Content uploaded by Adam Ockelford
Author content
All content in this area was uploaded by Adam Ockelford on Nov 05, 2017
Content may be subject to copyright.
UNCORRECTED PROOF
ARTICLE IN PRESS
1
2Learning and creativity in a prodigious
3musical savant
4A. Ockelford
a,
*, Linda Pring
b
5
a
Royal National Institute of the Blind, 105 Judd Street, London WC1H 9NE, UK
6
b
Goldsmith’s College, University of London, United Kingdom
7Abstract. Preliminary findings are presented of work undertaken with a prodigious dsavantT
8pianist–Derek Paravicini–in which he learnt a specially-composed piece (The Chromatic Blues)
9over a number of hearings. All Derek’s efforts were recorded digitally, which enabled precise
10transcriptions to be made. These were analysed in the context of a newly-developed theory of
11how musical structure is recognised and understood. The results are startling in musicological
12terms and (it is hypothesised) yield fascinating insights into the subtlety, ingenuity and creativity
13of Derek’s musical mind. It is hoped that the findings from this study will in due course inform
14the wider debate on the nature of musical learning, memory and creativity. D2005 Published by
15Elsevier B.V.
16Keywords: Musical learning; Creativity; Savant; Musical structure; Blues
17
181. Introduction
19This paper presents initial findings from a study that forms part of the dREMUST
20Project–dResearching Exceptional MUsical SkillT–which is a joint initiative of the
21Royal National Institute of the Blind, London and the Psychology Department of
22Goldsmith’s College, University of London. The project focuses on the abilities of
23musical dsavantsT: people with exceptional skill in the context of learning disabilities.
24The current study was undertaken with Derek Paravicini, 25, a prodigiously skilled
25pianist, despite have severe learning difficulties and being totally blind (through
0531-5131/ D2005 Published by Elsevier B.V.
doi:10.1016/j.ics.2005.05.051
* Corresponding author. Tel.: +44 20 7391 2149.
E-mail address: adam.ockelford@rnib.org.uk (A. Ockelford).
International Congress Series xx (2005) xxx – xxx
www.ics-elsevier.com
ICS-05572; No of Pages 5
DTD 5
UNCORRECTED PROOF
ARTICLE IN PRESS
26dretinopathy of prematurityT). It is one of a series of studies that aims to glean new
27insights into the nature of savants’ musical ability, in particular learning, memory,
28reproduction and creativity, and to use these findings to generate models of the
29processes involved, adopting a fusion of music-psychological and music-theoretical
30approaches [1]. The underlying methodological assumption is that the musical pieces
31and fragments that savants produce following exposure to controlled musical input–
32specifically the extent to and manner in which their responses can be considered to
33derive from the stimuli with which they are presented–provide powerful evidence of
34the cognitive processes involved [2]. In the longer term, it is intended that our findings
35will be used to underpin new didactic approaches in working with savants (and others),
36inform the wider debate on the nature of musical ability in general and stimulate
37further research.
382. Method
39A piece was especially composed for the experiment which was entitled The
40Chromatic Blues—dCBT, whose structure is shown in Fig. 1 (a score and digital
41recording of this and other material cited in the paper are available from the first
42author).
43Statistically, CB comprises 312 events occurring within 49 s (an average rate of 6.37
44events per second). The opening four bars are shown in Fig. 2.
45CB was presented to Derek over a number of sessions which were structured as
46follows:
47
48(a) Derek plays CB as well as he is able (in all except the first session)
49(b) Derek listens to CB
50(c) Derek plays CB (again)
51(d) Derek listens to CB again
52
Fig. 1. Structure of The Chromatic Blues.
Fig. 2. Opening bars of The Chromatic Blues (nAdam Ockelford, 2003).
A. Ockelford, L. Pring / International Congress Series xx (2005) xxx–xxx2
UNCORRECTED PROOF
ARTICLE IN PRESS
53All material was performed on a Korg touch-sensitive keyboard SP-200 (using dPiano
541Tsound), recorded as MIDI and digital sound files, and reproduced using the same
55system. The pattern of sessions is shown in Fig. 3.
563. Results
57Derek’s first attempt at playing CB (immediately having heard it for the first time) began as
58follows (Fig. 4). Overall, this version (dCB-D1T) was derived from CB with a strength of derivation
59of 0.31 (using the excerpts for analysis set out in Fig. 1)[1].
60As this excerpt illustrates, it is a feature of Derek’s playing (rather like a natural language
61speaker) that what he produces invariably makes musical sense. As he was asked to play CB as well
62as he could, because he achieved a derivation index of only 0.31, we can assume that the piece was
63beyond the capacity of his working memory and that he was forced to modify, introduce or create
64material to dplug the gapsT. It appears that he did this, quite intuitively, in a number of ways. First,
65within the general style of the opening of CB (including tonality, metre and phrase structure) he
66utilised a number of subtle transformations of its musical material, including re-ordering fragments,
67and separating pitch and rhythmic elements and re-combining them in new ways. Second, he
68appeared to introduce material from other pieces in similar style including the dbass riffTfound in
69bars 1 and 2 (typical of Count Basie, for example), whose harmonic structure is based on a common
70dBlues turnaroundT, with the opening four bars outlining a design reminiscent of the chorus of Paper
71Moon. It is extremely doubtful that these apparent borrowings were part of Derek’s conscious
72thinking, but they give an insight into the possible nature of his nonconscious musical processing
73that was in train.
Fig. 3. Pattern of sessions with Derek and The Chromatic Blues.
Fig. 4. The opening four bars of Derek’s first attempt at The Chromatic Blues (dCB-D1T).
A. Ockelford, L. Pring / International Congress Series xx (2005) xxx–xxx 3
UNCORRECTED PROOF
ARTICLE IN PRESS
74Derek’s next effort at reproducing CB (in Session 2, which was separated from Session 1 by 2
75days) resulted in a further version (dCB-D2T) that was comparable to, though distinct from, CB-D1.
76Having heard CB again, Derek then reverted to CB-D1. In Session 3 (after a further 5 days) he
77returned to CB-D2 (see Fig. 5). However, his second attempt in this session revealed a distinct and
78audible transition from CB-D1 to a rendition much closer to CB (Fig. 6).
79The indices of derivation that Derek achieved over the entire period are as follows (see Fig. 7). It
80is worthy of note that there are marked improvements following the first break of a month (0.66 to
810.74) and 3 months (0.76 to 0.84), and that even following the 6 months break, a level of 0.83 is
82achieved, followed by the highest score of all (0.88) at the second attempt in that session. For
83reference, an exponential trend is indicated. 84
854. Discussion
86These preliminary data from dREMUS Study 1Treveal a number of findings that are of
87interest. First, it is evident that Derek does indeed have an exceptional memory. Another
88jazz pianist (who, like Derek, had absolute pitch and was used to learning pieces by ear)
89attempted to learn CB under precisely the same conditions and, although his data await full
90analysis, it is evident that his performance was significantly inferior to Derek’s. However,
91Derek’s recall was by no means straightforward: musical analysis suggests that initially he
92was able to remember only certain salient fragments and features—details within an
93overall outline of what had occurred. Derek’s instinct to produce something that made
94musical sense subsequently drove him to manipulate what he could remember, whereby
95the moment-to-moment demands of musical coherence apparently took precedence over
96his powers of recall. Transformations of material from CB were made as well as the
97introduction of fragments and features from stylistically congeneric sources (memories of
98which hearing CB had evidently stimulated in Derek).
99Interestingly, analysis of data beyond this paper indicates that the most potent source of
100material for Derek at any given juncture was not CB itself (even immediately after having
101heard it), but his previous renditions, suggesting a strong interference effect. This was
102particular marked in the early stages of learning, when, as we have seen, three different
103versions of CB were maintained in parallel. Indeed, in renditions 7 and 9, Derek used CB-
104D2 as an introduction to a convincing performance of CB itself!
105Clearly, then, Derek’s developing model(s) of CB was far more than a series of surface
106percepts which he was able to reproduce, automaton-like at the keyboard. Evidently,
Fig. 5. The opening four bars of CB-D2.
Fig. 6. The audible switch in rendition 7 from CB-D1 to CB.
A. Ockelford, L. Pring / International Congress Series xx (2005) xxx–xxx4
UNCORRECTED PROOF
ARTICLE IN PRESS
107despite his inability to verbalise almost anything about his musical achievements, Derek
108has a lively and engaging musical mind with considerable strengths but with human
109frailties too. His efforts are informed by an intuitive understanding of musical syntax and
110structure that would do credit to any mature musician. His ability to take material from a
111range of sources and mould it in real time into something that is of musical value in its
112own right indicates a genuine creativity that merits further research. And, following
113completion of the current study, it is to this area that the REMUS project will turn its
114attention.
115References
116
117[1] A. Ockelford, Repetition in music: Theoretical and metatheoretical perspectives. London: Ashgate.
118[2] J. Sloboda, Exploring the Musical Mind: Cognition, Emotion, Ability, Function, Oxford University Press,
119Oxford, 2005.
120
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Series1 0.31 0.18 0.18 0.19 0.41 0.49 0.66 0.6 0.66 0.74 0.75 0.73 0.74 0.74 0.73 0.77 0.7 0.76 0.84 0.87 0.83 0.88
1 2a 2b 3a 3b 4a 4b 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b 10a 11a 11b 12a 12b
Fig. 7. Derek’s indices of derivation for CB over the test period (12 months).
A. Ockelford, L. Pring / International Congress Series xx (2005) xxx–xxx 5