This week’s blog post includes a linked audio file. Just click on the link below if you would like to hear the post read aloud. Scroll down to read the text.
We take what we know (declarative knowledge) and we make something out of it (procedural knowledge), and if we keep doing that thing enough times, it becomes part of who we are (automaticity).
Who says cognitive science isn’t poetic?
Whether I am processing visual information or the sounds of a language, exemplar theory tells me that what I know is computational in nature. I am a circuitry of memory banks functioning on the juice of high frequency and recency of exposure. We are all lean, mean, statistical machines, unconsciously crunching the numbers of our input, and creating hills of central tendency for the regularities, and valleys for the outliers.
Empirical research in perception tells us that we are born into this world already attuned to its linguistic rhythm. The sound of a mother’s voice is conducted through her bones shaping her fetus’s perception; her child is born not only in sync with the regularities found in his mother’s tongue, but even showing preference for her voice over anyone else’s.
It’s almost endearing, until you realize that Guinean baboons can also detect statistical patterns in the input, that cotton-top tamarin monkeys can perceive prosodic cues, and that rhesus macaques can hear melodic differences in music. Perceiving differences in the acoustic stream is easier than child’s play, heck, even the birds can do it.
So… what is man? A machine or a monkey with a nice set of articulators?
Language is commonly thought to be the thing that separates us from other mammals, but not for phonologists. It is our perceptual magnet effect, the aural-neural network within a speaker’s mind that becomes ‘warped’ in response to language input (Kuhl, 2000), that truly separates us from the primordial pack.
In this perceptual model, similar sounds cluster together because they share an acoustic ‘family resemblance’, and the best tokens among them emerge as prototypes that unyieldingly attract all similar sounds towards them. And while some sounds of a feather flock together, others that are deemed dissimilar are banished to the wings of the category’s boundary or form other sound categories altogether. Over time, these compartmentalized sound-clouds, or phonemic categories, flesh themselves out, becoming a connected network of sounds, or the phonological inventory of the language that we speak.
For some researchers, though, we don’t speak a language at all, we only speak a variety or dialect of it. Sound categories, much like humans, are not fixed entities—they like to travel. A category of sound can ‘split’ into separate categories or two categories can ‘merge’ into one; this is the phonological basis that differentiates dialects from each other.
In the communities of phonetic practice that I frequent, for example, many of us merge our mid-central vowels, making merry-Mary-marry phonetically identical. These variant productions of a word are often referred to as shibboleths, pronunciations able to mark members as insiders or outsiders of a particular linguistic group; in other words, acoustic markers of ‘us’ versus ‘them’.
This brings us right back to one of exemplar theory’s most striking insistence: the mind is quick to categorize any incoming data. Quite suddenly, one feels as if there’s no harm in asking: Where are you from…really? After all, it’s just a categorization thing that the brain is hardwired to do.
Exemplar theory is quite clear about what such a question means: it seeks explicit information about which ‘family’ the token resembles. In order for smooth processing of the input, critical information needs to be supplied to the perceiver.
Unfortunately, members of ambiguous categories, such as myself and so many others, fail to meet the centrality requirements that prototypical people seem to unquestionably possess. Upon fine-grained inspection, when my membership to any one category is weighted and graded to a particular ‘family’, I am unable to enrich a singular prototype; that is simply because by the theory’s own criteria: I am non-prototypical. Or maybe, I’m just not your prototype, baby. Whatever the case may be, this disorienting effect on the perceiver causes them to forget that no matter how non-prototypical a penguin may appear, it nevertheless remains a bird.
Exemplar theory is fundamentally premised on a notion that I believe we should all ruminate on:
There is no such thing as neutrality in perception.
Because we are born with ‘warped’ minds, and because our perceptions continue to twist and turn in frequency with the input that surrounds us and with we choose to surround ourselves by, being in a state of neutrality is something that we have to consciously work at—but we’ve got skills, baby!
We are extremely talented navigators of variation. We can accurately perceive and process a speech signal even in the noisiest of environments, regardless of how different each speaker we meet sounds.
The proof is in the acoustic pudding. What’s been turning our heads since we were infants and causing us to suck and salivate at the mouth have been noticeable differences in the acoustic input. It is the variation that naturally exists in our world which drives our learning, because without it, our ‘warped’ perceptions would simply flatline into monotony.
Unlike songbirds, we can learn each other’s song, so sing your heart out in your particular birdsong, and I will warp my perceptions to parrot a reply back to you. I mean, it’s not like birds can speak two languages, right?
Kuhl, P. (2000). A New View on Language Acquisition. Proceedings of the National Academy of Sciences of the United States of America, 97, 11850-11857 http://dx.doi.org/10.1073/pnas.97.22.11850