Currently working on my own I have decieded to talk some about orthography and various ways a language can be written, the options and general guidelines.
This is where graphs represents ideas, concept, items and other concrete things. They do not show any pronounciation or anything but merely give of an idea. Their major problem is the "failure" of being able to draw abstract items or grammatical things. How would one draw "love" for example? Most people think a heart but is that really how that culture percieve it? Is it really a good way? Words such as "And" "up" "down" "above" etc, words giving grammatical information, are also a huge problem because they are not easy drawing even as a detailed picture. This is usually solved by using a combination of symbols.
- Person + tie + Person = Marriage
- Sun + Earth = Up
- Sun + Earth + Thing = Above
and so on, the options there are endless and usually somewhat logical, another problem that do occure is even more integrated grammatical roles such as conjugations, declensions etc cannot be expressed with this and is hence often reserved for Analytical/Isolated languages where everything is 1 single word. Japanese solved the problem by having a second system soley for conjugations and declensions though.
Side notice is that all writtings starts out as Logographs.
Symbol Count: Over 2000
Syllabary is one where all graphs represent a possible syllable in the language in question. This kind of script is perfect for a language that has very few possible syllable combinations.Sych languages being for example (C)V(C), (C)V or V(C), these doesnt take too many syllable forms and offers few signs per word as most words wont be bigger than 4 syllables on avarage. The problem occures when you got a language like English where the structure is (C)(C)(C)V(C)(C)(C) which of course offers a huge range of syllables where its impossible to keep track of all of them and hence a syllabic writting is not a good idea. One criteria for a Syllabary is that between each and every symbol there must be no common symbol tieing for example all syllables with "k" in them into one.
This can though be used for large syllable structured languages by for example having what I call an "empty" vowel where a vowel is written out but not pronounced, saying "a" is empty "Flinsa" would be written as "Falinasa" and the first two a's are not pronounced while the last one is, this is just a memory thing like anything we are used to in spelling especially english.
Symbol Count: 40-200
A personal favorite of mine is the Abugidas, they share characteristics of syllabary and alphabet in a way providing the best of two worlds. They are syllabary in the sense each graph represent a vowel and a consonant but they are alphabet in the manner that all syllables with [k] sound in them are represented by a common symbol. The consonant is always the base symbol and vowel is a form of diatric added, a common thing with many real life abugidas is that the lone consonant has an inherent vowel, like [a] while all others are diatrics and even the removal of the vowel requires a diatric. In my own script, which is to be an abugida, is all vowels diatrics and lack of vowel is the natural base. The benifit of this script is that its always easy knowing the sounds from the symbols you see as all vowels and consonants are the same in each symbol, [ki][ka][ku][ko] will all have a symbol in common and all [na][da][pa][fa] will all have same diatric in common, it also offers the possibility of writting any language structure while playing a slight favor toward smaller syllable structure.
Symbol count: < 40 consonants and < 20 diatrics
an Abjad is in a way similar to Abugida in the way that the consonant is the base graph but unlike them it doesnt mark out in anyway the vowel. only consonants is written out as in many languages consonants play a more important role than the vowels due to their numeric superiority in any given language. They fit more for languages where the consonant dependence is even heavier such as arabic and hebrew where the fundament of everything is a three set of consonants that tells you if its about books, feet, people etc and where vowel only serves a secondary role of manipulating the main meaning, this makes an abjad suitible as context can take care of the rest. These have the benifit of being very fast in reading, less to look at and more. The disadvantage is that in languages where vowels play a more vital role in differens between words meanings and more they cannot give enough information for where context to take over and provide the rest. This is one of the reason older Abjads evolved into alphabets by the greeks as greek structure made an abjad impossible to use.
Arabic and hebrew does have ways to signal about vowels but they are secondary and unneccisery in most situations.
Symbol count: < 40
This is the one we are all used to using, an alphabet makes itself unique in the way that each graph represent ONE sound, its either a vowel or a a consonant and represent one sound of one kind. The first true real alphabet is often said to be the greek one which came from ancient abjads. But no alphabet truley hold that one graph is one sound, vowels are typicly the ones most excepted to this rule as they differ usually more than the consonants. It is more that each graph represent a sound of meaning as each sound of meaning can be a variation of multiple sounds which are all intepritated the same, an example is that in swedish "rs" is often said as "sh" instead without problem or lost of meaning. Digraphs might be used to supply additional sounds to not get a too wide alphabet.
Symbol count: <50
There are properties an orthography can have and others are neccisery, here is a few.
We are all so used to go from left (L) to right (R) in rows then up (U) to down (D), which I will call LU since it starts at left going in rows then moving downward from up, UL would be start in columns from up to down then move from left to right, that we sometime forget that there are other options, there are 10 options to be precise, what the additional 2 is I will tell soon. You got LU, RU, LD, RD, UR, UL, DR, DL. Though humans seem to usually seem to prefer going from up to down so in most cases the ones where you start from the bottom in any sense is eliminated, even though they are possible, leaving you with 6, LU, RU, UR, UL.
There is another 1, 2 if you count the "forbidden" one, which is acctually where direction is constantly switched, called Boustriphedon. You start at the top and writes to the right side then on the next row you write to the left side and on the next row again you go right, back and foruth. A common feature in these writings is that the symbol is mirrored in the direction which is switched and its facing depends on in which direction youre going on that particular row. Egyptians did this in their writting.
No I do not mean individual appearences of graphs but this means where shape differ depending on where it is in a word. This of course means it doesnt apply to Logographs and most likely not syllabaries either. In english we are used that all letters look the same no matter their placement that we might think thats how it should be. Though our Upper case is the original form and lower case is intended to ease writting which is also a shape feature that most languages dont share and hence should be taken into consideration. The one I was thinking on though is when the shape is depended on if its Isolated, In the begining, in the middle or the last letter. Arabic is a typical example of this. Hebrew shares a few letters like this but its not general through out their entire abjad like Arabic is. The benifit with this is that the shapes are adapted to ease writting making it quicker.
This is a huge problem for people in general, the question you should ask yourself is this first. Are you after a natural/realistic orthography or just want a perfect one for writting? If the answer is natural/realisitc you should read on otherwise this is pointless for you. Completeness refers to how accurately can your writting represent the acctual speaking of the word? This includes both meaning changing differenses and non-meaning making differenses though usually it is thsoe that changes meaning that is taken into account. All latin languages usually use more vowels than they acctually type. An example is swedish "Fred" which can be pronounced as [Fred], then meaning peace, or [Frɛd] which is a male name. Both are spelled the same and use the same symbol for the vowel but is clearly different meaning and pronounced differently. All systems that is natural has this inherent errors as the script often changes slower than acctual speaking hence cant keep up. Linear B which is an old greek writing didnt include the differens between voiced/unvoiced, aspired/unaspired which it had in speaking nor vowel lenght. You would write "A-to-ro-qo" to mean [anthro:kwo:i]. Alot is missing but context and prior knowledge fills the gap.
Orthography also may lack tone, questions etc, while containing things spoken language do not. There is no punctuation in spoken language. No one would in their sane mind say "I like her, dot." or something to mark the end of a sentence, they speak on so everything appears to be one giant clump of text with no meaning or begining. There are tones in a language, english writing has none of the sort. There are so many things one can add or substract into the writing because context takes care of it all.
This is something to think on as it is often quite important in most writtings and languages. Proper nouns are nouns refering to ONE single entity, London, Alise, The great war etc. In latin writing languages they are distinguished by using upper case for the first letter, in egypt they had a circle around it, does a language make a distinction or not?
Tips and tricks
We all need them =)
- Decied what kind of writting you want first
- Decied properties you would like it to have
- Decied what properties it should have and what it should lack
- Do not be scared of lacking certain things, keep in mind humans work mostly on context and it fill out most holes without problems
- Use a standard form of shape and figure that is modified rather than a random bunch that seems to have no relation to one another, p,b,q,d for example are turnings of one another
- Take time, no script is perfect from the start
- Always try writting them togather once you get a bunch because they all need to be togather before you can truley see how well they fit.
I think thats it for now, cant think of more =) Might update