How ‘Circumpolar’ is Ainu Music? Musical and Genetic Perspectives on the History of the Japanese Archipelago

Article (PDF Available)inEthnomusicology Forum 24(3):1-22 · September 2015with 845 Reads 
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
DOI: 10.1080/17411912.2015.1084236
Cite this publication
Understanding the cultural and genetic origins of the Ainu of northern Japan has important implications for understanding the history of the Japanese archipelago. Ethnomusicologists have tended to emphasise connections between Ainu music and a ‘circumpolar’ culture area. However, the ‘dual structure’ model from physical anthropology describes the Ainu as descendants of the first inhabitants of Japan with minimal circumpolar influence. To examine Ainu musical diversity empirically from a comparative perspective, we analysed 680 traditional songs from two Ainu and 33 surrounding East Asian and circumpolar populations. The Ainu repertoire contained a majority (∼50%) of unique stylistic song-types and lower frequencies of types shared with circumpolar (∼40%) and East Asian (∼10%) populations. These frequencies were similar to the corresponding frequencies of mitochondrial DNA types within the Ainu gene pool (∼50%, ∼30% and ∼20%, respectively), consistent with an emerging ‘triple structure’ model of Japanese archipelago history.
Figures - uploaded by Hiromi Matsumae
Author content
All content in this area was uploaded by Hiromi Matsumae
Content may be subject to copyright.
How circumpolaris Ainu music? Musical and genetic
perspectives on the history of the Japanese archipelago
Patrick E. Savage
, Hiromi Matsumae
, Hiroki Oota
, Mark Stoneking
Thomas E. Currie
, Atsushi Tajima
, Matt Gillan
and Steven Brown
Department of Musicology, Tokyo University of the Arts, Tokyo, Japan;
Department of Anatomy, Kitasato
University School of Medicine, Sagamihara, Japan;
Department of Evolutionary Genetics, Max Planck Institute
for Evolutionary Anthropology, Leipzig, Germany;
Centre for Ecology & Conservation, College of Life &
Environmental Sciences, University of Exeter, Penryn campus, Cornwall, UK;
Department of Bioinformatics
and Genomics, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan;
Department of Liberal Arts, International Christian University, Tokyo, Japan;
Department of Psychology,
Neuroscience & Behaviour, McMaster University, Hamilton, ON, Canada
Understanding the cultural and genetic origins of the Ainu of
northern Japan has important implications for understanding the
history of the Japanese archipelago. Ethnomusicologists have
tended to emphasise connections between Ainu music and a
circumpolarculture area. However, the dual structuremodel
from physical anthropology describes the Ainu as descendants of
the rst inhabitants of Japan with minimal circumpolar inuence.
To examine Ainu musical diversity empirically from a comparative
perspective, we analysed 680 traditional songs from two Ainu and
33 surrounding East Asian and circumpolar populations. The Ainu
repertoire contained a majority (50%) of unique stylistic song-
types and lower frequencies of types shared with circumpolar
(40%) and East Asian (10%) populations. These frequencies
were similar to the corresponding frequencies of mitochondrial
DNA types within the Ainu gene pool (50%, 30% and 20%,
respectively), consistent with an emerging triple structuremodel
of Japanese archipelago history.
Received 6 June 2014
Revised 16 February 2015
Accepted 12 August 2015
Music; Genes; Dual Structure;
Triple Structure; Ainu;
Genes and languages have provided powerful tools for studying human biological and cul-
tural history, but a fuller picture of human history requires the integration of other
markers (Cavalli-Sforza, Menozzi and Piazza 1994; Diamond and Bellwood 2003).
There are several reasons why music may provide a useful supplement in understanding
human history (for a review see Savage and Brown 2013). Music is a universal but highly
diverse aspect of culture that shows both similarities and differences with language (Feld
and Fox 1994; Patel 2008; Wallin, Merker and Brown 2000). Comparative musicologists
and ethnomusicologists have long proposed that peoples and their musics share deep
and complexly interwoven histories (Nettl 2015). Recently, new quantitative evidence
has emerged of correlations between musical and genetic diversity, suggesting that
© 2015 Taylor & Francis
CONTACT Patrick E. Savage
music may be preserving ancient connections in the order of thousands of years or more
(Callaway 2007; Pamjav et al. 2012). One such study (Brown et al. 2014) found that music
and language were both signicantly correlated with genes but not with one another,
suggesting that music and language can capture at least partially independent aspects of
human history. Furthermore, music can move independently of both genes and languages,
and such dissociations may capture important information about historical interactions,
such as those involving trade, religion, colonialism and/or globalisation.
The idea that music might preserve traces of ancient human history was explored most
comprehensively in Alan Lomaxs(1968,1976,1980,1989) Cantometrics Project. Initially,
Lomaxs proposals relating music and human history were overshadowed by controversy
over his causal interpretation of correlations between singing style and social structure
(e.g., sexual sanctions being reected in nasal singing; for overviews see Averill 2003;
Dubinskas 1983; Wood 2008). However, there has recently been renewed interest in reviv-
ing and rening the methods and goals of Cantometrics to revisit questions of music and
human history by integrating advances in genetic anthropology.
Cantometricsco-creator Victor Grauer (2006,2011) has argued that the global distribution
of specic stylistic acoustic features coded by Cantometrics (e.g., vocal interlock) matches the
proposed dispersal of anatomically modern humans out of Africa 50,000 years ago. On a
regional scale, ethnomusicologists such as Nattiez (1999)andKeeling(2012) have argued
that the distribution of certain musical features (e.g., throat-rasping, animal imitation
songs) match proposed dispersals of humans from Siberia across the Bering Strait to the
Americas some time after the last glacial maximum that occurred 15,000 years ago. These
qualitative approaches allow for sophisticated and complex musical analyses, but at the cost
of subjectivity of analysis and sample selection, because the reader must take the word of
these authors that the features and examples they have presented tomake their case are repre-
sentative of broader trends in the repertoire, and are not simply a subset that has been cherry-
picked to match existing genetic knowledge (Leroi and Swire 2006; Stock 2006).
In contrast, several recent scientic projects have attempted more quantitative
approaches to directly compare musical and genetic diversity. One study compared
matching Cantometric musical data and autosomal genetic data from 39 populations in
Africa (Callaway 2007). Another compared musical data from Cantometrics and a
closely related scheme, CantoCore (discussed later), with mitochondrial DNA
(mtDNA) from nine indigenous populations in Taiwan (Brown et al. 2014). A third
study compared musical data based on an automated contour analysis of musical tran-
scriptions with mitochondrial and Y-chromosome data from 30 Eurasian and North
American populations (Pamjav et al. 2012). Each of these studies used slightly different
methods of statistical analysis, but all argued that the substantial correspondence
between musical and genetic diversity reected the preservation of shared musical and
genetic histories over long spans of time.
These quantitative studies using standardised
sets of features and samples provide more rigorous evidence of a relationship between
music and genetic history than the qualitative studies mentioned earlier.
None of the authors argued for a causal connection, and neither do we (i.e., there is no suggestion that these correlations
reect genes for music).
Of course, these quantitative studies are still only partially objective because they are based on data from classications or
transcriptions that were done by ear (a situation analogous to the use of cognate data based on expert judgments in
quantitative linguistic studies). Their reliability is thus imperfect, although the limited empirical testing published thus
because these studies compared average trends across entire repertoires, they are unable to
explore the same kinds of musically nuanced comparisons in which specic songs or
genres can be linked with different historical inuences.
Recently, Savage and Brown (2014) proposed a method of musical cluster analysis that
attempts to combine the precision of the quantitative methods with the song-level focus of
qualitative approaches. This method allows broad cross-cultural quantitative comparison
without assuming that a given musical culture can be represented by a single favoured
song style’—a highly criticised assumption of the original Cantometrics Project (Henry
1976; Rzeszutek, Savage and Brown 2012). Applying this technique to the same Taiwanese
musical dataset analysed by Brown et al. (2014), they were able to group each song into
one of ve different cantogroups(i.e., stylistic song-types) and map the regional distri-
bution of these cantogroups throughout Taiwan. However, quantitative techniques are
most effective when used to test and extend nuanced hypotheses built from qualitative
intuitions. While there were pre-existing hypotheses about the genetic and linguistic
origins of Taiwanese populations, there were no clear pre-existing theories about their
musical origins. In contrast, our current study chooses a specic culturethe Ainu of
northern Japanwho have been the subject of a number of musicological and genetic
studies. This allows us to quantitatively test these musical proposals, as well as to
compare them with existing genetic data.
Ainu music and the circumpolar region
A number of scholars have noted possible ancient connections between the Ainu and a
circumpolar music culture(Keeling 2012; Malm 1967; Nattiez 1983,1999; Tanimoto
2000,2001a) encircling the Arctic from the Saami in Arctic Europe to the Inuit in Green-
land. The most striking example is the throat-rasping genre shared between the Ainu
(Figure 1), Siberian populations (e.g., Chukchi) and the Canadian Inuit, but other
genres include animal imitation songs and shamanic drumming. In the Cantometrics
Project (still the only quantitative attempt to characterise Ainu music to date), Lomax
grouped the Ainu as part of a Siberiansong-style area characterised by guttural,
raspy, punchy, slurred, nonsense-syllable kind of soloizing in extremely uneven phrases
(Lomax 1976: 38). In this group he included not only the Ainu and most indigenous Siber-
ian populations, but also the Saami of Arctic Europe and several indigenous populations of
the Americas.
In contrast, research in physical anthropology and linguistics has tended to emphasise
links between the Ainu and the other inhabitants of the Japanese archipelago, rather than
any circumpolar connections.
Early theories proposing that the Ainu were a lost Cauca-
soid group (Bickmore 1868) have now generally been rejected in favour of Haniharas
(1991)dual structuremodel. This model proposes that the major populations of the Japa-
nese archipelagoAinu, Ryukyu (Okinawan) and mainland Japaneseall originated from
two primary sources: the original hunter-gatherer inhabitants, who were already making
far has shown acceptable levels of inter-rater reliability (List 1974; Lomax 1976; Savage et al. 2012). Fully objective auto-
mated music analysis of audio recordings remains prohibitively difcult, particularly for non-western music (Tzanetakis
et al. 2007), and even then sample selection remains inherently subjective on some level.
The history of research on Ainu origins is inextricably linked with the political history of Japanese nationalism. Interested
readers can refer to book-length treatments by Hudson (1999) and Hudson et al. (2014).
distinctive Jomonpottery at least 15,000 years ago; and a later wave of Yayoiagricultur-
alist immigrants from mainland East Asia beginning 3000 years ago. The Ainu are pro-
posed to be relatively pure descendants of the Jomon, while the Ryukyu and mainland
Japanese are proposed to have experienced substantial admixture between the Yayoi
and Jomon, with the Ryukyu having relatively more Jomon ancestry and thus being
more similar to the Ainu than are the mainland Japanese, despite the geographic separ-
ation between the Ryukyu and the Ainu. Most genetic studies have supported the
general assertion of the dual structure model that the Ainu are most closely related to
the other populations of the Japanese archipelago, particularly the Ryukyu (Hudson
1999; Japanese Archipelago Human Population Genetics Consortium [JAHPGC] 2012;
Koganebuchi et al. 2012; Matsukusa et al. 2010; Omoto and Saitou 1997).
To our knowledge, no contemporary genetic studies support a purely circumpolar
origin of the Ainu. Hanihara in fact explicitly rejected a circumpolar contribution to
Ainu genetic history:
The view of Kiyono that the intermixture between Jomonese and north Asians gave rise to
the present-day Ainu is not acceptable because Ainu show almost none of the north Asian
characteristics found in the people of the extremely cold north. (Hanihara 1991: 22)
However, some genetic studies have in fact suggested some degree of northern inuence on
the Ainu (Adachi et al. 2011; Sato et al. 2009; Tajima et al. 2004). This evidence of northern
inuence has yet to be formally integrated into an alternative to the dual structure model.
With regard to languages, Greenberg (200002) proposed that the Ainu language is
most closely related to Japanese and Korean. This proposal is less widely accepted than
the dual structure model, but Ainu is generally viewed as not closely related to circumpolar
languages. The most widely accepted classication of Ainu is as a linguistic isolate (Lewis
Figure 1. Photograph of two Ainu musicians, Ogasawara Sayo (left) and Ehara Utae (right), performing
rekuhkara throat-rasping games. Source: Patrick E. Savage, 2013; still image from a video available
2009) that may be descended from one spoken by the Jomon-era hunter-gatherers
(Hudson 1999).
Thus, if the characterisation of Ainu music as being predominantly circumpolar is
correct, this would imply either that geneticists and linguists have underestimated the
degree of circumpolar inuence on the Ainu, or that the diffusion of music between
Siberia and the Ainu occurred on a large scale, mostly independently of human
If incorrect, it could imply that patterns of Ainu musical diversity do in fact
match those of genetic and linguistic diversity. In either case, this would have important
theoretical implications for understanding the history of the Japanese archipelago. Our
goal was to quantitatively re-examine Ainu musical diversity by analysing musical
samples from Ainu and surrounding East Asian and circumpolar populations. By using
new methods, we aimed to estimate the relative frequencies of different song-types
within populations and to compare these with published genetic data.
Song sample
To better understand Ainu musical diversity from a comparative perspective, we assembled
recordings of 680 traditional folk songs from 35 populations.
Table 1 presents the popu-
lation names, number of songs per population and recording source for each population.
Although the current study focuses primarily on music, the musical sample was selected
as part of a larger project with the eventual aim of directly comparing patterns of musical
and genetic diversity in populations relevant to understanding the history of the Japanese
archipelago. Thus, our primary goal was to nd as broad a sample as possible of traditional
songs from many populations that also had matching genetic data already published.
Rather than using many different smaller recording projects carried out by different
groups at different times with different goals and varying quality, we aimed to use a
small number of large-scale, high-quality recording projects to make the sample as amen-
able as possible to comparison. There were two recording projects that clearly provided the
most comprehensive coverage of the Japanese archipelago and circumpolar populations.
Both projects were coordinated across large geographic regions, including all traditional
folk song genres from a number of different populations.
The rst was a massive nation-
wide survey begun by Nippon HōsōKyōkai (NHK) in the 1940s and continuing through
the 1970s involving leading ethnomusicologists, such as Machida Kashō, Koizumi Fumio
and Tanimoto Kazuyuki (NHK 194494,1951,1965).
This included songs from the
Ryukyu Islands, mainland Japan, Hokkaido Ainu and Sakhalin Ainu who had recently
been relocated to Hokkaido from Sakhalin Island following its annexation by the USSR
at the end of World War II.
The second was a series of 11 CDs recorded between
1992 and 2011 by Henri Lecomte (19952012) featuring the music of Siberias indigenous
Convergent evolution is also a possibilitycold environments might conceivably lead to the independent adoption of
similar singing styles.
The details of Lomaxs Cantometric sample were never made publicly available, so it was not possible to directly re-analyse
his data.
These projects also included some purely instrumental tracks, but these were excluded because our classication schemes
were primarily designed to classify and compare only the vocal part.
Throughout this article, we use the standard surname-given name order for Japanese names (including Ainu).
Sakhalins Japanese name is Karafuto. Nattiez (1983,1999) refers to Sakhalin Ainu as Kraft Ainu. Ainu also formerly lived in
the north of the mainland Japanese island of Honshu, and in the Kurile (Chishima) Islands, Kamchatka peninsula and Amur
valley in Siberia, but historical information about them is scarce and none of their music was recorded.
peoples. Because these two recording projects did not include music of three populations
Korean, Saami and Inuitthat were crucial to evaluating proposals of Japanese archi-
pelago musical and genetic histories, we supplemented our sample with the most compre-
hensive recordings available for these populations (see Table 1 for detailed references).
minimise the number of additional recording sources included, we decided at this stage to
refrain from including additional recordings from other populations more broadly rel-
evant to understanding the regional history, such as those in China, Southeast Asia or
northwestern parts of North America.
To reduce these samples to manageable sizes for manual comparison, we followed the
sampling and analysis framework described in detail by Savage and Brown (2013,2014; see
Table 1. Breakdown of the 680-song sample, listing the number of songs, geographic region and
recording source for each of the 35 populations.
Region Population Number of songs Recording source
Ainu Ainu (Hokkaido) 30 NHK (Japan)
Ainu Ainu (Sakhalin) 30 NHK (Japan)
East Asian Japanese (Aomori) 30 NHK (Japan)
East Asian Japanese (Saga) 30 NHK (Japan)
East Asian Japanese (Shizuoka) 30 NHK (Japan)
East Asian Japanese (Tokushima) 30 NHK (Japan)
East Asian Ryukyu (Okinawa) 30 NHK (Japan)
East Asian Ryukyu (Yaeyama) 30 NHK (Japan)
East Asian Korean (Gangwon) 30 MBC (Korea)
East Asian Korean (South Jeolla) 30 MBC (Korea)
Circumpolar Buryat 30 Buda (Siberia)
Circumpolar Koryak 30 Buda (Siberia)
Circumpolar Evenk 28 Buda (Siberia)
Circumpolar Altai 26 Buda (Siberia)
Circumpolar Nivkh 19 Buda (Siberia)
Circumpolar Even 17 Buda (Siberia)
Circumpolar Nganasan 15 Buda (Siberia)
Circumpolar Chukchi 14 Buda (Siberia)
Circumpolar Nanai 13 Buda (Siberia)
Circumpolar Nenets 12 Buda (Siberia)
Circumpolar Selkup 12 Buda (Siberia)
Circumpolar Yukaghir 12 Buda (Siberia)
Circumpolar Khanty 11 Buda (Siberia)
Circumpolar Sakha/Yakut 8 Buda (Siberia)
Circumpolar Mansi 7 Buda (Siberia)
Circumpolar Ujlta/Orok 6 Buda (Siberia)
Circumpolar Udege 4 Buda (Siberia)
Circumpolar Ulch 4 Buda (Siberia)
Circumpolar Oroc 3 Buda (Siberia)
Circumpolar Saami 30 SF (Norway)
Circumpolar Inuit (Iglulik, Canada) 30 MFV (Canada)
Circumpolar Inuit (North Greenland) 20 ULO (Greenland)
Circumpolar Inuit (East Greenland) 12 ULO (Greenland)
Circumpolar Inuit (South Greenland) 9 ULO (Greenland)
Circumpolar Inuit (West Greenland) 8 ULO (Greenland)
Total 680
NHK = Nippon HōsōKyōkai (194494,1951,1965); MBC = Munhwa Broadcasting Corporation (Choi 2000); Buda = Buda
Musique (Lecomte 19952012); SF = Smithsonian Folkways (Laade 1956); MFV = Museum für Völkerkunde (Nattiez
and Conlon 1993); ULO = ULO (Hauser 1992).
Most recordings are available for purchase in digital form, with the exception of the NHK Ryukyu and Ainu recordings. The
Hokkaido Ainu recordings (NHK 1965) are available as records at many Japanese libraries and can sometimes be purchased
online secondhand. To our knowledge, the only remaining publicly accessible copies of the Sakhalin Ainu recordings (NHK
1951) are at the National Diet Library in Tokyo (call numbers Cas-671Cas-673). The unpublished Ryukyu recordings were
kindly provided by Uemura Yukio and Kaneshiro Atsumi and selected by Matt Gillan.
Table 2. Overview of the three-step cantogroup analysis framework (adapted from Savage and Brown
Level of
analysis Methods Description
1. Classication of songs
(Supplemental Data
CantoCore (Savage et al.
2012), Cantometrics
(Lomax 1976)
Classify each song according to 41
features of song structure and
performance style
2. Clustering of songs into
cantogroups (stylistic song-
types) (Figure 2)
All songs of a
Cluster analysis (Hartigan
and Wong 1979)
Assign songs to song-type clusters
based on their musical similarities,
irrespective of geography
3. Mapping of cantogroup
frequencies (Figure 3)
Pie charts Map frequencies of all cantogroups in
each populations repertoire onto
geographic regions
Table 3. Breakdown of the 41 CantoCore (Savage et al. 2012) and Cantometrics (Lomax 1976)
classication features used for the analysis.
Feature Domain Original line number
1. Metre Rhythm 1 (CantoCore)
2. Number of beats Rhythm 2 (CantoCore)
3. Beat sub-division Rhythm 3 (CantoCore)
4. Number of sub-beats Rhythm 4 (CantoCore)
5. Syncopation Rhythm 5 (CantoCore)
6. Motivic redundancy Rhythm 6 (CantoCore)
7. Durational variability Rhythm 7 (CantoCore)
8. Tonality Pitch 8 (CantoCore)
9. Mode Pitch 9 (CantoCore)
10. Number of pitch classes Pitch 10 (CantoCore)
11. Hemitonicity Pitch 11 (CantoCore)
12. Melodic interval size Pitch 12 (CantoCore)
13. Melodic range Pitch 13 (CantoCore)
14. Melodic contour Pitch 14 (CantoCore)
15. Melisma Text 15 (CantoCore)
16. Vocables Text 16 (CantoCore)
17. Number of vocal parts Texture 17 (CantoCore)
18. Rhythmic texture Texture 18 (CantoCore)
19. Harmonic texture Texture 19 (CantoCore)
20. Relative motion Texture 20 (CantoCore)
21. Phrase repetition Form 21 (CantoCore)
22. Phrase length Form 22 (CantoCore)
23. Phrase symmetry Form 23 (CantoCore)
24. Solo/group arrangement Form 24 (CantoCore)
25. Responsorial arrangement Form 25 (CantoCore)
26. Phrase overlap Form 26 (CantoCore)
27. Tonal blend Blend 5 (Cantometrics)
28. Rhythmic blend Blend 6 (Cantometrics)
29. Tempo Dynamics 24 (Cantometrics)
30. Volume Dynamics 25 (Cantometrics)
31. Rubato Dynamics 26 (Cantometrics)
32. Register Dynamics 32 (Cantometrics)
33. Embellishment Ornamentation 23 (Cantometrics)
34. Glissando Ornamentation 28 (Cantometrics)
35. Tremolo Ornamentation 30 (Cantometrics)
36. Glottal shake Ornamentation 31 (Cantometrics)
37. Vocal width Timbre 33 (Cantometrics)
38. Nasalisation Timbre 34 (Cantometrics)
39. Raspiness Timbre 35 (Cantometrics)
40. Accent Timbre 36 (Cantometrics)
41. Enunciation Timbre 37 (Cantometrics)
Tables 13). For populations represented by fewer than 30 songs, all available traditional
adult songs were used (including both solo and group songs). For those populations with
more than 30 songs available, only 30 randomly-selected songs were used. For complete-
ness, our analyses included all eligible populations, even those with small sample sizes of
less than ten songs or those whose circumpolarstatus was questionable (i.e., Buryat,
Altai). Follow-up analyses excluding these populations did not affect any of our
Classication of songs
P.E.S. analysed each song individually using 41 different classication characters (see
Table 3). These included 26 structural features from CantoCore (e.g., metre, melodic
range, phrase repetition; Savage et al. 2012) and 15 performance features from Canto-
metrics (e.g., tempo, embellishment, vocal rasp; Lomax 1976; Lomax and Grauer
Each character was classied manually by ear from a range of three to six different
character-states. For example, melodic range was classied as either small (less than 750
cents), medium (7501250 cents) or large (greater than 1250 cents). Full details of classi-
cation criteria and inter-rater reliability values for each feature can be found in Lomax
(1976) and Savage et al. (2012).
Clustering of songs into cantogroups (stylistic song-types)
Average similarities between songs based on these 41 classication characters were calcu-
lated using the method of Rzeszutek, Savage and Brown (2012), after which k-means
cluster analysis (Hartigan and Wong 1979) was used to group the songs into cantogroups.
Mapping of cantogroup frequencies
The frequencies of each cantogroup within each populations repertoire were mapped
onto geographic space using pie charts.
Song-level analysis and clustering
To demonstrate the musical relationships among the 680 songs used in our sample, we
used a multidimensional scaling plot, as shown in Figure 2. Cluster analysis identied
ve cantogroups (coloured in greyscale) as the optimal number of clusters. The high fre-
quencies of East Asian songs in cantogroups E1 and E2, circumpolar songs in cantogroups
C1 and C2, and Ainu songs in cantogroup A suggest that there are important musical
The one methodological difference from Savage and Brown (2014) was the inclusion of 15 Cantometric performance fea-
tures in addition to the 26 CantoCore structural features, as recommended by those authors. These 41 features were the
same ones used for musicgene comparison by Brown et al. (2014). The 15 Cantometric characters were chosen to avoid
structural characters that overlapped with CantoCore and instrumental characters that were unreliable to code (Savage
et al. 2012). Full metadata for all songs, including classications and discographic information, is listed in the Supplemen-
tal Data spreadsheet.
differences between these geographic regions. Thus, for ease of interpretation, we have
given each cantogroup a label according to its predominant geographic distribution (see
Figure 3). However, the cluster analysis does not incorporate any a priori information
about cultural or geographic afliations.
Table 4 presents a summary of some of the notable musical features that are most
regionally distinctive for each cantogroup, colour-coded as in Figure 2. To provide con-
crete examples of these abstract descriptions, we have included links to brief 30-second
excerpts from 11 recordings of prototypical Ainu and non-Ainu songs from each can-
togroup (these songs are highlighted with arrows in Figure 2).
It is impossible to give comprehensive descriptions of these songs, since cantogroups
are dened based on overall family resemblance and not by specic features. However,
the following trends emerge:
(1) East Asian. Cantogroup E1 contains many East Asian ornamented solo songs with
loose or non-existent metre, while cantogroup E2 contains many East Asian metric,
group-accompanied songs. These correspond to Koizumis(2009)
between two archetypal Japanese folksong types: Oiwake(ornamented, solo, a-
metric) and Yagibushi(metric, group-accompanied). In terms of Lomaxs global
Figure 2. Multidimensional-scaling visualisation of the cluster analysis of 680 traditional folk songs
from 35 populations in and around Japan. Each point represents a single song, with different
shapes representing different source regions (see legend to left). Songs that are closer together in
the plot have greater average musical similarityas based on their CantoCore/Cantometric codings
than songs that are further apart. Songs are colour-coded in greyscale according to their member-
ship in one of ve cantogroups (stylistic song-types), as identied through cluster analysis and shown in
the legend. Two or three example songs from each cantogroup are highlighted with arrows. These
songs are described in detail in Table 4, along with notable musical features of each cantogroup.
The 30-second audio clips and metadata for these examples can be found in the Supplemental Data.
Originally written in 196263 and rst published in 1984.
taxonomy, E1 and E2 could be thought of as contrasting sub-types within his Pan-
Eurasian Old High Culturetype, where he classied Korean, Ryukyu and mainland
Japanese music (Lomax 1968,1980).
(2) Circumpolar. Cantogroup C1 contains many solo, irregular, harsh-voiced circumpolar
types, while cantogroup C2 mostly contains a slightly more regular, vocable-heavy,
often drum-accompanied circumpolar type. These seem to correspond to Lomaxs
distinction between Siberianand Circum-Pacictypes, respectively.
Figure 3. A musical map of the 35 populations based on cantogroup frequencies. Relative frequencies
of all ve cantogroups (colour-coded in greyscale as in Figure 2) are shown separately for each of the 35
populations using pie charts. See Figure 2 for cantogroup identication. Population locations are based
on approximate recording locations.
Table 4. Simplied descriptions of notable musical features for each of the ve cantogroups.
Cantogroup Notable musical features Examples
(East Asian 1)
Irregular: metre loose/absent, solo a cappella, large melodic
range, pentatonic scales, long phrases, melismatic, nasal
E1-1: Yaykatekara (Ainu love song;
No. 33)
E1-2: Jongarabushi (Japanese
entertainment song; No. 75)
(East Asian 2)
Semi-regular: iso-metric, unison/call-and-response, large
melodic range, often chordophone accompaniment
E2-1: Sucho choy choy na (Ainu
seed-planting song; No. 17)
E2-2: Hatuma nakamuri (Ryukyu
sanshin song; No. 236)
(Circumpolar 1)
Irregular: metre loose/absent, solo a cappella, small melodic
range, sparse scales, microtonal glides, glottal shake
C1-1: Hawki (Ainu epic; No. 35)
C1-2: Tilgur (Nivkh epic; No. 439)
(Circumpolar 2)
Semi-regular: regular beat (but not necessarily regular metre),
non-lexical vocables, often self-accompaniment on (shaman)
C2-1: Tusu-sinotca (Ainu shamanic
drum song; No. 50)
C2-2: Shamanic song with uŋtu
drum (Ulc; No. 418)
A(Ainu) Regular: iso-metric, group canon (ukouk)/call-and-response
(iekaye), sparse scales, short/repetitive phrases, often
accompanied by clapping
A-1: Upopo (Ainu group song; No. 3)
A-2: Rekuhkara (Ainu throat-game;
No. 44)
A-3: Knylju (Koryak lullaby; No. 378)
Cantogroups colour-coded in greyscale as in Figures 2 and 3. Examples of Ainu and non-Ainu songs from each cantogroup
correspond to the labelled points in Figure 2. Song number (No.) refers to the songs number in the Supplemental Data
spreadsheet; refer to this for additional metadata for these songs. Not all songs in each cantogroup necessarily contain all
of these features, since cantogroups are dened based on overall similarity rather than any particular features.
(3) Ainu. Cantogroup A contains many Ainu songs known as upopo
in Hokkaido (Song
A-1) and heciri in Sakhalin that often use a canonic texture known as ukouk, as well as
other highly regular, overlapping group songs. Cantogroup A seems to be most similar
to the regular, overlapping group songs of LomaxsAfrican Gatherertype, common
among central African Pygmies and southern African Khoisan Bushmen. Lomax
(1976: 38) also mentions this African Gatherer type as appearing among the Ainu,
although he classies Ainu music as a whole as part of a Siberiansong-style region.
However, the vocal timbre of cantogroup A is generally much harsher than Lomaxs
prototypical African Gatherer style and also includes songs that contain little or no
overlap, particularly songs that were near the outer borders of this cantogroup.
This cantogroup also includes the Ainu rekuhkara throat-game genre (Song A-2)
and a handful of other circumpolar throat-rasping duets, such as from the Koryak
(Song A-3) and Iglulik Inuit, a connection suggested by Nattiez (1983,1999), Tanimoto
(2000,2001a) and Keeling (2012). However, it does not include some of the other songs
in the sample containing throat-rasping, such as the Ainu tusu-sinotca shamanic drum
song (Song C21) and the Ulc shamanic drum song (Song C2-2) found in cantogroup
C2, or the animal-imitation songs of the Even and Chukchi found in cantogroup C1.
Although each cantogroup has certain features that are regionally distinctive (summar-
ised in Table 4), there appears to be surprisingly little blending of individual features
between cantogroups. For example, while the overlapping ukouk texture is the most con-
sistent and distinctively Ainu feature of cantogroup A, ukouk is rarely found in Ainu
music outside of cantogroup A. Other notable features of cantogroup A presented in
Table 4 that are most unique to the Ainu (e.g., sparse scales, short/repetitive phrases, clap-
ping) are also rare in Ainu music outside cantogroup A. Instead, it seems that the stylistic
features of songs may tend to be transmitted as discrete packets, such that an Ainu tusu-
sinotca shamanic drum song (Song C2-1) in cantogroup C2 includes not only the use of
shaman drums common in the circumpolar region, but also other typically circumpolar
features (e.g., throat-rasping, irregular metres). Likewise, an Ainu yaykatekara love song
(Song E1-1) in cantogroup E1 shares not only pentatonic scales common in mainland
Japan, but also other typically mainland Japanese features (e.g., long, melismatic
phrases). In the future, we will explore more complex quantitative techniques such as mul-
tiple correspondence analysis (Greenacre and Blasius 2006) in order to more precisely
determine the degree to which each individual musical feature contributes to creating
regionally distinctive styles.
Cross-cultural distribution of cantogroups
Figure 3 uses pie charts to map the geographic distribution of the ve cantogroups. These
charts conrm the impression from Figure 2 that the East Asian,Circumpolarand
Ainucantogroups are indeed largely restricted to their respective regions:
(1) East Asian. Cantogroups E1 and E2 were largely restricted to East Asia, where they
were on average the most common types across all mainland Japanese, Ryukyu and
Korean populations (although when breaking populations down by sub-region, the
The Ainu language was traditionally transmitted orally without writing, and transliterations have not been widely stan-
dardised. In this article we adopt Chibas(2008) usage.
Korean province of South Jeolla had slightly more of Circumpolarcantogroups C1
and C2). E1 and E2 were rare within the Ainu repertoire (7% across both Hokkaido
and Sakhalin Ainu). When they did appear, it was with genres known to have had
recent inuence from mainland Japan within the last few hundred years, such as
songs involving agriculture (introduced in the eighteenth century; Song E2-1) and
sinotca lyric songs (including yaykatekara love songs as in Song E1-1) that incorpor-
ate stylistic and/or linguistic elements from the mainland Japanese (Chiba 2008;NHK
1965; Tanimoto 2000).
(2) Circumpolar. Likewise, cantogroups C1 and C2 were found predominantly among cir-
cumpolar populations in Siberia and the Arctic, where they are the most common
song-types in each populations repertoire. These types occur fairly frequently
within the Ainu repertoire (42% across both Hokkaido and Sakhalin Ainu), particu-
larly in that of the Sakhalin Ainu, where they actually outnumber the Ainucan-
togroup A by 53% to 43%, respectively. The only other non-circumpolar
population where these types are most common is in the Korean province of South
Jeolla. But when considering Korean and Ainu repertoires as a whole across all
sub-regions, the East Asianand Ainucantogroups, respectively, are most common.
Some of the Circumpolar types in the Ainu repertoire have known historical
connections to Siberia. For example, the tusu-sinotca shamanic drum songs (Song
C2-1) are considered to be related to Siberian shamanic drum song traditions
such as those of the Ulc (Song C2-2; Tanimoto 2000). Others, such as the epics
known as hawki in Sakhalin (Song C1-1) and yukara in Hokkaido, share stylistic
similarities with Siberian epics, such as the Nivkh tilgur (Song C1-2) and the
Sakha/Yakut olunkho.
(3) Ainu. Cantogroup A was the most common cantogroup within the Ainu repertoire,
accounting for 52% of all Ainu songs across both Hokkaido and Sakhalin Ainu. It
reached its highest frequency (60%) among the Hokkaido Ainu, followed by the
Sakhalin Ainu (43%). In all other populations, it was either absent or present at
very low frequencies, never more than 25%. The non-Ainu songs in this cantogroup
were almost always songs found near the border with cantogroup C2, except for the
cases of the throat-rasping duets shared with Koryak and the Iglulik Inuit mentioned
in the previous section.
Musicgene comparison
To compare patterns of musical and genetic diversity for the Ainu, we compared the fre-
quencies of the discussed Ainu-specic, circumpolar and East Asian cantogroups with fre-
quencies of mtDNA haplotypes (unique DNA sequences) published by Tajima et al.
(2004)(Table 5).
This comparison showed intriguing parallels between music and
genes, with both Ainu songs and Ainu mtDNA containing a majority of unique Ainu
types (52% vs. 51%, respectively), followed by circumpolar types (41% vs. 28%) and
then by East Asian types (7% vs. 18%). A small amount of Ainu mtDNA (4%) was also
Excluding these songs does not substantively affect any of our major conclusions regarding the frequencies of circum-
polar or Ainu cantogroups.
We plan to offer a more comprehensive musical and genetic comparison including all populations (cf. Brown et al. 2014)
and/or other genetic markers such as Y-chromosome DNA and autosomal DNA in a future publication.
shared with Southeast Asia, but we were unable to make the matching comparison with
our present musical sample.
Atriple structuremodel for Japanese archipelago cultural history
Our analyses provide the rst quantitative evidence that the Ainu musical repertoire con-
tains nearly equal parts that are uniquely Ainu versus shared with surrounding popu-
lations, and that sharing has been much more extensive with circumpolar populations
than with the mainland Japanese. These ndings do not contradict qualitative studies
(e.g., Keeling 2012; Malm 1967; Nattiez 1983,1999; Tanimoto 2000,2001a) that have
argued for a circumpolar component to the Ainu repertoire. In fact, our ndings comp-
lement these studies by providing quantitative support for such a component. However,
our ndings also extend previous studies by demonstrating that the circumpolar com-
ponent is only one part of the broader Ainu repertoire, thus helping to unify the apparent
discrepancy between Ainu musical and genetic histories. Such an integration of quantitat-
ive and qualitative approaches is essential for improving our understanding of music and
human history, and may help to resolve other controversies, such as Grauers(2006,2011)
proposal of an ancient out-of-Africa musical expansion.
By applying improved methods that allow us to quantify the relative frequencies of
different musical types within and between populations, we in fact nd that there is less
contradiction between the musical and genetic evidence than expected based on
Lomaxs(1980) musical analysis or Haniharas(1991) original dual structure model. In
fact, the presence of high frequencies of unique cantogroups (50%), medium frequencies
of circumpolar cantogroups (40%) and low frequencies of East Asian cantogroups
(10%) are similar to corresponding estimates of mtDNA diversity (50%, 30% and
20%, respectively) obtained by Tajima et al. (2004).
It seems logical that these parallels
could have arisen through the joint migration of people and their songs (including
Table 5. Comparison of Ainu musical and genetic diversity.
Putative source region Ainu songs (%) Ainu mtDNA (%)
Ainu 52 51
Circumpolar 42 28
East Asian 7 18
(Southeast Asian) (N/A) (4)
The percentages of songs derived from Ainu, circumpolar and East Asian sources are estimated based on the combined
percentages of cantogroups A (Ainu), C1 and C2 (Circumpolar) and E1 and E2 (East Asian) averaged across both the Hok-
kaido and Sakhalin Ainu repertoires. The percentage of mtDNA derived from these sources and from Southeast Asia is
taken from the published analysis of Tajima et al. (2004). The percentages add up to 101% rather than 100% owing to
rounding error.
Tajima et al. (2004) also report evidence for some (13%) circumpolar contribution to the Ainu paternally-inherited
Y-chromosome gene pool, but this evidence seems more equivocal, given the similar level (14%) of this haplogroup
(C-M217*) among Han Taiwanese found in the same study. Other genetic studies are still harder to interpret. Studies
supporting circumpolar inuence have shown that mtDNA haplogroup Y is the most common haplogroup among the
Ainu (20%), Nivkh (66%) and ancient Okhosk (43%), but not among ancient Jomon (0%) or contemporary mainland Japa-
nese (1%) (Adachi et al. 2011; Sato et al. 2009). Many of the studies supporting the dual structure model have not
included likely sources of circumpolar inuences such as the Nivkh or Okhotsk in the sample, and were thus unable
to evaluate the degree of potential circumpolar inuence (e.g., JAHPGC 2012; Koganebuchi et al. 2012; Matsukusa
et al. 2010; Omoto and Saitou 1997).
through intermarriage), but other explanations are also plausible, such as musical borrow-
ing and genetic isolation by distancetaking place independently due to geographic proxi-
mity (Savage and Brown 2013; Wright 1943). The current study has focused primarily on
characterising the musical patterns, with only limited direct comparison with genetic data.
In the future, we hope to more directly and comprehensively compare the regional pat-
terns of musical and genetic diversity in order to better understand how and why these
similarities and differences arose.
Many scholars have long felt that Haniharas dual structure modelwhile extremely
valuablewas overly simplistic in the case of the Ainu. Most historians believe that the
Ainu only emerged as a distinct culture around 1200 AD from the merging of the
Okhotsk culture from Siberia with the indigenous, Jomon-descended Satsumon culture
in Hokkaido (Hudson 1999).
Recently, a major consortium of population geneticists
published a diagram of Japanese archipelago population history that extends this idea
(JAHGPC 2012: 793). In addition to interactions that are recognised by the dual structure
model between the indigenous Jomon and mainland East Asian Yayoi and their descen-
dants, this diagram also included interactions with the Okhotsk and their descendants. We
propose terming this scenario the triple structuremodel (see Figure 4) to recognise the
Figure 4. Schematic diagram of the triple structuremodel of Japanese archipelago cultural history
(adapted from JAHPGC 2012). Approximate source-regions and dates of origin (ya = years ago) of
three major founding cultures are inferred from archaeological records, using the names given to
those archaeological cultures. The direction by which the creators of the Jomon culture entered the
Japanese archipelago is not shown, because there are disputes over this point (Omoto and Saitou
1997). Approximate historical distributions of three major contemporary ethnolinguistic groups in
Japan are shown to the right.
For instance, the different uniparental markers (matrilineally inherited mtDNA and patrilineally inherited Y-chromosome
DNA) may correlate better with songs performed by the respective sexes, just as sex-biased migration can affect the geo-
graphic distribution of genes (Oota et al. 2001) and genelanguage correlations (Forster and Renfrew 2011). Music-
making worldwide tends to be highly gendered (Koskoff 1987), and the circumpolar tradition of unilineal transmission
of personal songs from mother to daughter and from father to son (Tanimoto 2001b) in particular would be expected to
amplify sex-biased correlations between musical and genetic lineages.
Satsumon culture was descended directly from the Jomon culture via the epi-Jomon culture.
importance of Haniharas dual structure model as well as the need to acknowledge the cir-
cumpolar inuence on the Ainu that Hanihara initially rejected. This model does not
imply that each source population only contributed to one historical period; rather,
there have been multiple ongoing interactions. For example, East Asianinuences on
the Ainu may ultimately have their roots in the movement of Yayoi agriculturalists
from mainland Asia 3000 years ago, but most East Asian inuences did not become
incorporated by the Ainu until the assimilation policies introduced by the Meiji govern-
ment in the late nineteenth century. Likewise, circumpolar inuences that are pronounced
among the Sakhalin Ainu but not the Hokkaido Ainu, such as tusu-sinotca shamanic drum
songs, are likely to have been introduced not by the movement of the Okhotsk culture to
Hokkaido 1500 years ago but by more recent contact with the Nivkh and other inhabi-
tants of Sakhalin after the Ainu expanded there from Hokkaido 600 years ago (Ohyi
It is crucial to remember that all of these populations today have, and have had in the
past, a great deal of internal diversity which developed over the course of tens of thousands
of years of evolution and contact. Thus, even as the ancestors of the Ainu were experien-
cing complex cultural transitions and interactions, similar but distinct processesinclud-
ing continuing contact within the Japanese islands and with mainland Asiawere taking
place among the ancestors of the modern groups now known as Ryukyuand Japanese.
Indeed, these processes never endedif anything, they have intensied with globalisation.
The point of models is not to capture everything that happened historically, but to
provide tractable hypotheses that can explain a substantial portion of the observed data.
Simple models thus provide a way of helping us make sense of what are, in reality,
highly complex phenomena. In formulating models it is good practice to start with
simple models and then add in greater complexity, if it is justied by the data. While
the triple structure model seems to explain at least some of the genetic and musical
data better than the dual structure model, the true history was undoubtedly even more
complex (Sasaki 1997). Thus, with additional data and analyses there may very well be
further extensions needed to the triple structure model, just as Hanihara (1994) encour-
aged modications and extensions of his dual structure model.
Degrees of circumpolar-ness
While the musical and genetic similarities are intriguing, they are not identical, and there
is substantial variation even within the Ainu. In particular, the music of the Sakhalin Ainu
seems to have received particularly strong circumpolar inuence relative to the Hokkaido
a point also described in detail by Tanimoto (2000). He noted that a number of
circumpolar featuresincluding the kaco shaman frame drum used in tusu-sinotca sha-
manic drum songs, the tonkori ve-string zither and associated tonkori-heciri song
genre, and the throat-rasping technique used in tusu-sinotca and in rekuhkara throat-
game genreswere all traditionally found only in Sakhalin, not Hokkaido. Our analysis
conrmed all of these points within our sample.
It may be difcult to investigate regional genetic differences between Hokkaido and Sakhalin Ainu as, unlike for music, no
separate Sakhalin Ainu genetic sample has been collected (although some Sakhalin Ainu may have been included in the
major existing Ainu sample, which was collected in Biratori, Hokkaido in the early 1980s; JAHPGC 2012).
Some scholars have placed a strong emphasis on the rekuhkara throat-game genre as
being the key link between the Ainu and circumpolar cultures as far removed as the Cana-
dian Inuit (Malm 1967; Nattiez 1983,1999). In particular, Nattiez (1999) has inferred sha-
manic symbolism in the throat-rasping technique that connects the rekuhkara with other
circumpolar genres, including the pičeynen dance of the Chukchi and the katajjaq throat-
games of the Eastern Canadian Inuit. However, our analysis shows that, across all 41
classication features used, rekuhkara are generally quite similar to non-rekuhkara
Ainu songs, such as upopo/heciri, and do not always necessarily use throat rasping (e.g.,
Song A-2). They may thus be thought of as a special heciri variant, as proposed by
Chiri (1955) and Chiba (2008). If anything, it seems that the styles most similar to rekuh-
kara, such as the knylju lullaby of the neighbouring Koryak (Song A-3), could well have
originated from the Ainu rekuhkara and been transmitted to other circumpolar traditions,
rather than vice versa.
In any case, we believe that the seldom-discussed tusu-sinotca sha-
manic drum song genre is the strongest candidate for an Ainu genre of circumpolar origin,
because it contains not only throat-rasping and an unambiguous shamanic function, but
also the distinctive shaman frame drum that is perhaps the clearest musical circumpolar
marker (Tanimoto 2001b), being ubiquitous throughout Siberia and stretching as far west
as the Saami in Arctic Europe and as far east as the Greenland Inuit.
For the purposes of this study, we have not greatly explored the internal diversity of
repertoires other than that of the Ainu, but there is a great deal of it. In particular,
there seems to be a rough northeast/southwest geographic/cultural split in which
Altaic-speaking and Uralic-speaking populations from southern/western Siberia and
Arctic Europe tend to have higher frequencies of cantogroup C2, while northern/
eastern Siberian and Arctic American populations tend to have higher frequencies of can-
togroup C1. Given the massive size and sparse musicological literature in English on this
and the importance of Siberia to debates regarding the peopling of the Americas
(Cavalli-Sforza, Menozzi and Piazza 1994; Greenberg 1987; Reich et al. 2012), these
musical and extra-musical relationships deserve further investigation.
Sampling limitations
Savage and Brown (2013,2014) discussed in detail a number of caveats regarding Canto-
metrics, the cantogroup methodological approach and comparative musicology in general.
These issues include sampling, denitions of populations, choice of classication features,
modelling musical evolution and musical time-depth, assignment of songs to discrete can-
togroup clusters and the need for independent comparisons of multiple different lines of
evidence, such as genes and languages. Each of these issues applies to the current study to
some degree, but the issue most directly relevant to our analysis is that of comparability of
samples collected at different times and places by different people with different purposes.
We tried to limit sampling to a small number of recording projects with a similar
general purpose (i.e., to present an overview of the different traditional folk songs of a
It also seems plausible that the different cultures could have independently invented the genres by combining elements
of throat-rasping with a regular, canonic form.
No songs in our Saami sample included drumming because shaman drums were banned there by Christian missionaries
(Laade 1956). However, Fernandez and Jocelyne (1984) include one recording of a drum-accompanied Saami joik.
There is a substantial non-English literature (e.g., Dobžanskaja 2002; Lecomte 2012), particularly in Russian.
given population). However, projects like NHKs survey of folk songs throughout the Japa-
nese archipelago took place over a span of several decades and were carried out by differ-
ent teams of specialists in different places. Each had their own ideas about what they
wanted to record, and each worked with communities who had different ideas about
what they wanted to be recorded. Some of the songs recorded then are no longer per-
formed today, and some songs performed today did not exist at the time of these record-
ings. Needless to say, these problems become compounded when one then attempts to
further compare these recordings with others produced under still more different con-
ditions, especially when substantial musical changes have been documented in all of
these cultures during this time period (Hughes 2008;Kōchi 2001; Sheykin 2001; Tokita
and Hughes 2008). Similar (although generally less severe) issues, such as the effects of
urban migration and intermarriage, could be pointed out for the genetic studies against
which we are comparing our musical data.
While we have attempted to carry out as controlled a comparison as possible by limiting
our sample to a small number of large-scale, high-quality recording projects, the results
nonetheless need to be viewed with an awareness of the inevitable limitations. Still, the
results are unlikely to be merely artefacts of the sampling or analysis procedures, but
instead largely capture broad but meaningful patterns of diversity, even if they can
never fully account for all of the ner details that inevitably remain overlooked.
In particular, we believe that our primary conclusion regarding the triple structure of
Ainu musical diversity is robust to these issues. Although there has been a great deal of
change in the Ainu repertoire since these recordings were published in the 1950s and
1960s, most Ainu performances continue to contain diverse styles that still seem to
broadly fall within the same Ainu, circumpolar and East Asian stylistic types we identied,
and the relative frequencies of the types seem to remain fairly similar. Furthermore, we
know of no evidence from any time period that the Ainu-specic style we identied
was ever common in any of the non-Ainu populations included. However, future
studies of contemporary Ainu music performances will be important to formally assess
the degree of stability of these patterns and understand the mechanisms of preservation
and change.
We have shown quantitatively for the rst time that Ainu music is not simply Siberianor
circumpolarin style, but that it contains a majority of unique Ainu types not found at
high frequencies in the repertoires of other East Asian or circumpolar populations. We
have also shown that the relative frequencies of Ainu, circumpolar and East Asian
musical types parallel the frequencies of mtDNA haplotypes derived from these geo-
graphic sources.
These results imply that the Ainu are not simply modern-day descendants of the
ancient Jomon people, nor is their culture simply Siberianor circumpolar. Rather,
they have actively incorporated inuences from multiple directions while continuing to
maintain and adapt their own diverse cultural and genetic heritage. We hope that a
more nuanced triple structuremodel of Japanese archipelago cultural history that
acknowledges this diversity will help to provide more opportunities to promote, preserve
and rediscover both the shared and unique aspects of Ainu heritage.
We thank the performers and collectors who made this work possible. We would also like to thank
Chiba Nobuhiko, Kōchi Rie, Kitahara Jirōta Mokottunas, Richard Keeling, Henri Lecomte, Ehara
Utae, Ehara Rumiko, Ogasawara Sayo, Tateshita Naoko, Yamamoto Fumitoshi, Taniguchi Shigeru,
Taciana Cahielnik, Shimozaki Kumi, Yamashita Masami, Sakai Emi, Hugh DeFerranti, Uemura
Yukio, Tsukahara Yasuko, Ōsumi Kinya and the entire Haponetay staff for extensive advice and
support whileconducting this research and for helpful comments on earlierversions of this manuscript.
We would like to thank Kaneshiro Atsumi and Uemura Yukio for allowing us access to the unpublished
digitisedRyukyu recordings. Theaccompanying 30-second sample audio clips fromSiberia were kindly
authorised without charge by Gilles Fruchaux and Henri Lecomte at Buda Musique.
This work was supported by a Japanese MEXT scholarship to Patrick E. Savage, and by grants to
Steven Brown from the Social Sciences and Humanities Research Council of Canada and from
McMaster University. The authors declare no competing nancial interests.
Supplemental data
Supplemental data for this article, including raw data, metadata and 30-second audio
examples, can be accessed at
Patrick E. Savage is a PhD candidate in Musicology at the Tokyo University of the Arts,
Japan. Hiromi Matsumae conducted this research as a postdoctoral researcher in the
Department of Anatomy at the Kitasato University School of Medicine, Kanagawa,
Japan (currently at the Institute of Evolutionary Biology and Environmental Studies at
the University of Zurich, Switzerland). Hiroki Oota is an Associate Professor in the
Department of Anatomy at the Kitasato University School of Medicine, Kanagawa,
Japan. Mark Stoneking directs the Human Population History Group in the Department
of Evolutionary Genetics at the Max Planck Institute for Evolutionary Anthropology,
Leipzig, Germany, and is Honorary Professor of Biological Anthropology at the University
of Leipzig. Thomas E. Currie is a lecturer in Cultural Evolution in the Centre for Ecology
& Conservation at the University of Exeter, Penryn Campus, Cornwall, UK. Atsushi
Tajima is a Professor in the Department of Bioinformatics and Genomics at Kanazawa
University, Ishikawa, Japan. Matt Gillan is Senior Associate Professor of Music at Inter-
national Christian University, Tokyo, Japan. Steven Brown is the director of the Neu-
roArts Lab and an Associate Professor in the Department of Psychology, Neuroscience
& Behaviour at McMaster University, Hamilton, Ontario, Canada. Correspondence to:
Patrick E. Savage, Department of Musicology, Tokyo University of the Arts, 12-8 Ueno
Kouen, Taito-ku, Tokyo 110-8714, Japan. Email:
Patrick E. Savage
Adachi, Noboru, Ken-ichi Shinoda, Kazuo Umetsu, Takashi Kitano, Hirofumi Matsumura, Ryuzo
Fujiyama, Junmei Sawada and Masashi Tanaka. 2011. Mitochondrial DNA Analysis of
Hokkaido Jomon Skeletons: Remnants of Archaic Maternal Lineages at the Southwestern Edge of
Former Beringia.American Journal of Physical Anthropology 146(3): 346360.
Averill, Gage. 2003. Cantometrics and Cultural Equity: The Academic Years. In Alan Lomax:
Selected Writings, 19341997, edited by Ronald D. Cohen, 23348. New York: Routledge.
Bickmore, Albert S. 1868. The Ainos, or Hairy Men of Saghalin and the Kurile Islands.American
Journal of Science and Arts 45(135): 361377.
Brown, Steven, Patrick E. Savage, Albert Min-Shan Ko, Mark Stoneking, Ying-Chin Ko, Jun-Hun
Loo and Jean A. Trejaut. 2014. Correlations in the Population Structure of Music, Genes, and
Language.Proceedings of the Royal Society B: Biological Sciences 281(20132072): 17. doi:10.
Callaway, Ewen. 2007. Music is in Our Genes.Nature News.doi:10.1038/news.2007.359
Cavalli-Sforza, L. Luca, Paolo Menozzi and Alberto Piazza. 1994. The History and Geography of
Human Genes. Princeton: Princeton University Press.
Chiba, Nobuhiko. 2008. The Music of the Ainu. In The Ashgate Research Companion to Japanese
Music, edited by Alison McQueen Tokita and David W. Hughes, 32344. Farnham, UK: Ashgate.
Chiri, Mashiho. 1955. Ainu Bungaku [Ainu Literature]. Tokyo: Gengensha.
Choi, Sang-il, ed. 2000. Hanguk Minyo Taejon [Korean Folk Song Anthology; 12 CDs]. Seoul: MBC
(Munhwa Broadcasting Co).
Diamond, Jared and Peter Bellwood. 2003. Farmers and Their Languages: The First Expansions.
Science 300: 597603.
Dobžanskaja, O. É. 2002. Pesnja Chotarè [The Song of the Arctic Diver]. St. Petersburg: Drofa.
Dubinskas, Frank A. 1983. A Musical Josephs Coat: Patchwork Patterns and Social Signicance in
World Musics.Reviews in Anthropology 10(3): 2742.
Feld, Steven and Aaron A. Fox. 1994. Music and Language.Annual Review of Anthropology 23: 2553.
Fernandez, M., and M. Jocelyne, ed. 1984. Chants et Poésies Des Sames: Laponie [1 LP]. Orstom
Selaf CETO 806.
Forster, Peter and Colin Renfrew. 2011. Mother Tongue and Y Chromosomes.Science 333(6048):
Grauer, Victor A. 2006. Echoes of Our Forgotten Ancestors.The World of Music 48(2): 559.
——— . 2011. Sounding the Depths: Tradition and the Voices of History. Pittsburg, PA: CreateSpace.
Greenacre, Michael J., and Jörg Blasius. 2006. Multiple Correspondence Analysis and Related
Methods. Boca Raton: Chapman & Hall/CRC.
Greenberg, Joseph H. 1987. Language in the Americas. Stanford, CA: Stanford University Press.
——— . 20002002. Indo-European and Its Closest Relatives: The Eurasiatic Language Family.
Stanford: Stanford University Press.
Hanihara, Kazuro. 1991. Dual Structure Model for the Population History of the Japanese.Japan
Review 2: 133.
——— . 1994. Nijūkōzōmoderu: Nihonjin shūdan no keisei ni kakawaru hitokasetsu [Japanese
translation of Hanihara 1991 with additional afterword].Anthropological Science 102(5): 455
Hartigan, J. A., and M. A. Wong. 1979. Algorithm AS 136: A K-means Clustering Algorithm.
Journal of the Royal Statistical Society, Series C (Applied Statistics) 28(1): 100108.
Hauser, Michael, ed. 1992. Traditional Greenlandic Music [1 CD]. ULO.
Henry, Edward O. 1976. The Variety of Music in a North Indian Village: Reassessing
Cantometrics.Ethnomusicology 20(1): 4966.
Hudson, Mark J. 1999. Ruins of Identity: Ethnogenesis in the Japanese Islands. Honolulu: University
of Hawaii Press.
Hudson, Mark, Ann-elise Lewallen, and Mark K. Watson, eds. 2014. Beyond Ainu Studies:
Changing Academic and Public Perspectives. Honolulu: University of Hawaii Press.
Hughes, David W. 2008. Traditional Folk Song in Modern Japan: Sources, Sentiment and Society.
Folkestone: Global Oriental.
Japanese Archipelago Human Population Genetics Consortium (Timothy Jinam, Nao Nishida,
Momoki Hirai, Shoji Kawamura, Hiroki Oota, Kazuo Umetsu et al.). 2012. The History of
Human Populations in the Japanese Archipelago Inferred from Genome-wide SNP Data with a
Special Reference to the Ainu and the Ryukyuan Populations.Journal of Human Genetics 57:
Keeling, Richard. 2012. Animal Impersonation Songs as an Ancient Musical System in North
America, Northeast Asia, and Arctic Europe.Ethnomusicology 56(2): 234265.
Kōchi, Rie. 2001. On the Variety of Voices in the Musics of Northern Peoples: Focusing on Ainu
Music. In Proceedings of the 15th International Abashiri Symposium (ISSN 0918-7715),15.
Abashiri: HoppōBunka ShinkōKyōkai.
Koganebuchi, Kae, Takafumi Katsumura, Shigeki Nakagome, Hajime Ishida, Shoji Kawamura,
Hiroki Oota and The Asian Archival DNA Repository Consortium. 2012. Autosomal and Y-
Chromosomal STR Markers Reveal a Close Relationship Between Hokkaido Ainu and Ryukyu
Islanders.Anthropological Science 120(3): 199208.
Koizumi, Fumio. 2009. Nihon DentōOngaku No Kenkyū: Gappon [Japanese Traditional Music
Research: Complete], edited by Tomiko Kojima and Harumi Koshiba. Tokyo: Ongaku no
Koskoff, Ellen, ed. 1987. Women and Music in Cross-Cultural Perspective. New York: GreenwoodPress.
Laade, Wolfgang, ed. 1956. Lappish Joik Songs from Northern Norway [1 CD]. Folkways Records
Lecomte, Henri. 2012. Les Esprits Écoutent: Musiques Des Peuples Autochtones de Sibérie [The
Spirits Are Listening: Music of the Autochthonous Peoples of Siberia]. Sampzon: Éditions
Delatour France.
Lecomte, Henri, ed. 19952012. Musique Du Monde: Siberie [11 CDs]. Buda Musique.
Leroi, Armand M., and Jonathan Swire. 2006. The Recovery of the Past.The World of Music 48(3):
Lewis, M. Paul, ed. 2009. Ethnologue: Languages of the World. 16th ed. Dallas: SIL International.
List, George. 1974. The Reliability of Transcription.Ethnomusicology 18(3): 35377.
Lomax, Alan, ed. 1968. Folk Song Style and Culture. Washington, DC: American Association for the
Advancement of Science.
——— . 1976. Cantometrics: An Approach to the Anthropology of Music. Berkeley: University of
California Extension Media Center.
——— . 1980. Factors of Musical Style. In Theory & Practice: Essays Presented to Gene Weltsh,
edited by Stanley Diamond, 2958. The Hague: Mouton.
——— . 1989. Cantometrics.International Encyclopedia of Communications.
Lomax, Alan and Victor Grauer. 1968. The Cantometric Coding Book. In Folk Song Style and
Culture, edited by Alan Lomax, 3474. Washington, DC: American Association for the
Advancement of Science.
Malm, William P. 1967. Music Cultures of the Pacic, the Near East, and Asia. Englewood Cliffs, N.
J.: Prentice-Hall.
Matsukusa, Hirotaka, Hiroki Oota, Kuniaki Haneji, Takashi Toma, Shoji Kawamura and Hajime
Ishida. 2010. A Genetic Analysis of the Sakishima Islanders Reveals No Relationship with
Taiwan Aborigines but Shared Ancestry with Ainu and Main-Island Japanese.American
Journal of Physical Anthropology 142(2): 21133.
Nattiez, Jean-Jacques. 1983. The Rekkukara of the Ainu (Japan) and the Katajjaq of the Inuit
(Canada): A Comparison.The World of Music 25(2): 3344.
Nattiez, Jean-Jacques, and Paula Conlon, eds. 1993. Inuit Iglulik [1 CD]. Museum für Völkerkunde.
——— . 1999. Inuit Throat-games and Siberian Throat Singing: A Comparative, Historical, and
Semiological Approach.Ethnomusicology 43(3): 399418.
Nettl, Bruno. 2015. The Study of Ethnomusicology: Thirty-three Discussions. 3rd ed. Urbana:
University of Illinois Press.
NHK (Nippon HōsōKyōkai), ed. 194494. Nihon MinyōTaikan [Japanese Folk Song Anthology;
13 Volumes; 90 CDs + 660 Unpublished Tapes from the Ryukyu Islands]. Tokyo: NHK.
——— . 1951. Karafuto Ainu No Koyō[Traditional Songs of the Sakhalin Ainu; 21 SPs]. Tokyo:
——— . 1965. Ainu DentōOngaku [Ainu Traditional Music; 4 LPs]. Tokyo: NHK.
Ohyi, Haruo. 1985. On the Process of Crystallization of Sakhalin Ainu.Bulletin of the Institute for
the Study of North Eurasian Cultures, Hokkaido University 17: 16592.
Omoto, Keiichi and Naruya Saitou. 1997. Genetic Origins of the Japanese: A Partial Support for the
Dual Structure Hypothesis.American Journal of Physical Anthropology 102(4): 437446.
Oota, Hiroki, Wannapa Settheetham-Ishida, Danai Tiwawech, Takafumi Ishida and Mark
Stoneking. 2001. Human mtDNA and Y-Chromosome Variation Is Correlated with
Matrilocal Versus Patrilocal Residence.Nature Genetics 29: 2021.
Pamjav, Horolma, Zoltán Juhász, Andrea Zalán, Endre Németh and Bayarlkhagva Damdin. 2012.
A Comparative Phylogenetic Study of Genetics and Folk Music.Molecular Genetics and
Genomics 287(4): 337349.
Patel, Aniruddh D. 2008. Music, Language and the Brain. Oxford: Oxford University Press.
Reich, David, Nick Patterson, Desmond Campbell, Arti Tandon, Stéphane Mazieres, Nicolas Ray,
Maria V. Parra et al. 2012.Reconstructing Native American Population History.Nature 488:
Rzeszutek, Tom, Patrick E. Savage and Steven Brown. 2012. The Structure of Cross-cultural
Musical Diversity.Proceedings of the Royal Society B: Biological Sciences 279: 16061612.
Sasaki, Kōmei. 1997. Nihon Bunka No TajūKōzō: Ajia-Teki Shiya Kara Nihon Bunka o Saikōsuru
[Multiple Structure of Japanese Culture: Rethinking Japanese Culture from an Asian
Perspective]. Tokyo: Shōgakkan.
Sato, Takehiro, Tetsuya Amano, Hiroko Ono, Hajime Ishida, Haruto Kodera, Hirofumi
Matsumura, Minoru Yoneda and Ryuichi Masuda. 2009. Mitochondrial DNA Haplogrouping
of the Okhotsk People Based on Analysis of Ancient DNA: an Intermediate of Gene Flow
from the Continental Sakhalin People to the Ainu.Anthropological Science 117(3): 171180.
Savage, Patrick E., and Steven Brown. 2013. Toward a New Comparative Musicology.Analytical
Approaches to World Music 2(2): 14897.
——— . 2014. Mapping Music: Cluster Analysis of Song-type Frequencies Within and Between
Cultures.Ethnomusicology 58(1): 133155.
Savage, Patrick E., Emily Merritt, Tom Rzeszutek and Steven Brown. 2012. CantoCore: A New
Cross-cultural Song Classication Scheme.Analytical Approaches to World Music 2(1): 87137.
Sheykin, Yuri. 2001. Russian Federation II.3: Siberian Peoples. In The New Grove Dictionary of
Music and Musicians, Volume 22: Russian Federation II Scotland, edited by Stanley Sadie
and John Tyrrell, 2nd ed., 1927. New York: Grove.
Stock, Jonathan P., ed. 2006. Echoes of Our Forgotten Ancestors [Two Special Issues].The World
of Music 48(2): 1134 and 48(3): 163.
Tajima, Atsushi, Masanori Hayami, Katsushi Tokunaga, Takeo Juji, Masafumi Matsuo, Sangkot
Marzuki, Keiichi Omoto and Satoshi Horai. 2004. Genetic Origins of the Ainu Inferred from
Combined DNA Analyses of Maternal and Paternal Lineages.Journal of Human Genetics 49
(4): 187193.
Tanimoto, Kazuyuki. 2000. Ainu-e o Kiku: HenyōNo Minzoku Ongakushi [Listening to Ainu
Paintings: Musical Ethnography of a Cultural Transformation]. Sapporo: HokkaidōDaigaku
Tosho Kankōkai.
——— . 2001a. Music of the Ainu, Nivkhi, and Uilta. In The Garland Encyclopedia of World Music,
Volume 7, East Asia: China, Japan and Korea, edited by Robert C. Provine, Yosihiko Tokumaru
and J. Lawrence Witzleben, 7838. New York: Routledge.
——— . 2001b. Siberia. In The Garland Encyclopedia of World Music, Volume 7, East Asia: China,
Japan and Korea, edited by Robert C. Provine, Yosihiko Tokumaru and J. Lawrence Witzleben,
102731. New York: Routledge.
Tokita, Alison McQueen and David W. Hughes, eds. 2008. The Ashgate Research Companion to
Japanese Music. Aldershot, UK: Ashgate.
Tzanetakis, George, Ajay Kapur, W. Andrew Schloss and Matthew Wright. 2007. Computational
Ethnomusicology.Journal of Interdisciplinary Music Studies 1(2): 124.
Wallin, Nils L., Bjorn Merker and Steven Brown, eds. 2000. The Origins of Music. Cambridge: MIT
Wood, Anna L. 2008. Performance Style Research: Cantometrics and Related Studies of
Communication.In Rhythms of Earth [set of 4 DVDs and Accompanying Documentation],
edited by Alan Lomax, Forrestine Paulay, John Melville Bishop and Association for Cultural
Equity, 162. Portland, OR: Media Generation.
Wright, Sewall. 1943. Isolation by Distance.Genetics 28: 11438.
  • ... Yet language is only one out of many complex cultural traits that could serve as a proxy for deep history. It has been proposed that music may preserve even deeper cultural history than language (36)(37)(38)(39)(40)(41). Standardized musical classification schemes (based on features such as rhythm, pitch, and singing style) can be used to quantify patterns of musical diversity among populations for the sake of comparison with genetic and linguistic differences (36,37,40,42). ...
    ... It has been proposed that music may preserve even deeper cultural history than language (36)(37)(38)(39)(40)(41). Standardized musical classification schemes (based on features such as rhythm, pitch, and singing style) can be used to quantify patterns of musical diversity among populations for the sake of comparison with genetic and linguistic differences (36,37,40,42). Among indigenous Taiwanese populations speaking Austronesian languages, such analyses revealed significant correlations between music, mitochondrial DNA, and the lexicon (37), suggesting that music may indeed preserve population history. ...
    ... Asia provides a useful test region because it contains high levels of genetic and cultural diversityincluding a large number of small language families or linguistic isolates (e.g., Tungusic, Chukuto-Kamchatkan, Eskimo-Aleut, Yukagir, Ainu, Korean, Japanese) (34). Crucially, while genetic and linguistic data throughout much of the world have been published, Northeast Asia is the only region for which published musical data allow for direct matched comparison of musical, genetic, and linguistic diversity (40). ...
    Full-text available
    Culture evolves in ways that are analogous to, but distinct from, genetic evolution. Previous studies have demonstrated correlations between genetic and cultural diversity at small scales within language families, but few studies have empirically investigated parallels between genetic and cultural evolution across multiple language families using a diverse range of cultural data. Here we report an analysis comparing cultural and genetic data from 13 populations from in and around Northeast Asia spanning 10 different language families/isolates. We construct distance matrices for language (grammar, phonology, lexicon), music (song structure, performance style), and genomes (genome-wide SNPs) and test for correlations among them. After controlling for spatial autocorrelation and recent contact, robust correlations emerge between genetic and grammatical distances. Our results suggest that grammatical structure might be one of the strongest cultural indicators of human population history, while also demonstrating differences among cultural and genetic relationships that highlight the complex nature of human cultural and genetic evolution.
  • ... Since Lomax's death in 2002, however, his daughter Anna Lomax Wood (who took over administration of Lomax's Association for Cultural Equity [ACE]) and Victor Grauer (co-inventor of cantometrics) have worked toward this goal. This has taken place both independently (Grauer, 2005(Grauer, , 2006(Grauer, , 2007(Grauer, , 2009(Grauer, , 2011Wood, 2018, Forthcoming) and in tandem or in parallel with a wide variety of scientists, including Sarah Tishkoff's human genetics lab (Callaway, 2007), Armand Leroi's evolutionary biology lab (Busby, 2006;Leroi et al., 2015;Leroi & Swire, 2006), Simon Dixon's music information retrieval lab (Panteli, Benetos, & Dixon, 2016Panteli, Bittner, Bello, & Dixon, 2017), Quentin Atkinson's Cultural Evolution Lab (Du Toit, 2011;Savage & Atkinson, 2015), Steven Brown's NeuroArts Lab Ellis et al., 2018;Savage, Merritt, Rzeszutek, & Brown, 2012), and my own Comp-Music Lab (Savage & Brown, 2013Savage, Brown, Sakai, & Currie, 2015;Savage, Matsumae, et al., 2015). At the same time, others have investigated similar relationships with minimal reference to Lomax's ideas or methods (Le Bomin, Lecointre, & Heyer, 2016;Mehr, Singh, York, Glowacki, & Krasnow, 2018;Pamjav, Juhász, Zalán, Németh, & Damdin, 2012). ...
    ... Previously, Steven Brown and I (Savage & Brown, 2013) have suggested a modified version of Lomax's methodology, aiming to use 30 traditional songs per culture, selected strictly randomly from the pool of available songs to avoid confirmation bias in the selection. In our studies of musical/genetic diversity among indigenous Taiwanese Rzeszutek et al., 2012; and Northeast Asian/circumpolar populations (Savage, Matsumae, et al., 2015), we attempted to apply these methods to densely sampled sets of populations within geographically restricted regions in collaboration as much as possible with regional experts (including both ethnomusicologists and genetic anthropologists). In practice, this has proved extremely time-consuming (spending hundreds of hours coding songs by ear) and has not substantially reduced the sampling problems. ...
    ... It has proven hard to evaluate the degree of confirmation bias in Lomax's dataset, although this may be possible now that the Global Jukebox data have been published. My colleagues and I have striven to reduce the chance of confirmation bias by performing coding blind to extra-acoustic information as much as is practically possible, 8 having randomly selected quality-control sub-samples coded by raters who are both blind to extra-acoustic information and unaware of the research hypotheses, and by publishing full data spreadsheets as online supplemental material to allow others to scrutinize our sample and codings (e.g., Savage, Matsumae, et al., 2015). For maximum objectivity, all analyses would be performed by raters blind to the research hypotheses, as is sometimes done in biological or medical studies, but the level of time commitment involved has so far made this logistically impossible. ...
    Full-text available
    Alan Lomax’s Cantometrics Project was arguably both the most ambitious and the most controversial undertaking in music and science that the world has known. Its flagship component, Lomax’s “cantometric” analysis of approximately 1,800 songs from 148 worldwide populations using 36 classificatory features, sparked extensive debate. While Lomax responded to some criticisms, neither his final conclusions nor the evidence on which they were based were ever fully made clear. For decades, neither cantometrics nor Lomax’s related projects involving dance, speech, popular music, digital humanities, pedagogy, and activism were widely adopted by other researchers, but there has been a resurgence of interest since Lomax’s death in 2002. Here, I provide a comprehensive critical review of the Cantometrics Project, focusing on issues regarding the song sample, classification scheme, statistical analyses, interpretation, and ethnocentrism/reductionism. I identify misunderstandings, improvements that were made, and criticisms that remain to be addressed, and distil Lomax’s sometimes-conflicting claims into diagrams summarizing his three primary results: (1) ten regional song-style types, (2) nine musical factors representing intra-musical correlations, and (3) correlations between these musical factors and five factors of social structure. Although Lomax’s interpretations regarding correlations between song style and social structure appear weakly supported, his historical interpretations regarding connections ranging from colonial diaspora to ancient migrations provide a more promising starting point for both research and teaching about the global arts. While Lomax’s attempts to correlate features of social structure such as gender, religion, politics, and economics with stylistic features of musical performance largely failed to gain acceptance, the Cantometrics Project can still provide both inspiration and cautionary lessons for future exploration of relationships between music and culture.
  • ... Lomax then worked with anthropologists such as Conrad Arensberg to integrate his Cantometric data with social structure data from George Murdock's (1967) Ethnographic Atlas and run statistical correlations to test his hypothesis that a culture's song style was functionally connected to its social structure (Lomax 1968(Lomax , 1976(Lomax , 1980(Lomax , 1989. Brown, and myself have tried reviving/refining Lomax's Cantometric approach, focusing not on correlations with social structure but instead on the role of music as a marker of human history (Leroi and Swire 2006;Leroi et al. 2015;Wood 2018aWood , 2018bCallaway 2007;Grauer 2006;Grauer 2011;Savage et al. 2012;Rzeszutek, Savage, and Brown 2012;Brown et al. 2014;Savage and Brown 2014;Savage, Matsumae, et al. 2015). In 2017, the Association for Cultural Equity that Lomax founded finally publicly released 5,899 Cantometric codings along with associated recordings and metadata online at ...
    ... Likewise, my own research has identified significant correlations between musical, genetic, and linguistic diversity among nine indigenous populations of Taiwan ), which we argue suggests that music, like languages and genes, can act as a marker of human population history. However, because we have not yet been able to make this direct comparison for other regions, it remains to be seen whether this relationship holds at a global level (although this possibility is bolstered by some suggestive evidence of music-gene correlations from other regional studies ;Callaway 2007;Pamjav et al. 2012;Savage, Matsumae, et al. 2015). ...
    Full-text available
    The past few decades have seen a rapid increase in the availability and use of large music corpora. However, most music corpus studies remain limited to Western music, limiting our ability to understand the diversity and unity of human music throughout the world. I argue for the potential of cross-cultural corpus studies to contribute to comparative musicological studies in domains including music classification, evolution, universals, and human history. I highlight and discuss a number of important cross-cultural corpora, including the Berlin Phonogramm-Archiv, CompMusic Project, Essen Folksong Collection, Cantometrics Project, Garland Encyclopedia of World Music, Natural History of Song, and Deep History of Music projects. In the process, I discuss the pros and cons of music notation vs. recordings, automatic vs. manual analysis, regional vs. global analysis, and associated challenges regarding choosing appropriate analysis methods that can allow meaningful comparison across cultures. I argue for the need for bigger and better global music corpora and more emphasis on integration within and beyond academia, including to domains such as the music industry and cultural heritage organizations.
  • ... Moreover, and although I do not have the space to develop the point here, there is reason to think that music and species are candidates for a good conceptual analogy. Just as scientists create phylogenetic models of species, the study of phylogenetic models of musics is an emerging and fruitful research project (see e.g. Brown et al. 2013;Rzeszutek et al. 2012;Savage et al. 2015;Savage 2019). ...
    Full-text available
    According to pluralism about some concept, there are multiple non-equivalent, legitimate concepts pertaining to the (alleged) ontological category in question. It is an open question whether conceptual pluralism implies anti-realism about that category. In this article, I argue that at least for the case of music, it does not. To undermine the application of an influential move from pluralism (about music concepts) to anti-realism (about the music category), then, I provide an argument in support of indifference realism about music, by appeal to music archaeological research, via an analogy with Adrian Currie’s indifference realism about species licensed by paleobiological research.
  • ... Thus, it is conceivable that shared aspects related to song contour in humans and songbirds may also be convergently coevolved with vocal learning. Future comparisons analyzing song contour variation among different species of birds, including vocal nonlearners (e.g., common loons, Gavia immer) and in different modes of vocalization in humans (e.g., ingressive throat-rasping in Inuit/Ainu; Nattiez, 1999;Savage, Matsumae, et al., 2015) may help to distinguish between these or other potential alternatives, such as the adoption of learned ''templates'' (Ammirante & Russo, 2015). ...
    Full-text available
    There has recently been renewed interest in using quantitative data to explore questions about musical universals. One explanation for certain musical universals is that they reflect ways of singing that are most energetically efficient, as opposed to biological specializations for human music. Previous research found support for this ‘‘motor constraint hypothesis’’ by comparing pitch contour shapes in samples of human and avian songs, but the sample of human songs was limited to notated scores of European and Chinese folk songs from the Essen database. Here we test this hypothesis using a more diverse global sample of human music recordings from the Garland Encyclopedia of World Music. By directly comparing pitch contour shapes in a diverse sample of human songs and bird songs, we found that both human and bird songs tend to employ similar descending/arched melodic contours despite substantial differences in absolute pitch and duration. This preference was consistent for both Western and non-Western songs. Surprisingly, we also found that the global samples of human and bird song contours were significantly more correlated with one another than either was with the Essen contours. Our findings of broad cross-cultural and cross-species parallels support the motor constraint hypothesis for melodic contour. More generally, our findings demonstrate the importance of greater collaboration between ethnomusicology and music psychology.
  • ... 0-1. 1 0-2. 2 Savage and Atkinson 2015; Savage, Matsumae, et al. 2015; Brown, Savage, et al. 2014; Rzeszutek, Savage, and Brown 2012; Ellis et al. in press; Leroi et al. 2015; Takezawa et al. 2014) (Danchin et al. 2011: 483 ) 2 Cecil Sharp (1907Bertrand Bronson (1959Bronson ( -72, 1969Bronson ( , 1976) Samuel Bayard (1950Bayard ( , 1954) (Bartók 1931; Wiora 1953; 1965) 1 Bronson 1959Bronson -72, 1969Bronson , 1976) (Bayard 1950Bayard , 1954) Charles Seeger (Bayard 1954; Bronson 1959-72; 1965; Marett 1985) 172 where ...
    Full-text available
    Darwin's theory of evolution provided striking explanatory power that has come to unify biology and has been successfully extended to various social sciences. In this dissertation, I attempt to show how cultural evolutionary theory may also hold promise for explaining diverse musical phenomena, using a series of quantitative case studies from a variety of cultures and genres to demonstrate general laws governing musical change. Chapter one describes previous research and debates regarding music and cultural evolution. Drawing on major advances in the scientific understanding of cultural evolution over the past three decades, I clarify persistent misconceptions about the roles of genes and progress in definitions of evolution, showing that neither is required or assumed. I go on to review older and recent literature relevant to musical evolution at a variety of levels, from Lomax's macroevolutionary interpretation of global patterns of song-style to microevolutionary mechanisms by which minute melodic variations give rise to large tune families. To highlight the complex dynamics of musical evolution in action, I provide an autoethnographic case study of my own performance of folk songs in Japan. After addressing criticisms of the roles of individual agency and reductionism in studying musical evolution, I highlight ways in which cultural evolutionary theory can contribute to applied ethnomusicology in the domains of education, copyright, and sustainability. Chapter two introduces a new method for quantifying aspects of musical evolution. This method builds off of the tune family concept, but adds quantitative rigour by adapting tools from molecular genetics. In particular, I highlight analogies between protein evolution and melodic evolution. The former can be modeled as sequences constructed from an "alphabet" of 20 amino acids, while the latter can be modeled as sequences constructed from an "alphabet" of 12 notes, corresponding to the equal-tempered chromatic scale. This makes it possible to adapt sequence alignment methods from molecular genetics to quantify the evolution of any music - Western or non-Western - that can be approximated by standard staff notation, and to make automated comparisons on scales far beyond the capabilities of unassisted humans. I use examples of melodic evolution (Scarborough Fair) and protein evolution (avian influenza) to demonstrate analogies in the process of coding and analyzing their evolution. Chapter three tests hypotheses about general trends in musical evolution against a large sample of 4,125 British-American "Child ballads" notated between 1575-1972. Using the sequence alignment methods developed in Chapter two, I automatically identified and analysed 172 pairs of highly related (>85% identity) melodic variants encompassing a total of 15,786 notes. Mutation rates varied greatly (over 100-fold) in ways that followed general predictions of cultural evolutionary theory: 1) written notation evolves more slowly than oral transmission, 2) functional notes are more resistant to change than ornamental notes, and 3) substitutions are more likely to occur between small melodic distances than large ones. In addition, insertions and deletions greatly outnumbered substitutions, but there was no clear trend toward complexity (i.e., insertion) or simplification (i.e., deletion). I hypothesize that these trends are governed primarily by universal cognitive constraints, and thus are likely to characterize musical evolution cross-culturally. Chapter four extends the method developed in Chapter two to explore the generality of the trends identified in Chapter three using a diverse set of case studies in which the history of musical evolution has already been qualitatively documented: 1) the divergence of the Scottish 17th c. Lady Cassiles Lilt into nearly unrecognizable 20th c. American descendants, 2) the merging of work songs from distant prefectures into the Japanese folk song Esashi Oiwake, 3) the simultaneous performance of vestigial, inaudible 1,000-year-old Chinese melodies and their radically changed descendants in the Japanese gagaku piece Seigaiha, and 4) the legal cases finding George Harrison's My Sweet Lord (1970) and Robin Thicke and Pharrell Williams' Blurred Lines (2013) guilty of plagiarism. Although the precise mechanisms differ and absolute rates of evolution vary almost 400-fold within and between these case studies, the general trends supported my predictions from Chapter three regarding: 1) the relative ease of mutations to nearby pitches, 2) the relative predominance of insertions/deletions over substitutions, 3) the relative stability of functional notes (e.g., stressed vs. unstressed), and 4) the relative stability of written over oral traditions. Both increases and decreases in complexity were observed, with no clear trend favouring one or the other. This dissertation demonstrates that musical evolution can be rigorously measured by adapting theoretical and methodological tools from cultural evolution and molecular genetics, and applies this to show that musical evolution follows some general rules analogous to ones governing genes and non-musical domains of culture. Although cultural evolutionary theory and methods alone will not solve all the problems facing musicologists interested in the phenomenon of musical change, it does offer a new unified set of tools that help answer at least some longstanding questions of broad general concern. ダーウィンの進化論はその高い説明力で生物学を統一することに成功し、社会科学にも応用されてきた。本論文では、文化的進化論のアプローチが様々な音楽的対象にも応用できる可能性を論じ、幾つかの異なる文化とジャンルを含む多様な計量的ケース・スタディーを通して音楽的変化を制限する一般的な規則の存在を提示する。 第一章では、音楽と文化的進化の先行研究をまとめて説明する。この30年間での文化的進化においての科学的見解の進歩に基づき、根強い誤解に関して指摘する。特に、「進化」の定義には遺伝子も進歩も必要とされていない。そして、 世界の民謡様式におけるマクロ進化的パターンから、細かい旋律の変化による大きな曲族が生まれるミクロ進化的メカニズムまで、既存の先行研究について言及する。音楽進化における複雑な動力を見せるため、筆者の日本における民謡演奏の自民族詩的ケース・スタディーを提供する。音楽進化に対する個人のエイジェンシーと還元主義の役割の批判に答えた後、最後に文化的進化論が教育、著作権、持続可能性などの分野において応用音楽学に貢献できることを論じる。 第二章では、音楽進化を計量的に測る新しい方法を提示する。この方法は曲族研究に基づくが、分子遺伝学的な方法を応用することによって、曲族の概念を計量化することが可能になる。特に、たんぱく質進化と旋律進化の類似性を強調する。たんぱく質を20のアミノ酸の「アルファベット」から作られた配列としてモデル化することができ、旋律を(平均律半音階に当たる)12の「アルファベット」から作られた配列としてモデル化することができる。これによって、西洋・東洋にかかわらず、五線譜化さえできれば分子遺伝学のために作られた配列整列方法を音楽進化を測るために応用することができ、人の手だけでは不可能な量の比較を自動的にできるようになる。コーディングと分析の過程の類似を説明するために、旋律(「スカボロー・フェア Scarborough Fair 」)の進化とたんぱく質(鳥インフルエンザ)の進化の実例を示す。 第三章では、音楽進化の一般的傾向の仮説を大きなサンプルで検討する(1575〜1972年の間に記譜された 4,125 のブリティッシュ・アメリカンの「チャイルド・バラード」民謡)。第二章で提供した配列整列方法で、172の高い類似を持った旋律のペア(音符が85パーセント以上同じのもの同士)が見出された(全曲総量15,786 音符)。突然変異率は大きく(100倍以上)異なり、一般的な文化的進化論の仮説の予測通り、次のような結果になった:1)楽譜伝承は口頭伝承より進化が遅かった。2)機能的音符は装飾的音符より変化が認められなかった。3)旋律的に近い音程への変化が比較的多かった。また、置換より挿入・削除の方が多かったが、はっきりとした複雑さ(つまり、挿入)への傾向も、簡略化(つまり、削除)への傾向も見られなかった。これらの傾向は、主に普遍的な知覚的制限によるもので、異文化間でも音楽進化に見られるだろうという仮説を提示する。 第四章では、第三章で見出した仮説の一般性を検討するため、第二章において提示した方法を展開してゆく。本章では、音楽的進化の歴史が記録された以下の様々なケース・スタディーを用いる:1)17世紀スコットランドの「カシリス婦人のリルト Lady Cassiles Lilt 」の、20世紀アメリカにおけるその子孫へのほぼ認識不可能な変貌、2)遠く離れたいくつかの県の労働唄の合併からできた日本民謡の「江差追分」、3)雅楽の「青海波」における非常に異なった笛と笙の旋律の、一千年以上前の共通起源、そして4)ハリソン Harrison の「マイ・スィート・ロード My Sweet Lord 」(1970年) とシック Thicke とウィリアムズ Williams の「ブラード・ラインズ Blurred Lines 」(2013年)が剽窃と判断された法律事件。進化的メカニズムの詳細が異なったり、絶対的進化の「突然変異率」がほぼ400倍異なったりするも、一般的なパターンは第三章の仮説の予測通りであった。つまり、1)旋律的に近い音程への変化が比較的多く、2)置換より挿入・削除の方が多く、3)機能的音符は比較的変化が認められない(リズム的に強調された音符は不強調音符より変化が認められない)結果となり、4)口頭伝承より楽譜伝承の方が変化が少なかった。複雑さの程度が上下する事例もあり、どちらかに偏るはっきりとした傾向は見られなかった。 本論文は、文化的進化と分子遺伝学の理論や方法を適応することによって、音楽進化を計量的に測ることができることを明らかにし、遺伝子や言語と同じように、音楽進化は幾つかの一般的な規則に制限されていることを明らかにする。もちろん、音楽的変化に興味を持つ音楽学者が立ち向かう問題の全てを、文化的進化の理論と方法だけで解決できるわけではない。しかし、少なくとも長年論争してきた問題解決に役立つ、新しい統一されたツールとして貢献できることを本論文で論じる。
  • Article
    Full-text available
    The concept of cultural evolution was fundamental to the foundation of academic musicology and the subfield of comparative musicology, but largely disappeared from discussion after World War II despite a recent resurgence of interest in cultural evolution in other fields. I draw on recent advances in the scientific understanding of cultural evolution to clarify persistent misconceptions about the roles of genes and progress in musical evolution, and review literature relevant to musical evolution ranging from macroevolution of global song-style to microevolution of tune families. I also address criticisms regarding issues of musical agency, meaning, and reductionism, and highlight potential applications including music education and copyright. While cultural evolution will never explain all aspects of music, it offers a useful theoretical framework for understanding diversity and change in the world’s music.
  • Article
    Full-text available
    Part 1 of this article explored the development of Alan Lomax's theory of expressive style and culture in terms of his family background, temperament, cultural surroundings, influences, intellectual growth, and the nature and impact of his field experiences, the primary source of his ideas. It described the program of research established by Lomax and anthropologist Conrad Arensberg, codirector of the project, and set out its areas of investigation. Part 2 describes the project's hypotheses, methodology, and results; reviews its criticisms; and takes a personal, reflexive look at Lomax's personality, public persona, and presentational style as factors in the reception of his work. Finally, it touches upon the potential of a comparativist, cross-cultural approach to the current study of expressive culture.
  • Article
    The comparison of world music cultures has been a recurring topic in the field of musicology since the end of the nineteenth century. Recent advances in technology in the field of Music Information Retrieval allow for a large-scale analysis of music corpora. We review manual and computational approaches in the literature that fall within the scope of music corpus research and world music analysis. With a large-scale computational music corpus analysis in mind, we compare the tools and research questions addressed by each study and discuss strengths and weaknesses. Taking into account critical remarks from experts in the field and challenges involved in a large-scale computational analysis, we discuss how this line of research can be improved in future work.
  • Thesis
    Full-text available
    ダーウィンの進化論はその高い説明力で生物学を統一することに成功し、社会科学にも応用されてきた。本論文では、文化的進化論のアプローチが様々な音楽的対象にも応用できる可能性を論じ、幾つかの異なる文化とジャンルを含む多様な計量的ケース・スタディーを通して、音楽的変化を制限する一般的な規則の存在を提示した。 第一章では、音楽と文化的進化の先行研究をまとめて説明した。この30年間での文化的進化においての科学的見解の進歩に基づき、根強い誤解に関して指摘した。特に、「進化」の定義には遺伝子も進歩も必要とされていないことを強調した。そして、 世界の民謡様式におけるマクロ進化的パターンから、細かい旋律の変化による大きな「曲族」 (tune family) が生まれるミクロ進化的メカニズムまで、既存の先行研究について言及した。音楽進化における複雑な動力を示すため、筆者の日本における民謡演奏の自民族誌的ケース・スタディーを提供した。音楽進化に対する個人のエイジェンシーと還元主義の役割の批判に答えた後、最後に文化的進化論が教育、著作権、持続可能性などの分野において応用音楽学に貢献できることを論じた。 第二章では、音楽進化を計量的に測る新しい方法を提示した。この方法は曲族研究に基づくが、分子遺伝学的な方法を応用することによって、曲族の概念を計量化することが可能になる。特に、たんぱく質進化と旋律進化の類似性を強調した。たんぱく質を20のアミノ酸の「アルファベット」から作られた配列としてモデル化することができるように、旋律を(平均律半音階に当たる)12の「アルファベット」から作られた配列としてモデル化することができる。これによって、西洋・東洋にかかわらず、五線譜化さえできれば分子遺伝学のために作られた配列整列方法を音楽進化を測るために応用することができ、人の手だけでは不可能な量の比較を自動的にできるようになる。コーディングと分析の過程の類似を説明するために、旋律(「スカボロー・フェア Scarborough Fair 」)の進化とたんぱく質(鳥インフルエンザ)の進化の実例を示した。 第三章では、音楽進化の一般的傾向の仮説を大きなサンプルで検討した。ここで1575年から1972年の間に記譜された 4,125 のブリティッシュ・アメリカン(英米)の「チャイルド・バラード」民謡を事例とした。第二章で提供した配列整列方法で、全曲総量15,786 音符のうち、85パーセント以上の音符を同じくする高い類似を持った旋律のペアが 172 件見出された。突然変異率は大きく(100倍以上)異なり、一般的な文化的進化論の仮説の予測通り、次のような結果になった:1)楽譜伝承は口頭伝承より進化が遅かった。2)機能的音符は装飾的音符より変化が認められなかった。3)旋律的に近い音程への変化が比較的多かった。また、置換より挿入・削除の方が多かったが、はっきりとした複雑さ(つまり、挿入)への傾向も、簡略化(つまり、削除)への傾向も見られなかった。これらの傾向は、主に普遍的な知覚的制限によるもので、異文化間でも音楽進化に見られるだろうという仮説を提示した。 第四章では、第三章で見出した仮説の一般性を検討するため、第二章において提示した方法を展開した。本章では、音楽的進化の歴史が記録された以下の様々なケース・スタディーを用いる:1)17世紀スコットランドの「カシリス婦人のリルト Lady Cassiles Lilt 」の、20世紀アメリカにおけるその子孫へのほぼ認識不可能な変貌、2)遠く離れたいくつかの県の労働唄の合併からできた日本民謡の「江差追分」、3)雅楽の「青海波」における非常に異なった笛と笙の旋律の、一千年以上前の共通起源、そして4)ハリソン Harrison の「マイ・スィート・ロード My Sweet Lord 」(1970年) とシック Thicke とウィリアムズ Williams の「ブラード・ラインズ Blurred Lines 」(2013年)が剽窃と判断された法律事件。それぞれの例は進化的メカニズムの詳細が異なったり、絶対的進化の「突然変異率」がほぼ400倍異なったりしているが、一般的なパターンは第三章の仮説の予測通りであった。つまり、1)旋律的に近い音程への変化が比較的多く、2)置換より挿入・削除の方が多く、3)機能的音符は比較的変化が認められない(リズム的に強調された音符は強調されない音符より変化が認められない)結果となり、4)口頭伝承より楽譜伝承の方が変化が少なかった。複雑さの程度が上下する事例もあり、どちらかに偏るはっきりとした傾向は見られなかった。 本論文は、文化的進化と分子遺伝学の理論や方法を適用することによって、音楽進化を計量的に測ることができることを明らかにし、遺伝子や言語と同じように、音楽進化は幾つかの一般的な規則に制限されていることを明らかにした。もちろん、音楽的変化に興味を持つ音楽学者が立ち向かう問題の全てを、文化的進化の理論と方法だけで解決できるわけではない。しかし、少なくとも長年論争してきた問題解決に役立つ、新しい統一されたツールとして貢献できることを本論文で論じた。
  • Chapter
    This introductory chapter examines the passing of a resolution recognizing Ainu as “Indigenous to the northern part of the Japanese archipelago, and especially Hokkaido.” This legislative triumph was tempered by conditions attached to Japan's 2007 signing of the United Nations Declaration on the Rights of Indigenous Peoples. Firstly, it was made clear that the Japanese government took an exceptionalist position to international discourse by stating that what the international community regarded as “Indigenous” did not apply in Japan. Secondly, in a 2009 report drafted by a panel of experts charged with assessing the resolution, questions such as colonial history, Hokkaido settlement, and Ainu identity were carefully framed to sidestep calls for decolonization or recommendations for constitutional reform.
  • Article
    The Japanese say that 'folk song is the heart's home town'. Traditional folk songs (min'yo) from the countryside are strongly linked to their places of origin and continue to play a role there. Today, however, they are also taught as a quasi-art music, arranged for stage and television, quoted in Westernized popular songs and so forth.
  • Article
    In the first comprehensive study of the relationship between music and language from the standpoint of cognitive neuroscience, the author challenges the widespread belief that music and language are processed independently. Since Plato's time, the relationship between music and language has attracted interest and debate from a wide range of thinkers. Recently, scientific research on this topic has been growing rapidly, as scholars from diverse disciplines including linguistics, cognitive science, music cognition, and neuroscience are drawn to the music-language interface as one way to explore the extent to which different mental abilities are processed by separate brain mechanisms. Accordingly, the relevant data and theories have been spread across a range of disciplines. This book provides the first synthesis, arguing that music and language share deep and critical connections, and that comparative research provides a powerful way to study the cognitive and neural mechanisms underlying these uniquely human abilities.
  • Article
    L'A. demontre comment les jeux de gorge et les chants de gorge se ressemblent mais ont des significations differentes dans les societes polaires : chez les Inuit du Canada, le chant est appele katajjaq, chez les Ainu de l'ile de Sakhalin (un ancien territoire japonais avant son annexion par l'Union Sovietique en 1945), le chant est nomme rekutkar, et chez les Chukchi de la Siberie russe, le chant est le pic eynen. Chez les Inuit, ces chansons de gorge pratiquees seulement par les femmes sont generalement separees des chants accompagnes de tambours ou des musiques de danse. Ces trois genres de chant de gorge utilisent les memes techniques vocales. A partir de methodes phylogenetiques et semiologiques, l'A. explique les similarites et les differences entre ces genres.
  • Article
    Full-text available
    We propose a return to the forgotten agenda of comparative musicology, one that is updated with the paradigms of modern evolutionary theory and scientific methodology. Ever since the field of comparative musicology became redefined as ethnomusicology in the mid-20th century, its original research agenda has been all but abandoned by musicologists, not least the overarching goal of cross-cultural musical comparison. We outline here five major themes that underlie the re-establishment of comparative musicology: (1) classification, (2) cultural evolution, (3) human history, (4) universals, and (5) biological evolution. Throughout the article, we clarify key ideological, methodological and terminological objections that have been levied against musical comparison. Ultimately, we argue for an inclusive, constructive, and multidisciplinary field that analyzes the world's musical diversity, from the broadest of generalities to the most culture-specific particulars, with the aim of synthesizing the full range of theoretical perspectives and research methodologies available.