As many folks may experience, a sound sleep may be disturbed by a brain chasing some weird thoughts during the night. Mariner experienced this phenomenon last night when he was disturbed through the night by his brain struggling with language.
This time it was the troubles humans have because they must use the same sounds over and over again for different words because the human limits on making different sounds cannot handle the billions of words humans have invented. So humans have to use the same sound for many different words and meanings. Just a few simple one-syllable examples:
Cow, now, sow, plow, mean, bean, lean, tree, flee, see, flea, sea, etc.
If one sounds out the vowel letters in the alphabet (a,e,i,o,u and sometimes y, which is redundant to i), that’s about the limit of different noises a human can make. Humans try to stretch these sounds so they can invent more words. That is called slurring, or respectfully, dialect.
So humans use five sounds over and over and over again by attaching distortions of the throat, mouth, tongue, teeth, lips and jaw. These noises are called consonants.
Twenty-six letters constitute all the letters one is supposed to need to make words. But it isn’t that easy thanks to the Great Vowel Shift that occurred between 1400 and 1600. The shift made it easier to invent more words with the same vowel noises. Check these examples:
plough as well as plow, slough as well as slough – tricky, one uses the ‘ow’ sound, the other uses the ‘u’ sound; the other direction to make more words with no more vowels is to combine words, allowing vowels to be used more than once in one word. For example, tie dye as a simple one and Hippopotomonstrosesquippedaliophobia is a complex one that means ‘fear of big words’, something about which Icelandic and Welsh languages have no concern.
What would it be like to be a crow and have to develop an entire vocabulary based on ‘caw’ noises. They have, you know.
So mariner can understand the problems ChatGPT (which stands for Chat Generative Pre-trained Transformer) has trying to capture nuance and relevancy in human languages.
The dialect piece can be entertaining. Jimmie Carter and Fats Domino never end a word with the ‘ur’ sound; mariner’s friend from Boston puts ‘ur’ noises wherever he likes.
Thinking further, one wonders how the vowel noises are represented in sign language, There are only so many gyrations of the face, arms and hands.
Ancient Mariner
I don’t even know where to begin. It seems you are talking about English. Do you have any idea how many sounds are used in human speech? I don’t know exactly but there are hundreds. For example, take the vowels. They can be nasalized, whispered, unvoiced, made breathy, have many tones added, and glottalized. Take the consonant P. It can be glottalized, aspirated, unreleased, and velarized. These are just a few. I have no idea how Chat GPT works, but each language will have to work out its use of that system on its own. I don’t have space to discuss the Great Vowel Shift, but it was a doozy. All the extant English vowels were raised and if the vowel above it (in the tongue position) had nowhere to do, it was turned into a diphthong.
Well spoken, Robert. Where would our vowels and consonants be without aspiration slurring by the lips and all those other skeleton parts? You note that consonants also can be slurred with skeletal manipulations. An abstract example of mariner’s premise: The first wolf, Canus lupus appeared about 1 million years ago. Today there are several varieties of wolves and countless variations of dogs – even the wiener schnitzel. But they are all variations of lupus, that is, vowels and consonants constrained by the human genome.
While my insights were derived from a troublesome night’s sleep and somewhat simplistic, still I believe we are close in our definition of the issue – if only I were a world renowned philologist.
Great retort!
Skipper