From 74774.157@compuserve.com Thu Jun 8 15:10:19 1995 Date: 08 Jun 95 13:06:25 EDT From: Jeffrey Henning <74774.157@compuserve.com> To: BlindCopyReceiver: ; Subject: Copy of: MODEL LANGUAGES, 6/1/95, 1 of 2 MODEL LANGUAGES The newsletter discussing newly imagined words for newly imagined worlds ------------------------------------------------------------------------ Volume I, Issue 2 -- June 1, 1995 INVENTING A LANGUAGE FOR NAMING PEOPLE AND PLACES "My name is Alice, but-" "It's a stupid name enough!" Humpty Dumpty interrupted impatiently; "What does it mean?" "Must a name mean something?" Alice asked doubtfully. "Of course it must," Humpty Dumpty said with a short laugh: "my name means the shape I am -- and a good handsome shape it is, too. With a name like yours, you might be any shape, almost." -- from Lewis Carroll, _Through the Looking Glass_ Despite Humpty Dumpty's comment, Alice could not be just any shape -- her name actually summons forth an image of someone who is simple and proper, according to surveys conducted to determine the impressions people have of different names. All names have perceptions attached to them. Etymologically speaking, Alice's name is from the Greek for "truth". Most American and European names have become simple labels, their original meanings forgotten. How many people realize that a name like _Jeffrey Henning_, if translated literally, means "Godfriend Meadowlark"? Meanwhile, Indian names like "Dances With Wolves" (to take a bad example) wear their etymologies on their sleeves. If you are fascinated by the origins of names, then you will be happy to learn that a naming language is one of the most useful types of model languages to create -- and one of the easiest, making a great first language for the hobbyist. A naming language can be less complex than other model languages, since it does not need a detailed grammar and since it can get by with a small vocabulary: with just 150 words (revealed below), you can generate millions of names for imaginary people and places. Once you've read this issue, you'll be able to create two or three naming languages in as little as a half hour, though you'll end up fascinated by your creations and will spend many more hours on them. To begin creating any type of model language, you must be able to create words in that language. To create words, you need to understand sounds, meaning, sound change and so forth. This issue will introduce you to the basic aspects of language; subsequent issues of _Model Languages_ will explore each one in more depth. LANGUAGE CHANGE The vocabulary of languages is constantly changing, as technology changes and as our understanding changes. Twenty years ago no one talked of faxes, PCs or being on-line. No one had heard of perestroika. Things were still groovy, nizza, happening. Besides adding and retiring words, languages put new spins on old words: _gay_ now primarily refers to "homosexuality", not "happiness"; _liberal_ now is almost a curse, referring to "favoring governmental power" when it once meant "favoring governmental power to promote social progress". These word changes are not surprising. Any of us can look over the linguistic landscape of our lives and see how the terrain has changed. If you project this forward a thousand years, it is easy to see how the shape of a language's vocabulary will go through major upheaval. It's harder to see that the grammar of the language, the way we put words together, will change too. While saying _hopefully_ is still frowned upon, it is no longer viewed as completely ungrammatical. The pronoun _them_ is often used to refer to one person, rather than the plural it is formally meant to refer to; in casual conversation and writing, _them_ is now the gender-indifferent alternative to _he_ or _she_ (incidentally, as it was four hundred years ago, before pedantic grammarians -- yes, _them_ -- stepped in). Looking a thousand years out, other grammatical distinctions will have been leveled, revealing new horizons behind them. Finally, it can be hard to realize that the very sounds we use for words change. It's not hard to believe the occasional word changes, such as knowing that _cup board_ is now pronounced _cupboard_, the [p] sound having assimilated to the following [b]. It is harder to believe that English words that now begin with [p] and date from Indo-European all began with [b] in Indo-European times. Such systemic changes, where a sound changes throughout the entire vocabulary, happen gradually. To imagine how it happens, think of a dialect, such as the Bostonian's "idear about whether the cah is pahked in Hahvahd yahd". Sound changes systematically when these dialectal differences become emulated and become the new accepted pronunciations. Imagine an alternate universe where JFK served out 8 years as the U.S. President, and was succeeded by 8 years of RFK, who was followed by 8 years of Teddy (it had to happen in some universe!). No doubt in that universe the Bostonian accent became American English's new standahd. Basic sound changes do not happen suddenly like earthquakes buckling the landscape, but gradually like water eroding a shoreline. Language change is for the most part slow, since change is on the whole discouraged. The whole point of language is for people to be able to make themselves understood to each other, and this happens best in an environment where the language changes no faster than the land at the water's edge. Language change is important because it shows the best way for you to invent a model language -- by making changes to an existing language (whether natural or a model). AN ANCESTRAL LANGUAGE -- THE GRANDMOTHER TONGUE Every person alive today has or had a mother. Similarly, every mother tongue spoken by all these people had an ancestral language that it evolved out of. Even Proto-Indo-European, the reconstructed ancestor language of hundreds of European and Indian languages, had an ancestral language it evolved out of: Nostratic, which some linguists hypothesize was also the ancestor to five other proto-languages. Since Nostratic itself is most likely descended from another language, records of the first language are no more knowable than records of Adam. The ramifications for the language modeller are that the language he or she creates should not spring fully armed from the head of Zeus like Athena, but should derive from its own parent language. Most model languages are unknown orphans, when a pedigree would not have been hard to provide. Tolkien is one of the few modelers to actually create an ancestor tongue, which he used to derive many different Elvish languages for _The Lord of the Rings_, of which the best known are Quenya and Sindarin. "Wait a minute," you might be thinking, "are you saying that to create a model language I first have to create another model language? Where does that language come from? When does it end?" Tolkien again provides the best example; he created root words in a _proto-language_; he imagined that the elves would have reconstructed their ancestral language, much as Europeans reconstructed Indo-European. Proto- languages are elaborate hypothetical constructions and, as hypotheses, are fuzzy around the edges: nothing but the bones of an extinct dinosaur, while the exact color of its flesh can never be known. A proto-language, therefore, can be a simpler form of model language. The benefit of creating a proto-language is that it makes it easier to create sister languages to the model language you are chiefly interested in (what, more languages?!), enabling you to formulate new words based on regularly sound changes (more on this in it a minute). It also makes it easier to coin words in your desired model language, providing a rich system of root words to use to derive new words. So creating a proto- language can save you time. The easiest way to save time on your first model language is to use an existing language as the proto-language. I once worked on a science fiction story set aboard a colony whose original settlers had been 20th- century Italians and Spaniards, who -- through centuries of living together -- had created a new, simpler language. By using Italian as the ancestor language, with many borrowings from Spanish, I not only made it easier to create a new language but I taught myself some Italian and Spanish as well! If you are writing about a story that has taken place in the last 10,000 years and is set in Europe or India, you might even use Proto-Indo- European as the ancestral language for your languages. Check out _The Roots Of English_ by Robert Claiborne for an easily readable discussion of Indo-European roots, or check out the appendix to _The American Heritage Dictionary of the English Language_, published by Houghton Mifflin; both works are biased in emphasizing those roots from which English words descended, but make good starting points for devising a language. SOUND To create your language, you need to decide which sounds you want speakers to distinguish. Basically, while it would be easy to think that the sound [t] is exactly the same, [t] actually describes a range of sounds, all closely approximating one another. The way you position your tongue when saying [t] will vary depending on what other sounds you say before or after it, but we both articulate [t] similarly enough to recognize it as the same thing. There is no objective reference that says a language must have any particular sound. For instance, Old English did not distinguish between the sounds [f] and [v] or [s] and [z]. The plural of [hoof] was pronounced [hoovz] but it was not until later times that speakers treated the \f\ sound in the singular as different from the \f\ sound in the plural. In Old English times, there could be no word [vat] different from [fat] -- such a distinction was just not made. Gradually, the sounds came to be heard as distinct. So when creating the sounds of your language, you need to realize that they will only approximate English sounds, not exactly match them, and might not reflect distinctions currently made in English. The [hw] sound in _whale_ might be regarded by your speakers as the same as the [w] sound in _wail_ (yes, they are different sounds, but you might have to listen closely as you pronounce them to tell the difference). You can certainly include in your language sounds that are not part of English, say the French vowels, typically pronounced with the lips rounded, or the expectorating [kh] of Hebrew and Yiddish, let alone the clicking sounds of the Hottentots and Bushmen. However, you should refrain from having too many unusual sounds in your language; you want your readers to be able to pronounce your words without too much difficulty. Simply having regular sounds combined in unique ways (e.g., _sretan_, or _tsedet_) will be enough to convince them it is a unique language anyway. Languages are very strict about how sounds are combined. English, for instance, allows words to begin with [sn-], but never [zn-]. The rules English uses could fill pages, but as a modeler you want to just hint at complexity. You may want to have a combination that is unusual in English and make it frequent in your language: for instance, have some words begin with [sr-], [kn-], [kth-], [tl-], but here again restraint is the order of the day. As you specify how sounds can be combined, you may want to outline valid syllables. Your language might only allow syllables of CVC (Consonant+Vowel+Consonant) or just CV or VC. Some languages, like Japanese or Korean, have very strict limits on how syllables can be formed, making it possible to list all the valid syllables of the language. But where Hawaiian allows just 162 different syllables, Thai has 23,638 syllables. Two languages can have the exact same consonants and vowels and yet sound very different, depending on the syllable patterns and on the frequency of the consonants and vowels. You may want to list the sounds that occur most often. By paying rigorous attention to this when developing the proto-language, you can relax a little more during creation of the descendant language, which will carry on many of the same frequency patterns, though applied to different sounds as the sounds change. Many languages have very simple vowel systems. Eskimo-Aleut has just three vowels (the smallest number ever observed), while Spanish and Japanese each has five vowels. The typical language has between 5 and 7 vowels, but Indo-European languages usually have more; English has 12, and German has 14. The African language Khoisan has the record with 24 vowels. Languages have been observed to have anywhere from six consonants (Rotokas) to 95 (Khoisan), with an average of 22.8 consonants. The typical language has twice as many consonants as vowels. The most common consonants include [p], [b], [t], [d], [k], [g], [gh], [f], [s], [sh], [m], [n], [ng], [gng], [w], [l], [r], [j] and [h]. For a great discussion of the sound structure of languages, check out _The Cambridge Encyclopedia of Language_ by David Crystal. SOUND CHANGE Over time, sounds gradually change in certain circumstances. John F. Kennedy, like many Bostonians, would drop his last [-r] from words like [car], while adding an [-r] to _Cuba_ [cubar] and _idea_ [idear]. As alluded to before, had enough Americans adopted this, it would have been considered a regular sound change and many other words might have undergone this change. Or listen to the dialect of Brooklyn, where [bird] becomes [boyd], for instance; someday all English speakers might pronounce [ir] as [oy]. No doubt, through the rise of one dialect in Old English, the sound [sk] was gradually becoming [sh]. Over great periods of time, these changes become more pronounced. Literally and figuratively. Here are some common ways consonants evolve into one another: b <---> gw b <---> p b <---> v ch <---> kw d <---> g d <---> t d <---> th f <---> p f <---> v g <---> d g <---> k g <---> w g <---> y g <---> z gu <---> gw gw <---> b gw <---> d gw <---> g gw <---> gu gw <---> k gw <---> ku gw <---> kw gw <---> v gw <---> y gw <---> zh h <---> hy h <---> k h <---> s h <---> y hv <---> hw hw <---> hv hw <---> kw hw <---> p k <---> g k <---> gw k <---> h k <---> kw k <---> s k <---> th kh <---> kw ku <---> gw ku <---> kw kv <---> kw kw <---> ch kw <---> gw kw <---> hw kw <---> k kw <---> kh kw <---> ku kw <---> kv kw <---> p kw <---> sh kw <---> t l <---> r p <---> *- p <---> b p <---> f p <---> hw p <---> pf pf <---> p r <---> l s <---> h s <---> k sh <---> kw t <---> d t <---> th t <---> z th <---> d th <---> k th <---> t v <---> b v <---> f v <---> gw v <---> w w <---> g w <---> v y <---> *- y <---> g y <---> gw y <---> h y <---> z z <---> g z <---> t z <---> y zh <---> gw *- (lost) This list is not meant to be all inclusive, just representative of changes that occurred in Indo-European. Likelihood Of Sound Change # Of IE Languages Where IE Initial Consonant Changed gh 12 gw 12 gwh 12 bh 11 dh 11 kw 11 g 9 w 9 k 7 b 4 d 4 s 4 p 3 t 2 y 2 l 1 r 1 m 0 n 0 You can use the above table as a rough guide to determine which consonants are more likely to undergo change. It is not representative of all languages, being an analysis of 12 languages descended from Proto-Indo-European and showing the number of languages where the consonant in the word-initial position changed. The languages analyzed were Armenian, Avestan, Common Germanic, Greek, Hittite, Latin, Lithuanian, Old Church Slavonic, Old Irish, Old Persian, Sanskrit, and Tocharian. The nasals, [n] and [m], are fairly stable, as are the liquids [l] and [r]. The stops [p], [t] and their voiced counterparts [b] and [d] change in only a third of the languages. All aspirated consonants changed in every language analyzed, being markedly unstable; [k] and [g] and their glide forms [kw] and [gw] were also more likely to change than not. Sound changes actually vary by position, with a sound change applying to different places -- the [s] might become [h] at the beginning of a word, [k] in the middle of a word and [z] at the end of a word (though this is an extreme example). For simplicity's sake, you may just want to apply the same changes regardless of position. Besides these phonetic changes, there are often "environmental" changes in words, where sounds change because of the sounds they are near. The following examples illustrate the major types of sound change. Assimilation -- Regressive or anticipatory, a sound is influenced by the following next sound: English [cupbord] became [cubbord]; the word _assimilation_ is itself an example: Latin __adsimula_-re_ became _assimula_-re_, since [ad-]_ regularly assimilated to [as-]_ before the [s] sound. -- Progressive, a sound is influenced by a preceding sound -- Coalescent or reciprocal, when two neighboring sounds influence one another: _don't you_ becomes pronounced [donchu] Dissimulation -- sound moves away from the pronunciation of neighboring sound: French _marbre_ became English _marble_ as the second [r] became dissimilar from the first. Split - a sound becomes regarded as two distinct sounds, such as Old English \s\ compared to Modern English \s\ and \z\ (Old English's failure to distinguish between the sounds is one of the reasons many Modern English words are written with 's' when [z] is pronounced) Metathesis -- two sounds change places, _third_ from Old English _thridda_ Elision -- sounds are omitted (elided) in rapid speed, often dropping a consonant from a cluster of consonants: [cubbord] became [cubord]; elision specifically refers to loss of an unstressed vowel or syllable: _elementary_ becomes pronounced [elementry] when the final schaw sound is elided. Loss -- a sound disappears from the language altogether, as the velar fricative, a variant of /h/ (and the final sound of Scottish _loch_), did in English, with only a vestige remaining in English spelling: the common silent 'gh' of English words like _light_, _night_, _sight_, which were once pronounced [likht], [nikht] and [sikht]. Haplology -- the loss of a sequence of sounds because of similarity of neighboring sounds: should this ever be called _haplogy_ it will have undergone haplology itself. Syncope -- the loss of medial sounds, as _boatswain_ lost the [t] sound as it was shortened to _bosun_ ([bosun] is the correct pronunciation of _boatswain_, by the way, never [bo_-tswa_-n]). Apocope -- the loss of final sounds, as in the silent 'e' in words like _love_ and _hate_; of course, the silent 'e' used to be pronounced. Liaison -- introduction of a sound between words, as in French when the silent final consonant of a word is pronounced when the next word begins with a vowel. Prothesis -- introduction of an extra initial sound, as occurred in Spanish and Old French, which frequently inserted an [e] sound before an initial [sp]: for instance, Latin _specia_-is_ became Old French _especial_. Epenthesis -- introduction of extra medial sound, as Old English _bre_- mel_ became Old English _braembel_. You can quickly generate more than one language by inventing different sound change rules for each language. So perhaps the Dilbertian [d] becomes [t] in Dogbertian, whereas it becomes [th] in Dinobertian. Or take a look at how the names James, John and Katherine have evolved in seven different languages: English James John Katherine French Jacques Jean Catherine German Jakob Johann Katharina Italian Giacomo Giovanni Caterina Spanish Jaime Juan Catalina Swedish Jakob John, Johan Karin, Katerina Yiddish Dzheymz Yohan Katerine Source: _Webster's Third New International Dictionary_ Names vary idiosyncratically and do not always evolve according to the regular sound changes that affect other words. Thus the English towns of _Luton_ and _Leyton_ are -- despite their differences -- both derived from the same word, _Lygetun_, "farm by the river Lea" (the river Lea, incidentally, may either mean "bright one" or may represent the name of a river god, _Lugus_). Names get shortened frequently; for instance, _Johann_, _Giovanni_ and _Yohan_ all indicate that there used to be an [a] sound after before the [n] in _John_ and that the silent [h] in _John_ used to be pronounced, and still is in German, Swedish and Yiddish. ---- CONTINUED ----