Was muziek vroeger complexer? [Dutch]

Foto’s Getty Images, uit het besproken NRC artikel.

Hoe je muziek representeert op een manier die werkelijk iets zegt over haar structuur en ervaring, is al decennialang een centraal probleem binnen de muziektheorie, cognitieve musicologie en computationele muziekwetenschap. Wie ooit een ISMIR- of ICMPC-conferentie heeft bezocht, weet hoe principieel die vraag is: elke representatie (denk bijvoorbeeld aan de westerse muzieknotatie) maakt sommige eigenschappen zichtbaar en duwt andere naar de achtergrond. Wat een model kan vinden, hangt rechtstreeks samen met wat je vooraf besluit als muziek te coderen.

MIDI is daarvan misschien wel het bekendste voorbeeld. Het is in de jaren tachtig ontwikkeld als praktisch communicatieprotocol voor elektronische toetsinstrumenten — synthesizers, keyboards, digitale piano’s: welke toets wordt wanneer ingedrukt, hoe hard, en hoe lang. Een pianouitvoering is er goed mee te vast te leggen en te repliceren, maar het is natuurlijk alles behalve een realistische afspiegeling van hoe muziek gehoord, onthouden, gemaakt of gewaardeerd wordt.

 

Tegen die achtergrond leest het artikel in Scientific Reports, dat deze week in de wetenschapsbijlage van het NRC werd besproken, als een technologisch indrukwekkende, maar theoretisch opvallend naïeve exercitie. De auteurs analyseerden zo’n twintigduizend MIDI-bestanden met behulp van netwerkanalyse, waarbij noten worden voorgesteld als knopen en overgangen tussen noten als verbindingen. Het resultaat oogt mathematisch elegant. Alleen: de muzikale werkelijkheid waarop die elegantie zou moeten slaan, is onderweg grotendeels verdwenen.

Want wat hier “muzikale complexiteit” heet, blijkt in feite vooral variatie in toonhoogteovergangen te meten (welke toets werd na een voorgaande toets ingedrukt). Ritme, metrum, dynamiek, articulatie, timbre, meerstemmingheid, vormopbouw — de dimensies waarin veel muziek haar spanning, identiteit en emotionele werking ontleent — spelen nauwelijks of geen rol. Een Bach-fuga, een compositie voor Javaanse gamelan of een solo van John Coltrane zijn hopeloos verloren.

Dat is geen onschuldige vereenvoudiging. Het bepaalt direct welke conclusies je überhaupt kunt trekken. Wie muziek reduceert tot toonhoogtenetwerken, zal uiteindelijk vooral patronen in toonhoogtenetwerken ontdekken. Dat klinkt triviaal, maar precies die cirkelredenering sluipt hier voortdurend binnen.

Ook het gebruik van MIDI-data zelf wringt. Partituurachtige invoer en expressieve uitvoeringen worden vrijwel identiek behandeld, waardoor rubato, microtiming en uitvoeringsnuance effectief worden weggefilterd. Polyfonie wordt bovendien teruggebracht tot opeenvolgingen van nootclusters, zonder serieuze modellering van stemvoering, contrapunt of harmonische functie. Muziek wordt daarmee behandeld alsof zij primair een lineair pad door toonhoogteruimte is. Een opmerkelijke aanname voor een studie die pretendeert iets algemeens te zeggen over muzikale evolutie.

De omgang met transpositie maakt het probleem nog scherper. De kernnetwerken zijn gebaseerd op absolute toonhoogtes, terwijl slechts een deel van de embeddings intervalrelaties transpositie-invariant behandelt. Daarmee raakt het model verstrikt in een van de oudste inzichten uit de muziekcognitie: mensen luisteren doorgaans relatief, niet absoluut. Een melodie blijft herkenbaar wanneer zij een toon hoger wordt gespeeld. Het model lijkt dat inzicht maar half te begrijpen.

Toch trekken de auteurs vervolgens de forse conclusie dat muziek door de tijd heen “eenvoudiger” zou zijn geworden. Dat is een nogal spectaculaire claim voor een analyse die blind blijft voor productie, timbre, ritmische gelaagdheid, studiotechniek en performatieve subtiliteit — kortom: voor vrijwel alles waarin veel hedendaagse muziek juist haar complexiteit organiseert. Misschien is complexiteit helemaal niet verdwenen, maar verplaatst. Van melodische variatie naar klankkleur. Van harmonische modulatie naar microtiming. Van noot naar sound.

En daar wringt uiteindelijk het diepste probleem. Muziek bestaat niet uit symbolen. Een MIDI-bestand is geen muziek, hooguit een recept, een uitvoeringsprotocol (in het geval van een pianouitvoering). Wie de geschiedenis van muziek probeert te begrijpen via zulke abstracties, loopt het risico hetzelfde te doen als iemand die de evolutie van spreektaal reconstrueert op basis van spreadsheets met letterfrequenties. Je vindt ongetwijfeld patronen. Maar de stem zelf — timing, intonatie, aarzeling, nadruk, ademhaling — is dan al verdwenen voordat de analyse begint.

Referenties
Di Marco, N., Loru, E., Galeazzi, A. et al. (2026). Decoding the evolution of melodic and harmonic structure of Western music through the lens of network science. Sci Rep 16, 11121. https://doi.org/10.1038/s41598-026-42872-7

Do bumble bees sense rhythmic patterns?

Zeng et al. (2026, Science) reported an intriguing study of rhythmic pattern discrimination in bumble bees (Bombus terrestris). However, the claim that “[bumble bees] form robust abstract rhythm representations” may be somewhat premature.

Overall, the study is fascinating: bees learned to discriminate flashing temporal patterns and appeared to generalize across tempi and sensory modalities. But a key question is whether this shows rhythm abstraction, or whether simpler temporal cues could explain the results.

A defining feature of rhythm cognition is tempo invariance: recognizing a temporal pattern when all its intervals are proportionally stretched or compressed. In Zeng et al.’s tempo-generalization experiment, however, flash durations varied while the silent gaps reportedly remained fixed at 100 ms. This means the test stimuli were not true proportional transformations of the training stimuli. Instead, they combined changing flash durations with fixed inter-flash gaps. That weakens the interpretation that bees recognized an abstract rhythmic relation.

There is also a simpler possible strategy. Bees may not have encoded the full pattern, but instead relied on local cues such as immediate element repetition or matching familiar temporal fragments such as a particular flash-plus-gap combination. Such strategies would still be cognitively interesting, but they are not the same as forming a global abstract rhythm representation.

The authors also suggest their findings challenge the hypothesis that vocal learning and flexible rhythm perception are linked. But that hypothesis concerns advantages for auditory rhythm processing in vocal-learning species; visual and vibrational discrimination in bees does not directly test it.

Bumble bees may indeed have impressive temporal abilities. But to demonstrate rhythm abstraction, future experiments should use proportionally scaled rhythms, including gaps, and rule out local-cue strategies. For now, rhythm abstraction in bumble bees remains an exciting possibility — but not yet a settled conclusion.

(For more see news article from Science, Zeng et al.  study, and Comment.)

References

Zeng, Z., Barron, A. B., Peng, F., & Solvi, C. (2026). Flexible, abstract rhythm perception in bumble bees. Science, 392(6793), 93–95. https://doi.org/10.1126/science.adz2894

Ning, Z.-Y., ten Cate, C., Patel, A. D., & Honing, H. (2026). Rhythm Abstraction in Bumble Bees Remains Inconclusive. PsyArXiv preprint. https://doi.org/10.31234/osf.io/m7rph_v2

Isn't musicality more than rhythm?

Last month, we organized a follow-up to the 2014 Lorentz Workshop on Musicality in Leiden, The Netherlands. Twelve years later, it felt both exciting and meaningful to return to Leiden with a renewed focus: spectral percepts. 

While rhythm cognition has received substantial attention over the past decade, key perceptual dimensions of melodic cognition—especially timbre and pitch—remain comparatively underexplored. Many comparative studies still rely on simplified stimuli, such as pure tones, which may limit our understanding of how non-human animals perceive melodic structure. Recent findings suggest that pitch and timbre do not map uniformly across species, inviting us to rethink how these percepts are studied. 

We therefore deliberately shifted attention away from rhythm perception and production toward the perceptual and affective dimensions of melody, harmony, and timbre. In doing so, we revisited Darwin’s idea that animals may not only perceive melodies, but may also take pleasure in them (see workshop proposal). 

What made this workshop especially rewarding was the remarkable diversity of backgrounds and expertise in the room. Researchers from neurobiology, psychology, ethnomusicology, musicology, and evolutionary theory came together to examine the evolutionary and perceptual roles of pitch, timbre, and consonance. This breadth of perspectives allowed us to explore how these percepts vary across species, cultures, and contexts in ways no single discipline could address alone. 

By bringing together such a broad and inspiring group of researchers, the workshop generated new insights, strengthened interdisciplinary collaborations, and laid the groundwork for a more coherent framework on the evolution and cognition of musicality. 

A special issue is planned for Spring 2027, in which we will summarize the workshop’s findings, develop new research ideas, and outline a future agenda for musicality research. 

Photo credits: (cc) 2026 Bas Cornelissen and Lorentz Center.

If musicality did not arise from language, where did it come from?

Recent interdisciplinary advances have transformed the study of the evolution of music. Rather than treating music as a cultural artifact, current research targets musicality — the biological capacity enabling humans to perceive, produce, and enjoy structured sound. Evidence from observations of infants, cross-cultural studies, and neuroscience shows that humans possess innate predispositions for rhythm, pitch, and temporal expectation that arise independently of training. Comparative studies have revealed that components of musicality have distinct evolutionary histories: primate research supports gradual development of rhythmic and audiomotor integration, while convergent traits in vocal-learning species highlight shared biological constraints. Neuropsychological and developmental findings have further shown that musicality is not reducible to language, drawing instead on perceptual, motor, and affective systems that likely predate speech. Collectively, these insights establish musicality as a fundamental cognitive capacity and provide a robust framework for investigating how its components evolved, how they function across species, and why music is central to human life.

But, if musicality did not arise from language, where did it come from? 

[Published in Current Biology as Honing (2026)]
 
Honing, H. (2026) The biology of musicality. Current Biology, 36(5), R177-R180;
Preprint DOI: 10.31234/osf.io/j8x4w_v6;
Drawing courtesy of Marianne de Heer Kloots (mdhk.net).

No progress since Darwin and Spencer?

Darwin and Spencer.

Asif Ghazanfar and Gavin Steingo open their recent Commentary in Science, by asserting that –because no fossil or archaeological record of early music-making exists– modern musicality researchers “rely as much on conjecture as they did in Darwin and Spencer’s time.” 

This characterization is inaccurate. 

The evolution of musicality can be reconstructed using methods from comparative biology, genetics, and cross-cultural analyses, empirical domains that were unavailable to Darwin and Spencer. 

Over the past twenty years, musicality research has shown that virtually all humans have a natural capacity for music (1, 2), comparable to our innate capacity for language. Examples include beat processing in human newborns (3), species-specific precursors of both rhythmic and pitch processing (4, 5), and showing cross-cultural ‘universals’ in the structural aspects of human music (6–8), suggesting a biological basis. Additionally, recent neuroscientific findings indicate that humans process speech and music through distinct — and possibly independently evolved — neural pathways (9). Together, these findings constitute a robust empirical foundation rather than conjecture and have substantially reshaped our understanding of musicality. 

While trained tapping in macaques (10)—as discussed in Ghazanfar and Steingo’s Perspective—addresses only one subcomponent of musicality, it nonetheless offers a valuable window into its evolution, particularly within the framework of the Gradual Audiomotor Evolution (GAE) hypothesis (11). This hypothesis proposes that beat perception and synchronization emerged through incremental increases in the connection between cortical and subcortical motor planning regions. Probing beat perception and isochrony perception in animals is still in its infancy, but it appears, at least within the primate lineage, that beat perception has evolved gradually, peaking in humans and present only with some limitations in chimpanzees and other non-human primates (12, 13). 

Lastly, the relevant object of inquiry here is not music per se, but musicality. For this reason, Ghazanfar and Steingo’s analogy comparing the study of music evolution to ‘human bike evolution’ is unhelpful. Riding a bike requires explicit training even in humans, whereas moving to a musical beat emerges spontaneously and effortlessly, often before the onset of language. This spontaneity is precisely what places beat perception so prominently within musicality research. In other primates, beat perception is not effortless but can be acquired through training, suggesting that for them it is analogous to bike riding in humans. As the authors note, studying trained abilities can nevertheless reveal the basic processes underlying those abilities. More generally, both spontaneous and trained behaviors in animals offer complementary insights into their evolutionary capacities: humans spontaneously acquire speech but can be trained to imitate bird calls, indicating a specialized drive for conspecific communication alongside a broader capacity for vocal imitation. Similarly, non-human primates possess timing and pattern-detection abilities that may form the evolutionary substrate from which human beat induction emerged. Overall, comparative research across cultures and across species provides a powerful framework for uncovering the biological foundations and evolutionary history of musicality. 

As a result, investigating the origins of musicality has become increasingly feasible. What was once a largely speculative corner of musicology has developed into a rapidly advancing interdisciplinary field, rich with compelling new research questions.  

Published as eLetter in Science on December 9, 2025; Written by Henkjan Honing - University of Amsterdam, NL; W. Tecumseh Fitch - University of Vienna, AT; Marisa Hoeschele - Austrian Academy of Sciences, AT; Hugo Merchant - Universidad Nacional Autónoma de México, MX; A preprint is available at OSF.

 

References

1. H. Honing, C. ten Cate, I. Peretz, S. E. Trehub, Without it no music: cognition, biology and evolution of musicality. Philosophical Transactions of the Royal Society of London B: Biological Sciences 370, 20140088 (2015).
2. W. T. Fitch, Four principles of bio-musicology. Philos Trans R Soc Lond B Biol Sci 370, 197–202 (2015).
3. I. Winkler, G. P. Háden, O. Ladinig, I. Sziller, H. Honing, Newborn infants detect the beat in music. Proc Natl Acad Sci U S A 106, 2468–71 (2009).
4. C. ten Cate, H. Honing, “Precursors of music and language in animals” in The Oxford Handbook of Language and Music, D. Sammler, Ed. (Oxford University Press, Oxford, 2025; https://academic.oup.com/edited-volume/59773).
5. M. Hoeschele, H. Merchant, Y. Kikuchi, Y. Hattori, C. ten Cate, Searching for the origins of musicality across species. Philosophical Transactions of the Royal Society B: Biological Sciences 370 (2015).
6. J. H. McDermott, A. F. Schultz, E. A. Undurraga, R. A. Godoy, Indifference to dissonance in native Amazonians reveals cultural variation in music perception. Nature 25, 21–25 (2016).
7. P. E. Savage, S. Brown, E. Sakai, T. E. Currie, Statistical universals reveal the structures and functions of human music. Proceedings of the National Academy of Sciences 112, 8987–8992 (2015).
8. N. Jacoby, E. H. Margulis, M. Clayton, E. Hannon, H. Honing, J. Iversen, T. R. Klein, S. A. Mehr, L. Pearson, I. Peretz, M. Perlman, R. Polak, A. Ravignani, P. E. Savage, G. Steingo, C. J. Stevens, L. Trainor, S. Trehub, M. Veal, M. Wald-Fuhrmann, Cross-cultural work in music cognition: Challenges, insights and recommendations. Music Percept 37, 185–195 (2020).
9. P. Albouy, L. Benjamin, B. Morillon, R. J. Zatorre, Distinct sensitivity to spectrotemporal modulation supports brain asymmetry for speech and melody. Science (1979) 367, 1043–1047 (2020).
10. V. G. Rajendran, L. Prado, J. Pablo Marquez, H. Merchant, Monkeys have rhythm. Science (1979) 390, 940–944 (2025).
11. H. Merchant, H. Honing, Are non-human primates capable of rhythmic entrainment? Evidence for the gradual audiomotor evolution hypothesis. Front Neurosci 7, 1–8 (2014).
12. H. Honing, F. L. Bouwer, L. Prado, H. Merchant, Rhesus monkeys (Macaca mulatta) sense isochrony in rhythm, but not the beat. Front Neurosci 12 (2018).
13. Y. Hattori, M. Tomonaga, Rhythmic swaying induced by sound in chimpanzees (Pan troglodytes). Proc Natl Acad Sci U S A 117, 936–942 (2019).