers separate! that so cruelly sunders us.
How shall I live, my darling, How shall I be able to live, How shall I be able to live, How shall I be able to live,
thus far from you? without you, oh maiden? without you, oh maiden? O Damon, without you?
I shall live ever in pain, A stranger to all joys, A stranger to all joys, I want to go with you,
I shall have no more good things; I live on through the sorrow, I live on through the sorrow! in spirit by your side
to hover near you.
and you, who knows if by chance And you? and you? And you? and you? And you, and you,
you should think again of me! Perhaps for ever Perhaps for ever perhaps for ever
Daphne will forget me! Louisa will forget me! will forget me!
But no!
How can I think such a thing?
You surely cannot be untrue,
STEVEN JAN

ah no, ah no.
THE EVOLUTION OF A ‘MEMEPLEX’ IN LATE MOZART 369

of regret – that taking the ‘meme’s eye view’61 of these pieces calls into
question traditional conceptions of Mozart’s conscious intentionality –
and indeed, by extension, that of every composer – and at the same
time illustrates how fragile are the boundaries of his works. This is
because, as was implied at the end of Section 3,
The self is a vast memeplex – perhaps the most insidious and pervasive
memeplex of all. I shall call it the ‘selfplex’. The selfplex permeates all our
experience and all our thinking so that we are unable to see it clearly for
what it is – a bunch of memes. It comes about because our brains provide
the ideal machinery on which to construct it, and our society provides the
selective environment in which it thrives.62
‘If this memetic analysis is correct’, Blackmore asserts,
the choices you make are not made by an inner self who has free will, but
are just the consequence of the replicators playing out their competition in
a particular environment. In the process they create the illusion of a self
who is in control.63
In a similar vein, Dennett asks ‘Cui bono?’, apropos of ‘those cases
when push comes to shove, and the interests of the body . . . conflict
with the interests of the genes’.64 Memetics applies the same principle
to the meme and therefore motivates a model of the composer as
vessel, a conception which maintains that, essentially, Mozart was the
conduit through which the selfish memes he imitated and mutated
passed. In short, I suggest the evidence considered here speaks, ulti-
mately, for the power of the replicator over that of the vehicle, to
employ Dawkins’s dichotomy.65
Fourthly, as implied in Section 4, I believe that the memetic
paradigm can motivate and guide a rethinking of the theory and
practice of analytical musicology, leading to a reconsideration of the
notion of structural value in analysis. A significant conclusion deriving
from the material considered here seems to be that a replicated struc-
ture common to two or more contexts should be privileged analytically,
irrespective of whether elements of that structure satisfy, for instance,
the Fuxian/Schenkerian voice-leading considerations (or, for that
matter, the Fortean/neo-Riemannian set/group-theoretical criteria)
normally imposed as conditions of structural importance in analysis.
In this sense, to a consideration of Satz (the surface profile of the
music and its articulation into units of structure and expression) and
of Ursatz (the deep-structural products – to reverse Schenker’s aeti-
ology – of voice-leading processes at shallower levels of organization),
analysis might profit from incorporating a sensitivity to what might be
termed the Memesatz. This dimension of music might be defined as
those salient intertextual structures generated and defined by repli-
cation. Such a focus on replication tends automatically to emphasize
61 Blackmore, The Meme Machine, Chapter 4.
62 Ibid., 231.
63 Susan Blackmore, ‘Meme, Myself, I’, New Scientist, 161/2177 (1999), 40–2 (p. 42).
64 Dennett, Darwin’s Dangerous Idea, 325.
65 Dawkins, The Extended Phenotype, 112.
370 STEVEN JAN

questions of perception and cognition of musical patterning, and

thereby highlights those elements of the music which have the greatest
reality and immediacy to the listener.
Finally, as suggested in Section 6, an examination of musico-concep-
tual memeplexes – those associations between memes in the purely
musical realm and those in the verbal-conceptual dimension – can
offer revealing hermeneutic insights into music. These rich memetic
relationships may help elucidate some of music’s veiled referential and
connotative meanings and offer a means of unravelling the memetic
threads that are the psychological motivations of the composer and the
emotional responses of the listener.

ABSTRACT
‘Memetics’, a concept most elegantly expounded by Richard Dawkins in The
Selfish Gene, asserts that human culture consists of a multitude of units trans-
mitted between individuals by imitation and subject to evolutionary pressures.
Such particles, ‘memes’, are broadly analogous to the genes of biological
transmission. Four late pieces of Mozart’s, including Pamina’s aria ‘Ach ich
fühl’s’ from Die Zauberflöte, are examined in terms of the meme concept and
a conglomeration, or ‘memeplex’, consisting of seven memes is identified
within them. The nature of the musical memeplex, in this specific case and
also more generally, is considered, particularly from the perspective of its
location at different levels of the structural hierarchy. The evolutionary
history of some of Mozart’s memes is examined with reference to selected
passages from works of the eighteenth and nineteenth centuries. Finally,
relationships between the musical memes under investigation and memes in
the verbal-conceptual realm are explored.