*hackles raise instinctively*Frislander wrote: ↑12 Nov 2018 22:27Because I'm an exemplar theory man, who believes "phonology" is merely an emergent property of a multitude of remembered phonetic word tokens, so while ideas like "phonemes", "allophones", "features" and so forth may be useful in describing patterns in languages, they do not necessarily reflect an actual underlying mental representation.
OK, I'm sorry to have to go all philosopher here, but I really do.
A linguist saying that phonemes (etc) are 'complete shit' because they "don't necessarily reflect an actual underlying mental representation" is like a physicist saying that gravity is complete shit because gravity doesn't necessarily reflect an actual world-thought in the supreme mind of the All-Lord Saurgothmor. Three objections necessarily spring to mind:
a) nothing in the theory of gravity requires there to be a true representation of the actual world-thoughts in the supreme mind of the All-Lord Saurgothmor;
b) so far as anyone in physics or theology has been able to demonstrate, there are no such things as actual world-thoughts in the supreme mind of the All-Lord Saurgothmor, and if there were we'd have no robust way to determine what they were;
c) the existence, and if then necessary the properties, of actual world-thoughts in the supreme mind of the All-Lord Saurgothmor, are matters for theologians, as physicists have neither the empirical tools nor the theological training to explore the non-Euclidian interior hypostasis-sphere of the All-Lord's unfathomable thought-essence, and are probably not even employing commensurable concepts of truth.
Likewise, the people who should be building theories of what is and isn't 'an actual mental representation' are neuroscientists, guided by psychologists and evolutionary biologists, under the watchful supervision of philosophers and associated experts (logicians, information theorists, etc). Not linguists.
And likewise, there are, so far as we can tell, no such things as "actual mental representations", and if there were we wouldn't have any way to determine what they were; this is a superstition, albeit a popular one. What's more, the entire concept of "actual" representations is naive and incoherent - put simplistically, "actual" (as opposed to non-actual) is an absolute and objective property, but a "representation" is a subjective, relative, and insubstantial object. It's like saying "the cuboid immorality" or "the spherical irony". How many representations are actually in one landscape photograph, and what are they actually representations of?
And likewise: nothing in the idea of a phoneme requires any sort of 'actual mental representation'. A phoneme is a concept in science - it is an element in a narrative representation of the world that we employ for economic advantage, and gains its value as a scientific concept from its utility within that representation. The validity of the phoneme-concept rests upon its ability to enable people to make predictions about observable linguistic behaviours - which is to say, from the extent to which those who employ the phoneme-concept are able to conduct empirical behaviours without encountering unpleasant surprises. Likewise, for example, the validity (and limitations) of the 'brick' concept may be tested by seeing whether people who believe in bricks do or do not meet with unpleasant surprises in their attempts to construct tall buildings. The property "is (or is not) reflective of an actual mental representation", on the other hand, is not a scientific concept but a metaphysical one. It is therefore not only something that cannot be tested by science, but something that is irrelevent to science, even if it is a coherent an truth-bearing concept in its own right (which this is not). In the same way, whether quarks 'represent a true thought in the mind of God' (or the All-Lord Saurgothmor for that matter) is not only untestable, but unimportant to the particle physicist. Whether a green dot partakes in the universal property of greenness, or imperfectly instantiates an ideal of greenness, or exhibits a singular trope to which the name 'green' has been given by virtue of an intertropical resemblance, is entirely irrelevant to the question of the validity of the concept of red-green colourblindness. Empirical theories are, by the fact of their being empirical, necessarily metaphysically agnostic.
And in that way, the fact that something is "emergent" is by no means an indicator that it is "complete shit". Other concepts that describe emergent properties with no, or questionable, actuality include "harmonic resolution", "emotion", "Frislander", "fun", and of course "brick". Nonetheless, these concepts are not complete shit, but rather play important roles within their own specialised forms of discourse.
Come now, you're a smart guy - let's not wander back into the poorly-lit middle-ages, to argue about angels on pinheads and suchlike!
EDIT: sorry, not trying to be aggressive. Just, occasionally something riles my philosophical instincts. Imagine seeing someone explain that, actually, indo-european descends from tamil. You know you're not actually going to change their mind, but still...
[no, I'm not saying frislander's naivity is on that level exactly, it's just the principle of the thing]