Cross-Cultural Validity is a Philosophy of Science Problem

In 2010, Henrich, Heine, and Norenzayan produced an insightful, if damning paper exposing the extreme ethnic, cultural, geographical, and social class biases that pervade most psychology experiments. The authors estimate that virtually all participants are ‘Western, educated, and from industrialized, rich, and democratic countries’, or ‘WEIRD’. As the authors point out, nobody wants to publish in the ‘Journal of Personality and Social Psychology of American Undergraduate Psychology Students’, but even today, few scientists have the grounds to claim their findings generalise across socio-cultural-experiential barriers. And despite a somewhat non-establishment disciplinary foundation, music science is just as vulnerable to selectively representative samples as the next field. Nonetheless, diversifying the subject pool is increasingly recognised as a problem, and one that won’t go away on its own.

But what about the rest of the scientific process? In scientific experiments, we often force choices, or attempt to reveal contrasting features between supposedly objective categories – but do our designs reflect something factual, a physical constant, or a belief we hold because, well, everyone else we know believes the same thing? Does the research question itself reflect a universal human experience, or just ours? I don’t think that diversity in the recruitment process is enough. Rather, considering scientific questions from another cultural (or experiential) lense permits the space and detachment we need to determine what is truly ‘objective’ – or, at least, what amounts to a shared subjectivity across human societies.

Note to Abel Tasman and crew: the Maori are not here to jam with you

This isn’t just an issue for the so-called ‘soft science’ of psychology. Take the concept of species. It doesn’t sound like a topic that biologists would do much philosophizing over, until you examine the term in its historical context, and adopt a data-driven, rather than ‘common-sense’ approach. As Darwin put it, the status of ‘species’ is “one arbitrarily given for the sake of convenience to a set of individuals closely resembling each other”, a sentiment that has only gained credibility as time passes, according to the Smithsonian:

The idea of a species is ultimately a human construct. With advancing DNA technology, scientists are now able to draw finer and finer lines between what they consider species by looking at the genetic code that defines them. How scientists choose to draw that line depends on whether their subject is an animal or plant; the tools available; and the scientist’s own preference and expertise.

By the way, this also applies to hard sex/gender and racial categories; in short, they can be useful—alternatively, destructive, and sometimes deadly—for quotidian communication, and capture or drive many of our social structures and group identities. But these classifications are far from universal or elemental, and should be handled with care in a scientific setting, particularly if you don’t take factors such as culture, experience-driven neuroplasticity (e.g., a lifetime of performing the social role of ‘female’ may unsurprisingly make your brain look more ‘female’), and epigenetics into account when explaining between-group differences. And let’s not forget the effects an experimenter can introduce by unintentionally treating her subjects differently, as a result of her unconscious bias and/or personal stake in a specific experimental outcome.

Anyways. Music science. Sure, my colleagues and I are all pretty clear on what happy and sad songs should sound like, or how they differ. In the West, we substitute one for the other and it’s hilarious.

We acknowledge these happy/sad characteristics may not remain constant across culture, context, or the mood one’s in. And maybe that’s what we precisely want to test.

But what about the imposition of a happy/sad binary? And, on another level, is the very concept of emotions (happy, sad, fearful, millenial) appropriate, given the context (and that’s without getting into the hot mess that is the science of emotions)? We could try to avoid these problems by taking an implicit measure of emotion or valence. But even without directly asking our participants about the happy/sad nature of our musical stimuli, the obvious presence of happy/sad categorisation could interact with their experience of the task, resulting in reactivity, i.e., maybe happiness/sadness wouldn’t really be ecologically relevant, if we hadn’t introduced the idea in our design. After all, humans are adept at noticing patterns, even—make that especiallyon an unconscious level.

As soon as we decide we want to test emotions and music-listening, and that we furthermore are interested in happy/sad feelings, we’ve already answered at least two potentially uncontemplated questions before collecting any data. But should we really assume that this basic happy/sad division is a valid dichotomy, or that it’s relevant across musical cultures? Or that emotion as a topic is appropriate here, period? Researchers have taken their music-emotion matching paradigms to geographically and culturally isolated peoples in an attempt to determine universals, and found that individuals across cultures indeed ‘recognize’ the ‘intended emotion’ of a musical excerpt above chance level, an interpretation that typically reflects some kind of a ‘happy-uptempo-simple melody’ or ‘sad-slow-complex’ pattern of results. It wouldn’t surprise me if this really was universal across music functioning in a narrative context, wherein phrasing and expressive timing dominate structurally, because this is also how prosody works. Even chimpanzees and squirrel monkey communication systems share these characteristics. But when I am jamming along to James Brown’s Funky Drummer, for me, it’s about the groove, the nod. I guess I would say that track is ‘happy’ if I were forced to use a likert scale, but this datapoint would grossly misrepresent the social-behavioural function, the affective nuance, the process and sensoriality of funk. It feels good; I don’t necessarily feel happy. And I know this experience has very little in common with my understanding, as a receptive audience member, that Dido’s Lament is … a lament.

Okay, so why even science then, too much postmodernism, facts don’t care about your feelings, you say, stroking your lifesize Jordan Peterson cardboard cutout.

World of Drums fullsize
Peter’s got it covered

A lack of objectivity is precisely why we need to explore cultural and experiential difference. Like it or not, there is way more subjectivity built in to brain sciences than headlines would suggest. We can say with objectivity that, in the modern Western tuning system, the musical note A above middle C is defined by a frequency of 440 Hz. But saying that a song is happy is not objective. Saying that a song is a song is not even objective. I’m not being facetious! The idea of a stand-alone, self-contained ‘song’ that should be performed in a particular way only makes sense for specific times, places, and peoples. This is starkly self-evident to translators, anthropologists, and ethnomusicologists, but if you were from, say, North America and only talked to people like you, and tested out your ideas on other people like you, it wouldn’t be.

Whereas most would agree that emotions and body-states are by definition subjective, we shouldn’t blindly trust the ‘hard’, ‘objective’ concepts (‘song’) that are tossed around, unchallenged, in behavioural sciences, without some point of reference, or at least an attempt at neutrality. For that matter, the same way I don’t know what it is like to be a bat, I don’t actually know what it is like to be happy in another person’s body, even if we share a culture. Or even in my own body, since there are as many ways to be happy as there are instances of the psychological present in a lifetime. I can recognize that some snippet of music is meant to communicate conceptual happiness in a theatrical narrative sense, but that doesn’t necessarily have anything to do with my own feelings. Happiness is a dynamic, amorphous, multifaceted concept, making it shitty for science and great for expression (or bad for expression if you’re a postmodern artist). We approximate and simulate each other’s takes on reality, and that’s as far as language can take us, let alone when we don’t share the same words.  Acknowledging this is not about being politically correct; rather, it’s about not assuming that a socially homogeneous group of Greek dudes from over two-thousand years ago with next to zero equipment and no scientific method serendipitously got it comprehensively right for everyone. Because that is where a ridiculously large number of Western truisms came from, and were filtered down via Islamic Golden Age scholars, towards the Renaissance, the Enlightenment, industrialisation, and modern science as we know it.

Words some faculties are ‘higher’ than others! Five senses! Animals don’t think! [Evil cackling]
For example, Plato and Aristotle believed that vision was the most significant sense (also, ‘the senses’: another hot mess), associating it with reason, the mind, and humanity. In a deforested, urban environment, this is intuitive, but ‘seeing is believing’ might not be the case if you live in a dense jungle and and you can only see four feet in front of your face (for emerging empirical support of this idea, see the recent cross-cultural dismantling of the Aristotelian hierarchy of the senses by Majid and colleagues). Awareness of cultural variation as regards The Senses was charmingly contorted into the same ridiculous 18th- and 19th-century systems of classification I have complained about in previous blog posts, wherein the European way is conveniently also the more cerebral and evolutionarily advanced way, and therefore invokable as justification for slavery and colonialisation. Take, for instance, German proto-scientist Lorenz Oken, who claimed that, by natural order, the European ‘eye-man’ reigned supreme over the Asian ‘ear-man’, Indigenous American ‘nose-man’, Indigenous Australian ‘tongue-man’, and African ‘skin-man’. Unlike peak racist Victorian pseudoscience, I actually really enjoy referring to classical writing, and I’m not suggesting we villanise or erase ancient Greek contribution. What I mean is: thanks for the logic, astronomy, political theory, detailed observation of natural phenomena, and erotic pottery, but maybe no thanks to the mind-body divide, anthropocentrism, and the preternatural valorisation of ‘rational faculty’ at the expense of other important stuff (e.g., prosody, nonverbal social interaction, interoception, motor cognition in general).

Lorenz Oken, obviously a ‘chin-man’

For the record, I think it’s still worth running the happy/sad song experiment, but with the ontological and epistemological machinery driving our research question worked solidly over, and any findings suitably qualified. We shouldn’t just accept fuzzy, emergent linguistic-cultural concepts (‘happy’, ‘sad’; ‘language’, ‘music’) as though they were plainly demarcated, physical properties that will remain constant across time, place, and actors (440 Hz). Yes, for historical and pragmatic reasons, let’s have music science and linguistics; keep neuroscience of speech perception separate from neuroscience of speech production; say that one person researches gesture and the other person researches speech; and create other taxonomies that may reflect our culturally-bound intuitions moreso than function or biology. Just as long as we see the disciplinary boundaries and experimental contrasts for what they are, so we’re less surprised to find that musical and linguistic abilities (and disabilities) tend to overlap; ‘production’ brain areas often activate during ‘perception’ tasks; or that the visual processing of gesture and expression modulates auditory speech processing. Each of these relatively new and startling discoveries may not have seemed unlikely in an alternative scientific universe with less Greek metaphysical influence. For instance, if our society’s worldview essentially did not differentiate between the concepts of ‘dance’ and ‘music’, our research questions (and answers) would almost certainly reflect this in some way. Even the term ‘WEIRD’ reflects what Western academics notice about themselves (‘educated’, ‘industrialised’, ‘democratic’), whereas an outsider might have pointed out ‘height’, ‘materialism’, and ‘physical ineptitude’.

When I became seriously interested in African-American music as a teenager, it completely transformed my world of experiencing and knowing music, sound, feeling and movement – not just in terms of jazz or hip hop, but even the European classical music I grew up studying. Anecdotally, I suspect it may have also altered the way I track time on the scale of seconds, ‘in’ and ‘outside’ of musical settings (that’s right, ‘everything’ ‘gets’ ‘air’ ‘quotes’ ‘now’!). I sought and valued certain features I wouldn’t have recognized previously, and my affordances for music changed, too. For me, this underscores why it isn’t enough to recruit participants from different backgrounds. I wonder what scientific endeavours I’m missing out on, concerning rhythm or timbre or action or some quality I don’t even know about yet. In unsettling ways, cognitive and behavioural science is a philosophical practice, and there is no universally correct answer. but if the problem is that we are limited to our own language, culture, and beliefs that are shaped by experiences both individual and collective, the solution is diversity, not just in terms of participants, but within and across researchers.