Historical linguistics is basically what everyone thinks linguists do. Actually, that's not quite true. People who have heard of linguistics as a field of study, think that it involves historical linguistics. Most people, on hearing "I'm a linguist", immediately ask which languages you speak. Which, let's be fair, is a valid question even when linguist means studier-of-languages, not student-of-languages; we are perhaps more likely than the average English speaker to be interested in other languages.
However! There is an idea that the study of language involves figuring out where words come from, how languages are related to one another, and how they've changed. This is, indeed, what linguistics historically involved, though it is no longer the exclusive, or even main, field of study. But it is very interesting, and more accessible than any other facet of the subject - except perhaps infant language acquisition.
Here are my notes on the first three chapters of the book.
Chapter One: Introduction
In which the book is introduced. Onwards!
Chapter Two: Sound Changes
What are the different kinds of sound changes that happen over time?
They can either be regular (applying everywhere the phonetic conditions are met) or sporadic, applying unpredictably to only a few words.
They can either be conditioned or unconditioned. Conditioned changes are contingent on neighbouring sounds, where the sound is in a stressed syllable, etc, whereas unconditioned changes apply to all instances of a sound.
Changes can also be either phonetic or phonemic. Non-phonemic changes may also be called allophonic changes - the total number of phonemes does not change, nor which phoneme is used in a given word, but only the pronunciation.
They can either be conditioned or unconditioned. Conditioned changes are contingent on neighbouring sounds, where the sound is in a stressed syllable, etc, whereas unconditioned changes apply to all instances of a sound.
Changes can also be either phonetic or phonemic. Non-phonemic changes may also be called allophonic changes - the total number of phonemes does not change, nor which phoneme is used in a given word, but only the pronunciation.
Another way of classifying phonemic changes is as mergers and splits. Mergers are where either two phonemes are replaced with a third, new one, or an existing phoneme is substituted for all examples of a current one.
An example of the latter type is happening in modern British English: words such as cure /kjʊə/ are pronounced by younger speakers to rhyme with 'your': /kyɔː/.
Mergers cannot be reversed. Children acquiring a language have no way of knowing that two sounds were originally one, so they treat the results of a merger as the same sound.
Another example: British English "wh". Older British speakers - and other dialects of English - pronounce words such as "what" and "when" starting with /ʍ/, "hw-", whereas for most British speakers, this sound has merged with /w/, "w", making the wh- spelling unpredictable.
Splits follow mergers. In general, new sounds are not introduced to cause a split, but sounds that used to be allophonic varients are no longer predictable after a merger. For example, English used not to have the sound /v/, only /f/. /f/ was voiced between vowels, so words like "leaf" /liːf/ became /liːvəz/ "leav-ers". Then the /ə/ at the end of the word was dropped, and so the /v/ was no longer between two vowels: /liːvz/, leaves. So - along with some borrowing of French words, which had /v/ already - English acquired a contrast between voiced and voiceless /f/: that is, between /v/ and /f/.
An example of the latter type is happening in modern British English: words such as cure /kjʊə/ are pronounced by younger speakers to rhyme with 'your': /kyɔː/.
Mergers cannot be reversed. Children acquiring a language have no way of knowing that two sounds were originally one, so they treat the results of a merger as the same sound.
Another example: British English "wh". Older British speakers - and other dialects of English - pronounce words such as "what" and "when" starting with /ʍ/, "hw-", whereas for most British speakers, this sound has merged with /w/, "w", making the wh- spelling unpredictable.
Splits follow mergers. In general, new sounds are not introduced to cause a split, but sounds that used to be allophonic varients are no longer predictable after a merger. For example, English used not to have the sound /v/, only /f/. /f/ was voiced between vowels, so words like "leaf" /liːf/ became /liːvəz/ "leav-ers". Then the /ə/ at the end of the word was dropped, and so the /v/ was no longer between two vowels: /liːvz/, leaves. So - along with some borrowing of French words, which had /v/ already - English acquired a contrast between voiced and voiceless /f/: that is, between /v/ and /f/.
Thirdly, we could talk about assimilation and dissimilation.
Assimilation is where sounds gain properties of their neighbours - like their location, their manner, their voicing. Dissimilation is where the opposite happens, and is much rarer, and usually sporadic.
An example of dissimilation is "chim(b)ley" as a pronunciation of chimney - the /n/ becomes an /l/ or /bl/, losing its nasal manner, dissimilating it from the preceding /m/.
Assimilationa and dissimilation can either be total (gaining/losing all features of the neighbour) or partial (gaining/losing only some); contact (with an immediate neighbour) or at a distance; and progressive (following a neighbour) or regressive (before a neighbour; to the left as it is written).
Assimilationa and dissimilation can either be total (gaining/losing all features of the neighbour) or partial (gaining/losing only some); contact (with an immediate neighbour) or at a distance; and progressive (following a neighbour) or regressive (before a neighbour; to the left as it is written).
A list of some common types of sound change. (There are subtypes given, but I shall gloss over them):
- Deletion (usually of a vowel)
- Epenthesis (insertion, often of a vowel)
- Rhotacism (changing /s/ or /z/ into 'r')
- Metathesis (swapping the position of two sounds)
- Haplology (deleting a repeated sound e.g. pronouncing 'library' as 'libry')
- Voicing and devoicing (especially intervocalic voicing, i.e. between vowels, and word-final obstruent devoicing)
- Nasal assimilation (the nasal is articulated at the same place as a following stop e.g. in-possible = impossible)
- Palatalisation (place of articulation moves to the palate before front vowels, e.g. /s/ becomes /ʃ/, 'sh',. Or, a secondary articulation is added - basically /j/, 'y' is inserted)
- Diphthongisation / monophthongisation (a single vowel becomes two, or vice versa. A diphthong is e.g. 'I' - to say it slowly, you must say ah-ee, it isn't a single sound)
- Vowel raising / lowering
- Nasalisation (of vowels, around nasal consonants)
- Lenition (weakening) / strengthening (this concept is not very rigorously defined, and deserves an entire post of its own. Or lecture course, as the case may have been.)
- Gemination / degemination (doubling a single consonant. This does not exist in English; think Italian 'notte')
- Affrication / spirantisation / deaffrication (changes between stops, affricates and fricavites. Fricatives used to be called spirants.)
- Lengthening / shortening (including compensatory lengthening, when another sound is lost)
What is a chain shift?
Basically, one sound moves, and forces a whole string of others to do the same. There are two types: pull (or drag) chains, and push chains.
Pull chains are where a sound leaves a gap when it moves, which another sound moves to fill, leaving another gap, etc.
Push chains are where a sound moves too close to another, forcing it to move away.
Chain shifts rely on the idea that sound systems tend to be, well, systematic.
Sounds are distributed in articulatory / acoustic space in a non-random fashion, to facilitate communication.
If you only have 3 vowels, for example, they will be at the extreme edges of the high/low and front/back grid, maximally spaced from one another.
An example of a chain shift is the Great Vowel Shift in English, which happened somewhere between Chaucer and Shakespeare (1400 and late 1500s).
Chapter Three: Borrowing
What kind of things can languages borrow? Words are the most obvious, but also:
- Sounds can be introduced, lost, kept or shifted in pronunciation based on contact with other languages
- Phonological rules can be borrowed e.g. which syllable in a word is stressed
- Sound changes: if a language undergoes a sound change, a neighbour might too
- Calques, or loan translations, in which elements of a phrase are translated literally (e.g. courriel, courrier électronique, a Canadian French calque of e-mail, electronic mail). This can also be a 'semantic' borrowing, where a word has two meanings in one language, and another language borrows the metaphor e.g. Spanish estrella, star, has borrowed the English meaning of 'movie star'.
- Words can be "emphatically foreignised" - Campbell gives as an example the pronunciation of "Azerbaijan", with the un-English stress on "jan", and "j" being pronounced as /ʒ/.
Why are they borrowed? Need or prestige:
We have borrowed words for chocolate, veranda, and zebra because English didn't have words for them before they were encountered. Very few words are made up wholesale when they can be borrowed.
We also borrowed the French words for beef and pork when we already had the Anglo-Saxon cow and pig because of the prestige associated with speaking like the ruling Normans. (Notice who had the word for the animals, and who the word for the meat...)
And on a humorous note, I learnt that the French word for "ceiling fan" is "vasistas". From the German "Was ist das?" - What is it?
Some outstanding questions:
- When and how can sound changes be borrowed?
- Why do people prefer to borrow words than invent new ones?
- Can languages borrow syntax as well as phonetic and phonology?