In a well-known and generally debunked story, Inuit people have around 50 words for snow. Or so the argument by anthropologist Franz Boas goes. In fact, people who engage with the phenomena of their environments often make distinctions that those who rarely or never engage in the same way with those phenomena don’t. Snow is a salient aspect of boreal environments; it matters whether the snow is packed, or loose, falling or not, just as those in less severe environments make distinctions about rain. Are these terms or just adjectival nouns? How much categorisation of the world is too much, or too little?
The idea that our world is constructed from our linguistic distinctions is an old one, going back to Alexander Humboldt in the nineteenth century (and possibly much earlier, all the way back to the Sophists). Sapir (a student of Boas’) and Whorf (Sapir’s student) proposed a thesis, known obviously as the Sapir-Whorf hypothesis, that our categories determine how we experience the world.This is a kind of linguistic determinism, either of our cognition, or our ontology, of the world. The ruling notion here is that ideas (or terms) specify our experience.
Science finds itself dealing with this issue constantly. Obviously I am most interested in species, but similar issues arise with gene (see Portin 1993), individual (see Wilson and Barker 2017; and note Matt Barker’s comments here), niche (Wikipedia) and even such fundamental terms as ecology, environment, and even development. And that is just in biology. Chemistry has its categorical issues (Scerri 2007), as does physics. Even mathematics has it (Lakatos 1963).
Why is this? Science is supposed, or so I was taught, to be increasingly specific and exact in its terms. Each term is refined and redefined to apply to classes of phenomena that are real, causal, and important theoretically. We know that the social sciences have the problem: terms like religion, society, folkway, class, and so on are routinely held to be constructed kinds, and terms like gender, race, and other identity terms have a strong political overtone as well. And yet, they persist. Why?
There are a few explanations on offer. One is that our language evolved for functional, usually economic, reasons. We identify the categories we do due to something akin to Marx’s notion of false consciousness, in which we speak of the world in ways that are imposed for socioeconomic reasons (which serve the interests of the powerful). Let us call that the Marxian explanation. Another is that we identify the categories that suit our metaphysical, ontological or theoretical worldview. I call that the Weltanschauung explanation. Both have their defenders.
The Marxian account is employed often by those called postmodernists, or structuralists. It suggests that concepts are relative to interests. Change the interests and you change the concepts. Such conceptual relativity gives primary causal roles to the conditions in which ideas develop. It is often used to suggest that truth is a matter of functional coherence in a social sense. The truth of the scientific image (van Fraassen 1980) is not comparable with the truth of, say, the magical thinking of the Asante people in West Africa. They are in their own ways equally “valid”.
The Weltanschauung account implies that our best theory in science, religion or economics (which is neither) licenses the categorisation of the world. In short, as Quine once wrote, to be is to be the value of a variable in some theory. This implies, as Ludwig Fleck (Löwy 1988) and Thomas Kuhn noted (as part of another old tradition in philosophy) that if your theory changes, so too do your commitments to what is in the world. And moreover, these commitments are not comparable. Newton’s use of the term mass is not the same as Einstein’s. This is an essentialistic account of categories: the theory defines the class in terms of theoretical properties. This view is also known as scientific realism and holds that things exist only if they are aspects of our best theory.
Both of these views are set against what has come to be known as Empiricism. The empiricist view, which was widely held until the nineteenth century, despite Hume and Kant rejecting it, was that one merely has to observe the world in order to categorise it. Pierre Duhem, the physicist of the late nineteenth century, attacked naive empiricism in favour of his Kantian view that phenomena are determined by one’s theoretical concepts (Duhem 1954, chapter II), and this became the default view after the collapse of Logical Positivism in the 1950s. The logical positivists held there were two languages in science – one based upon observational operations, and one based upon theoretical concepts and properties. This theory-observation dichotomy failed in the fact of the highly theoretical nature of observations in, specifically, physics. Phenomena were not “ready-made”, but constructed.
Constructionism of categories is thus the consensus among philosophers of science, just as it was among anthropologists and linguists for a time. Sometimes it is social construction, sometimes it is theoretical construction. Empiricists are naive.
In recent years, though, empiricism has been making a comeback. In particular Michaela Massimi (2008, 2011) has argued that while constructive empiricism such as van Fraassen’s fails, the distinction made by Bogen and Woodward (1988) between data and phenomena is correct:
Thus, the metaphysical framework is close to that of experimental realism, whereby (i) phenomena such as weak neutral currents exist in the world ‘out there’; (ii) they manifest themselves by causally producing data such as bubble chamber photographs, which we then (iii) learn how to recognise from other data due to background noise via reliable procedures. (2011, 103)
Massimi offers what she calls “scientific perspectivism” in which
Knowledge claims in science are perspective-dependent when their truth-conditions (understood as rules for determining truth-values based on features of the context of use) depend on the scientific perspective in which such claims are made. Yet such knowledge claims must also be assessable from the point of view of other (subsequent or rival) scientific perspectives. (2016, 13)
Hence, while a perspective depends upon the theoretical issues and rules of a discipline or investigation, it must be cross checked by other perspectives. Massimi holds this is a kind of realism.
If claims to know the reality of things depend on our prior knowledge in this way, by cross checking from other fields and theories, can we ever say that we do really know things? Isn’t it the case that all we know is what coheres with our experiences and our existing knowledge? Are our categories of the world just social constructions? This debate has raged for decades among linguists, philosophers and social scientists. Those who go full constructivist say that each culture, or even each individual, has their authorities, sources, beliefs and religions. In other words, it’s perspectives all the way down.
Those who go full realist, though, want to anchor our categories in hard facts, universally accessible and confirmable. Sure, we have perspectives, but get enough of them together and you converge upon the joints of nature (appealing to Plato’s “cut nature at its joints” comment, Phaedro 265e). Popper once approvingly quoted Churchill about the mathematical prediction of a solar eclipse:
You … look, and your sense of sight immediately tells you that their calculations are vindicated … We have taken what is called in military map-making ‘a cross bearing’. We have got independent testimony to the reality of the sun. When my metaphysical friends tell me that the data on which the astronomers made their calculations [of an eclipse] were necessarily obtained originally through the evidence of their senses, I say ‘No’. [1972, p43]
Such cross bearings are held by Popper among others to converge upon a theory-observation pairing of the world that indeed cuts at the joints. Theory comes to present a structural description of the world that breaks things into their real objects, classes and relations (this is known as “structural realism”, Psillos 1999). This way, we develop a number of categories of things like atoms, fields, orbits and other aspects of the world-as-it-is, even if incompletely or partially. Or so the story goes.
Both of these accounts may be true of some categories, but I doubt they hold for all of our scientific categories. In fact, even when categorical structure converges, it retains a constructed aspect (why wouldn’t it? The terminology of science is hardly discovered “out there”), and no constructed category that is empirically inadequate, as van Fraassen noted, is scientific.
But there is another way we gather our concepts and categories. It is termed “folk” science. For instance, living kinds are usually distinguished by cultures in a fairly predictable fashion, as Berlin and his collaborators have shown (Berlin, 1973, 1976, 1992, Berlin et. al. 1973). According to this view, traditional societies begin with what they term a “unique beginner”, then “life form”, then generic, then specific, then varietal (see Figure 1 from Berlin et. al. 1973). This closely approximates Linnaeus’ kingdom-class-genus-species-variety hierarchy.
Similar folk hierarchies have been proposed by Bulmer and Tyler (1968) and Scott Atran (1985, 1990, 1998, 1999, Medin and Atran 1999). Moreover, folk psychology and folk rules for weather prediction and so forth have similar hierarchies, which itself is a reason why the hierarchies of traditional logic held such sway, as they formalised folk categorisation and practices.
Now, in the case of ethnobiology (or ethnopsychology, etc.), clearly these shared categories are not based on science and theory, and yet again and again such categories have been shown to have close correlations with scientific categories (and of course some surprising differences). Moreover, Linnaeus and his predecessors named species in the absence of much scientific theory that have remained species up to 500 years later. In my book (2009, 2018) I document cases of medieval classifications being “natural”, in the sense of being categories still accepted in modern biology. If empiricism is false, and theory is absent, how did these categories come to be? Is it just accidental? Or are the phenomena ready-made in some non-trivial sense?
By focusing on the theory-observation dichotomy so exclusively, philosophers of science have tended to overlook the phenomenality of categories in both wider culture and science. Initially, when a field has no theory to speak of (other than ancillary theories in related or technical disciplines like optics), the guiding principle for categorisation is experienced observation. To illustrate with an anecdote or two:
A few years back I happened to host a world-renowned American bryophytologist in Sydney, so I took him to the location in the Blue Mountains where Darwin had gone while on his Beagle voyage. I expected naively that he would find the vistas breathtaking and the sense of history would be his focus. Instead, we hardly got out of the carpark, as he found some liverworts in gametophore stages. He thanked me profusely, much to my consternation, for taking him there. I could barely see them except as background plant-like things on the car park embankment.
A year or so later, a well-known coleopterist, specialising in a group of beetles that have many representatives in Australia, visited from the US, and so his host and I took him out into the bush. As we wandered, I was impressed that he not only knew where to look for the beetles, but that he could even see them in the undergrowth and litter. I couldn’t. To him, they leapt out in his field of vision – he had the search image in his head and literally saw them as patterns in the mess. In some ways, I was like James’ newborn child, in a blooming, buzzing confusion of experiences, while he was an expert observer, filtering out the noise to spot the targets. In a later trip I got to see something of how he did it, but not much. I lacked the professional and experience learning.
In many ways, this is like the ways a traditional hunter hunts. A trained hunter sees the prey even when camouflaged or obscured. They know where to look and what to look for. Successes in the past, and folk lore, reinforce those observational techniques that will find the natural types of things being hunted. A forager is the same: they see the indications of useful plants and animals that can be used for food, medicine and cultural purposes. If your living depends on getting the natural world (mostly) right, experience tends towards the right phenomena.
There is not much in the way of theory here. Neither is the observation naively empiricist, nor is it just evolved predispositions, or else I would have had no problems at least seeing the differences that identified these types. This is culturally-scaffolded and experienced observation that converges, out of necessity, on natural phenomena. There are several aspects here. One is that of course observation of any kind relies initially upon our evolved predispositions to respond to certain types, scales, and duration of phenomena. We do not respond observationally to very slow processes, which we tend to normalise as the “natural” state of affairs (akin to the supposed quote of Einstein’s, that “Common sense is nothing more than a deposit of prejudices laid down by the mind before you reach eighteen”). We do not respond to the very rapid either, treating it as noise or static. But at the mesoscale, we do respond observationally to a good many natural phenomena, which are handed to us by our evolved senses, and these can be truth tracking even though they evolved, mostly, as fitness enhancers. When our fitness depends on getting observations right, we evolve truthful senses.
But that is not enough. No single set of observations or the experience of a single person is going to engage with enough of the world to develop what we might call “well-formed categories”. Evolution sacrifices false negatives on the altar of false positives (Wilkins and Griffiths 2013, Griffiths and Wilkins 2014). So a different process aggregates and selects out the experiences of many: cultural evolution. We are trained by our peers, and the expertise of others is passed on, sometimes as rules, yes, and definitions, but mostly as setting up the framework in which individual learning, trial and error, will home in on the “right” categories. Thus, three factors determine observing phenomena: biology, culture and individual experience. Each selects out as much error as it is worth eliding, but the cultural and individual experience is faster and much more efficient at removing error than biology, so long as the success is sufficiently cogent. Finding food, for example, is much more significant than finding a decorative feather.
All that is (conceptually and historically) before science even enters the race. However, it explains why folk taxonomies are often quite robust in the light of the conceptual selection processes of technically advanced sciences; as a first cut, folk taxonomies are not too bad at identifying real phenomena. If they were, people would die, or lose fitness overall.
But ethnotaxonomy is not foolproof; which comes as no surprise, and it can be overturned by cultural factors if the costs are outweighed by the benefits. To give the classical example, let us consider Bulmer’s classical paper: “Why is the Cassowary is not a Bird?” (Bulmer 1967). Bulmer noted that of the New Guinea tribe, the Karam, that
… at the smallest units which Karam discriminate, the ‘terminal taxa’ … Karam show an enormous, detailed and on the whole highly accurate knowledge of natural history, and that though, even with vertebrate animals, their terminal taxa only correspond well in about 60 per cent. of cases with the species recognised by the scientific zoologist, they are nevertheless in general well aware of species differences among larger and more familiar creatures.
However, he said, at the upper level of classification
… objective biological facts no longer dominate the scene. They are still important, but they allow a far greater, almost infinitely varied, set of possibilities to the taxonomist. This is the level at which culture takes over and determines the selection of taxonomically significant characters.
For cultural reasons, the cassowary, which walks on two legs but does not fly, is regarded by the Karam as “human” and thus is forbidden to be killed or eaten. Clearly cultural exigencies overcome the natural in this case. But, and this is not often noted, adopting the arbitrary categorical standards of one’s culture, particularly when it costs you some missed opportunity, is a good way to enhance your standing in the community by demonstrating your commitment, and it raises your fitness (vis á vis social aid) more than it lowers it (in terms of lost protein); so it is not surprising that this occurs.
Consequently, we must expect that while a culture will categorise the environment in which it exists in ways that to a degree track truth (that is, delineate real phenomena) not all the natural categories of a culture will do so. Inuit may have multiple terms for snow based on the ways they interact with snow, but the having of multiple terms doesn’t mean that the categories they name are somehow dividing nature at its joints.
I just used the term phenomenon. What does it mean? The etymology gives us a clue: it comes from the Greek word phainomai, meaning “appearance to the senses”. Leaving aside phenomenological philosophy, which is mostly about the subjective elements of cognition [see comments], in our context, this means a phenomenon is not self-standing. It is a two place predicate: P appears to O. The world has many, possibly an infinite number of clusterings of things. A phenomenon is one that an observer O observes. Now a phenomenon can be something that exists, such as when I note that doors permit egress when open but not when closed. Or it can be something that bears no truthful relation to the world, such as when I see eyes in leaves while on LSD [which the Greeks called phantasia]. Scientific phenomena tend to eliminate the latter kind in favour of the former. A phenomenon in science relies on a good observational system as well as real clusters of properties in the world.
So what counts as a scientific phenomenon is not defined by the categories of folk science/culture/psychology. Instead it depends upon the use of scientific instruments, techniques and methodologies, as well as the use, when available, of theories. If we get 50 words for snow out of that, we can have a reasonable confidence that we are delineating natural properties, given that these operations of science have been honed over time to be successful at exactly that.
There’s much more to say on this, but this post is already too long.
Atran, Scott. 1985. “The early history of the species concept: an anthropological reading.” In Histoire du Concept D’Espece dans les Sciences de la Vie, 1–36. Paris: Fondation Singer-Polignac.
——. 1990. The cognitive foundations of natural history. New York: Cambridge University Press.
——. 1998. “Folk biology and the anthropology of science: cognitive universals and the cultural particulars.” Behavioral and Brain Sciences 21 (4):547–609.
——. 1999. “The universal primacy of generic species in folkbiological taxonomy: Implications for human biological, cultural and scientific evolution.” In Species, New interdisciplinary essays, edited by R. A. Wilson, 231–261. Cambridge, MA: Bradford/MIT Press.
Berlin, Brent. 1973. “Folk Systematics in Relation to Biological Classification and Nomenclature.” Annual Review of Ecology and Systematics 4 (1):259–271.
——. 1976. “The concept of rank in ethnobiological classification: Some evidence from Aguaruna folk botany.” American Ethnologist 3 (3):381–399.
——. 1992. Ethnobiological classification: principles of categorization of plants and animals in traditional societies. Princeton, N.J. ; Oxford: Princeton University Press.
Berlin, Brent, et al. 1973. “General Principles of Classification and Nomenclature in Folk Biology.” American Anthropologist 75 (1):214–242.
Bogen, James, and James Woodward. 1988. “Saving the phenomena.” The Philosophical Review 67 (3):303–352.
Bulmer, Ralph. 1967. “Why is the cassowary not a bird? A problem among the Karam of the New Guinea highlands.” Journal of the Royal Anthropological Institute 2 (1):5–25.
Bulmer, R. N. H., and M. J. Tyler. 1968. “Karam classification of frogs.” The Journal of the Polynesian Society 77 (4):333-385.
Duhem, Pierre M. M. (1954 ). The aim and structure of physical theory. Princeton, Princeton University Press.
van Fraassen, Bas C. 1980. The scientific image. Oxford: Clarendon Press.
Griffiths, Paul E, and John S Wilkins. 2014. “When do evolutionary explanations of belief debunk belief?” In Darwin in the 21st Century: Nature, Humanity, and God, edited by Philip Sloan. Notre Dame, IN: Notre Dame University Press.
Lakatos, Imre. 1963. Proofs and refutations. London: Nelson.
Löwy, Iiana. 1988. “Ludwik Fleck on the social construction of medical knowledge.” Sociology of Health & Illness 10 (2):133-155.
Massimi, Michela. 2008. “Why there are no ready-made phenomena: what philosophers of science should learn from Kant.” Kant and Philosophy of Science Today, Royal Institute of Philosophy Supplement 63:1–35.
——. 2011. “From data to phenomena: a Kantian stance.” Synthese 182 (1):101–116.
——. 2016. “Four kinds of perspectival truth.” Philosophy and Phenomenological Research, http://dx.doi.org/10.1111/phpr.12300.
Medin, Douglas L., and Scott Atran, eds. 1999. Folkbiology. Cambridge MA: MIT Press.
Popper, Karl R. 1972. Objective knowledge; an evolutionary approach. Oxford: Clarendon Press.
Portin, Petter. 1993. “The concept of the gene: Short history and present status.” The Quarterly Review of Biology 68 (2):173-223.
Psillos, Stathis. 1999. Scientific Realism: How Science Tracks Truth. London, New York: Routledge.
Rosch, Eleanor, et al. 1976. “Basic objects in natural categories.” Cognitive Psychology 8 (3):382-439.
Scerri, Eric R. 2007. The periodic table: Its story and its significance. New York: Oxford University Press.
Wilkins, John S. 2009. Species: a history of the idea, Species and Systematics. Berkeley: University of California Press.
——. 2018 forthcoming. Species: the evolution of the idea. 2nd ed. Boca Raton, FL: CRC Press.
Wilkins, John S., and Paul E. Griffiths. 2013. “Evolutionary debunking arguments in three domains: Fact, value, and religion.” In A New Science of Religion, edited by J. Maclaurin and G. Dawes, 133–146. Chicago: University of Chicago Press.
Wilson, Robert A. and Barker, Matthew. 2017. “The Biological Notion of Individual.” In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta. Metaphysics Research Lab, Stanford University.