All natural languages change over time in a number of ways.
Some do it faster, some do it slower,
but no natural language is totally immune to change. If you're aiming to create a
naturalistic language, the best way to ensure your language looks and behaves like a real language is to simulate this kind of change yourself
One avenue of change is through phonological evolution, when the pronunciation of sounds in a language slowly drifts.
This occurs primarily through historical sound changes. These occur when a certain sound and languages inventory
undergoes a change Where it is pronounced differently in some or all environments.
Fundamental to this definition is that it affects ALL
instances of the sound in the described environment across the entire language in all areas.
It pays no attention to the grammar or parts of speech.
If you have a sound change that says that [a] becomes [e] before nasals then that means that ALL
instances of [a] before nasals becomes [e], no exceptions.
This is your opportunity to shape the language's sound to get the exact phonetic character you want.
Languages with similar or even the exact same sound inventory can be made to sound completely different
based on the sound changes that occur and how their sounds are distributed.
Recall that at this point in our language. We have purely open syllables. How can we go about introducing some more complex syllable shapes?
Very often in rapid speech some sounds can get skipped over, and, over time, they may get worn down and lost completely.
This very often happens with vowels in unstressed syllables.
(I mean, when was the last time you actually pronounced it 'EV-E-RY'.)
Let's create a sound change in which vowels are lost between voiceless obstruents in unstressed syllables.
So 'Akato' would become 'Akto'.
Voila! Suddenly we have consonant clusters.
We've taken a purely CV language and turned it into a CVC language.
Note that because vowels were only lost between voiceless obstruents,
only voiceless obstruents can serve as codas.
It's critically important to mention that word internal vowel loss usually does not
affect stressed syllables, since those are the ones you're putting all the emphasis on, and so you're less likely to skip over them.
Another sound that is very frequently lost is [h].
It's a pretty weak sound so many languages tend to lose it, as in French, Spanish, and more than a few dialects of English.
Let's say in our sample language that [h] gets lost in between vowels. So, 'ahi' would become 'ai'.
Now, in deleting H, we've now got a whole load of vowels coming into contact with each other
where there was a consonant separating them before.
Now they could stay as they are, but frequently vowels in contact with each other undergo a number of changes,
most commonly combining to form diphthongs, or two vowel sounds expressed in the same syllable.
A simple and very common way that this can manifest is for an unstressed [i] and [u]
To go to [j] and [w] respectively when next to another vowel
So 'kuhani' became 'kuani' when the /h/ was lost, and then became 'kwani'.
While 'ihaku' became 'iaku', which became 'yaku'.
The same thing can happen if the [i] or [u] is the second vowel on the pair.
So, what was 'ahi' became 'ai' became 'ay',
And 'ahu' became 'au' which is now 'aw'
What if we want to add some more sounds to our language?
Some sound changes have the effect of producing new sounds that weren't in our proto-language.
For example, remember we don't have any voiced obstruents in this language so far.
Well, what if we say that the voiceless stops become voiced between vowels?
This kind of process is called Assimilation,
in which a sound will change to be more like the sounds surrounding it.
When a voiceless stop is sandwiched between two voiced sounds like vowels,
it's quite likely to become voiced to be more like the vowels that surround it.
This is extremely common in basically all languages from all parts of the world.
In fact, it's a standard feature of American English pronunciation.
There are two important things to note here:
Firstly, notice that the order in which these sound changes are implemented matters.
If we make all voice stops voiced first and then get rid of vowels between voiceless obstruents,
We arrive at a very different result than if we did the reverse,
one where stops can't occur in clusters at all.
Make sure to keep track of the order in which the sound changes occur and implement them accordingly.
But there's something else we have to consider;
It might seem like it but actually by voicing stops between vowels,
we haven't added anything to our sound inventory.
Why?
Because we've simply changed how sounds are pronounced depending on environment rather than creating something new
In other words, there is never a point where the distributions of the voiced and voiceless variants overlap
Voiced stops can only occur between vowels and voiceless stops can only occur in clusters.
You can always predict which variant will be used based on the environment.
In linguistics, this is called Allophony,
When a single phoneme, or sound used in making meaning distinctions, can surface as two or more different allophones
depending on its environment
So, if we were okay with this, then we could just keep it how it is,
But if for whatever reason we really really wanted to add voiced stops to this language. How can we fix this?
Well, we'll have to change the distribution of one or both variants so that they overlap in some cases.
Take a look at these clusters. They consist of two voiceless stops in a row.
These sorts of clusters are comparatively rare in the world's languages because
1. The coda position is the hardest part of the word to hear
and 2. Stops are the hardest sounds to distinguish acoustically making it comparatively difficult to hear coda stops
compared with other coda sounds
think about the words 'set', 'sept' and 'sect'.
If there were no context clues
It would be a little difficult to differentiate between them in rapid speech. As such, many languages don't allow stop clusters at all,
and in those that do, they might not last very long.
There's a couple of ways these clusters might simplify.
Number one, the coda stop could become reduced to a glottal stop
and therefore all voiceless stops would have a glottal stop as an allophone in coda positions which would minimize the number of forms that need distinguishing.
Or the coda stop could change to become identical to the following stop, producing something called a geminate,
sssentially a consonant pronounced for twice the length of a normal one.
Think about the English words 'unnamed' and 'penknife'
(nobody says a 'unamed' or 'penife').
Another possibility is dissimilation
This is the opposite of assimilation, where a sound becomes more different from its neighbouring sound to make it easier to hear.
Dissimilation isn't nearly as common as assimilation and is most likely to occur
if, for whatever reason, your language really needs to distinguish between these particular sounds that they deem as being too similar.
So if we have a stop cluster, the coda stop could dissimilate from the following stop by becoming a fricative.
This exact sound change happened in Dothraki.
Again, this type of change is quite rare and should be used sparingly.
Finally, and most simply, the coda stops could just be dropped entirely.
This is exactly what happens in certain varieties of British English. For example, in many parts of England,
This word isn't 'recognize', it's 'reco'nize', because it's just a little bit too difficult to catch that coda /g/,
So it's just left out completely.
If we choose this option,
suddenly we've reintroduced voiceless stops between vowels again.
Now the distributions of voiceless and voiced stops overlap and the voice variants
officially qualify as new phonemes and we can add them to our sound chart.
It is also quite likely that this sound change will cause the preceding vowel to lengthen,
since the time that used to be used to pronounce a vowel and two consonants is
Now occupied by a vowel and just one constant, so the vowel may lengthen to fill in the missing time,
a process called compensatory lengthening.
And speaking of compensatory lengthening, that's something that might happen if these coda /h/s are lost.
Again, H is quite a weak sound, so when it occurs before another consonant, it may disappear completely.
If this is the case, it may trigger compensator a lengthening of the previous vowel,
which would also keep our stress system intact. /h/ may also disappear when immediately following another consonant,
since this is pretty hard to enunciate properly when it occurs.
If it's lost in this environment it may cause compensatory lengthening of the previous consonant, triggering gemination.
Let's implement a sound change in which /h/ is deleted and whatever sound comes before the /h/
undergoes compensatory lengthening, either a short vowel to a long vowel or a single consonant to a geminate.
Now /h/ can't occur between vowels,
in codas, or following another consonant, meaning the only place it can occur is the very beginning of a word.
It is possible that /h/ could remain in this state,
In fact is actually not too far off from how it is in English,
But at this rate it may be lost from the language entirely.
This actually happened in many of the Romance languages, and it's not just /h/ that's liable to be lost.
Many languages go through stages of ditching sounds that are a little too hard to distinguish.
Languages lose sounds just as often as they gain them.
So by implementing this change we've introduced geminates to our phonology, but lost one of the sounds from our proto-language.
Now let's implement another round of vowel loss, this time between nasals and stops.
Once again, the vowel loss does not affect our stressed syllables, but something else has happened;
Take a look at this word.
There's something about it that makes it look distinctly non English,
probably the fact that it's got an M right next to a D. In English this almost never happens.
In fact in most languages it almost never happens.
Why?
Well, [m] is a labial sound, produced with both of the lips,
whereas [d] is an alveolar sound, produced behind the teeth.
To produce an [md] consonant cluster, we have to start at the lips and then pull the tongue back to pronounce the [d],
Which is quite a complex maneuver, so most languages will make it easier by changing the [m] to a [n],
Which is pronounced in the same place as the [d] making it much easier to pronounce.
This sound change is called 'nasal assimilation', and happens in almost every language that allows nasal codas.
In English, when we take the word 'possible', and stick on the negative prefix 'in-' we get
'Impossible', where the [n] changes to an [m] to match the following [p] in its place of articulation.
So for our language,
Immediately after the evolution of nasal codas, we implement nasal assimilation,
Meaning that any nasal coda will go to [m] before [p] and [b],
[n] before [t] and [d] and the velar nasal [ŋ] before [k] and [g].
The velar nasal is the same sound as the final sound in the English 'sing',
But notice once again that the sound is not phonemic, as it only occurs before Velar stops where the sound [n]
can never occur, and therefore it is just considered an allophone of [n].
What about including some more sounds? What if we wanted to create a whole new column in our sound chart?
For that we'll use something called 'palatalization'.
See front vowels like [i] and [e] are pronounced near an area of the mouth called the palate.
But if a front vowel gets paired with a consonant that's produced either just in front of or behind
The soft palate, the vowel may cause the consonant drift further towards the palate
until the sound takes on a whole new place of articulation.
So for example in the syllable /ki/ the K is a velar sound, produced far back in the throat,
But the /i/ is pronounced at the palate, and so over time, the vowel may start to drag the velar sound forward towards the palate
Going from /ki/ to /chi/.
This can also affect alveolar sounds like /ti/ to /chi/ or /si/ to /shi/,
Which explains how the Latin '-tion' suffix got turned into [ʃən] in modern English,
or why in English, French, and Spanish, the letter C is pronounced
like [k] before an /a/, /o/, or /u/, but as [s] before an /i/ or an /e/.
Palatalization is extremely common.
Some languages palatalize any sound that comes before any front vowel
and had a huge inventory of palatal consonants like Russian and Irish Gaelic
but other languages are a bit more selective.
A common place for palatalization to occur is in dipthongs
If the first and the pair is a [j], the previous consonant may be palatalized and the [j] deleted
This happens in English too: in many British dialects the words 'tune' and 'Duke' are
pronounced as 'chune' and 'juke'.
The advantage with this form of palatalization is that it automatically makes the new palatal sounds phonemic as their
distribution cannot be predicted, since the vowels that caused the palatalization get deleted.
So let's go ahead and add a palatal series to our language.
Let's say that when [t] and [k] occur before a diphthong beginning with a [j], the [j] is deleted and the
[t] or [k] becomes [t̠ʃ],
and, under the same circumstances, we can say that [s] becomes [ʃ].
Notice that the palatalization also turns the stops into affricates rather than into other stops.
There is a palatal stop, but it's actually much rarer than its corresponding affricate,
and this kind of palatalization frequently turns stops into affricates anyway, so we'll just stick with affricates at this point.
Whenever you add new sounds to your language, you need to consider your romanization system again.
How are you going to romanize these sounds in a way that makes sense?
For our sample language, we can simply use the English digraphs
/ch/ for [t̠ʃ] an /sh/ for [ʃ], since that would be pretty intuitive and [h] no longer exists in our language,
so there's no possibility it could be misinterpreted as a consonant cluster.
Finally, let's get rid of all word-final short vowels, and shorten all word final long vowels.
We'll also say that this same change causes word-final diphthongs to simplify so that [aj] and [aʊ]
Both become [a:].
This happens pretty often, since it's just easier to make fewer distinctions at the end of a word.
This change will mean that now any consonant can serve as a word final coda.
This also has the side effect of breaking our stress system.
Which syllable is stressed is likely to stay the same before and after the loss of final vowels,
meaning that there are now words stressed on the second to last and last syllables.
So now we need to keep track of where the stress on a word is.
Following our stress rules before word final vowel loss, if the vowel in any given words
second-to-last syllable was long then it received stress.
So when the final vowel was lost now the final syllable is stressed.
So we can amend our stress rules to say that now if a word's final syllable is long and ends in a consonant,
Then it receives the stress.
Notice how there needs to be a final consonant for this rule to apply.
Because old word-final diphthongs were reduced to long vowels without altering the word's stress.
If the word was previously stressed on the third to last syllable then after vowel loss
It will be stressed on the second-to-last syllable
UNLESS the word ended in a long vowel or diphthong,
in which case the final syllable wasn't lost so it still receives the stress on the third to last syllable.
So if a word that ends in a consonant doesn't have a long vowel in the final syllable,
then it is stressed on the second-to-last syllable.
Let's also say that word final vowels are not lost when they are preceded by a consonant cluster,
meaning that word final consonant clusters are not permitted. It doesn't have to have this way, but we'll say it does here.
How many sound changes you implement simply depends on personal choice and how long of a time depth you're dealing with.
You could theoretically keep applying sound changes indefinitely until you get the language to a place you like.
For this sample language. I'm going to stick with this pretty small selection just to demonstrate the principles at work here.
Write down a list of all your sound changes in the order in which they occurred and stick to it.
Now we need to apply these changes to all the words in our proto-language to get the forms for the modern language.
Once you've applied them, you'll notice that you've created a set of phonetic constraints that delimit what your words can sound like.
Compared with our original syllable shape of purely open syllables,
We now have a language where nasals and voiceless fricatives and affricates, but not stops are allowed to serve as codas,
except at the end of the word where any consonant can serve as a coda.
It's important to note that all of these sound changes have a phonetic motivation behind them,
that is, there's a reason why they occur the way they do, most often because it makes things easier to say.
It's not a matter of choosing any random change you like if it doesn't have any phonetic motivation to it.
A great way to avoid Infeasible sound changes is to look at sound changes that have happened in languages in the real world to get an idea of what's possible.
Now we know how our sounds are going to change over time,
It's time to talk about how our grammar is going to change over time.
Join me next time when we talk about evolving new grammar, and this is where you really get to have some endless fun.
For more infomation >> No Make a Mistake - A Fé Move Montanhas (WebVídeo) - Duration: 4:27. 
For more infomation >> Several schools make high school football schedule changes ahead of upcoming weather - Duration: 0:58. 
Không có nhận xét nào:
Đăng nhận xét