rickharrison.com → Artificial Language Lab
(9th draft, ©2001)
1.1 synopsis
This essay advocates the creation of an international auxiliary language (IAL) for the world according to relatively objective criteria, and discusses how such criteria can be specified.
1.2 premise
When people who speak different languages need to communicate, they often choose to use an “interlingua” or “lingua franca.” Some examples: Latin served as Europe’s scholarly interlingua for several centuries. Citizens of Israel, who have many different mother tongues, use deliberately-revived Hebrew as their common language. Malay-Indonesian is the national language and common means of communication in Indonesia, Malaysia and Singapore, even though it is the “native” or “home” language of a minority of their people. Pidgins – simplified versions or mixtures of languages – often arise to bridge the gap between groups of people who need to communicate across a language barrier.
The adoption of an interlingua for the entire world appears to be desirable. Such a language would make it easier for scholars and scientists to gain access to information, and to participate in conferences and journals. A world lingua franca would facilitate commerce and make it easier for businesses to participate in the international marketplace. Diplomats, tourists and others would benefit. The costs of providing translations for items ranging from computer software instruction manuals to movies and novels would be minimized, and the potential market for such items would become global rather than regional.
It is possible that a “natural” language such as English could play the role of auxiliary language to some extent. However, natural languages (and particularly English) have some characteristics which make them difficult to learn: unpredictable spellings, sounds that are difficult for people from different language backgrounds to pronounce, irregularities in the derivations of related words, etc. (Edward Sapir offers an especially eloquent and convincing argument along these lines.{1}) Many persons of average intelligence do not have the time and/or special talent needed to master English (or some other natural language) as a second tongue. They can learn to “get the gist” of texts and to have simple “how are you” conversations, but cannot expect to gain the ability to confidently express themselves after a moderate amount of study. Also, there are political and cultural objections to granting any natural language the status of world interlingua.
For these reasons the intentional creation of a suitable planned language seems desirable. This essay explores the idea of creating the best possible design according to relatively objective criteria, rather than proceeding in a whimsical or unplanned way, although we must understand that perfection is impossible. Complete consensus among all interested parties is also impossible because many of the people who support the idea of an IAL are firmly attached to existing projects.
In order to determine what features an optimal IAL design would possess, we must first identify the main goals of the project, and then examine all aspects of language design and choose those options which are most suitable in view of the goals. This requires us to set aside personal preferences, as much as is humanly possible, and proceed in an objective manner. I propose that the following three goals be considered the prime, paramount considerations in evaluating IAL design possibilities:
1. An optimal IAL will be relatively easy for most children and adults to learn as a second language.
2. An optimal IAL will have the ability to handle both mundane conversation and highly technical information.
3. An optimal IAL will be culturally neutral; it will not provide advantages of word recognition or other special favors to one or two ethnic groups at the expense of all others.
These three goals are listed in order of declining importance. If there seems to be a conflict between item 1 and item 3, item 1 takes precedence. Other considerations in language design, such as the desirability of brevity versus the desirability of redundancy to prevent errors, questions of aesthetic appeal, and so forth are interesting and worthy of consideration. However, adding more items to our list of criteria would increase the number of conflicts between the criteria and reduce the likelihood of agreement or acceptance among interested persons. Therefore I suggest that we limit ourselves to the three goals listed above.
1.3 the “ease of learning” question
The question of what makes a language easy to learn, and for whom, is a complex matter. However, it does seem to be possible to objectively specify some of the characteristics which make languages easy or difficult for most people to learn.
There are strong indications that some languages are more difficult than others, even for children learning them as native languages. According to Slobin, an average Russian child requires several more years than an English child to master the morphology of his own language. {2}
Rick Morneau wrote: “Ease-of-learning is definitely measurable. A lot of work has been done that, although not aimed at (language) design, can certainly be applied to judging whether a language is easy to learn. As one of many examples of such work, several studies have been sponsored by the Foreign Service Institute of the U.S. State Department about the problems encountered by teachers in teaching certain languages to students of different linguistic backgrounds. For instance, Chinese students have a difficult time with English tenses, because they are not used to having to indicate tense all the time (in Chinese, tense is usually indicated only when it cannot be determined from context). Conversely, English speakers have difficulties with Chinese tones, because they are not used to using pitch contours to make phonemic distinctions. Yet English speakers have no difficulty at all in learning to use verbs without marking them for tense.
“Language educators would agree unanimously and without reservation that it always comes down to this: a language becomes more difficult to learn when the student must learn to make distinctions that he is not used to making. That’s why students have difficulties with things like accusatives, mandatory tenses, tones, noun/adjective agreement, honorific inflection, consonant and vowel harmony, mandatory gender distinctions, polysynthesism, ad nauseam.” {3}
Jacques Guy commented: “There definitely are languages that are simpler than others, and by a long shot, too. If you don’t believe me, just try learning Navaho, or French, for that matter. If learnability is one thing we are looking for, we ought to examine those simple, that is, easily learnt, languages, and draw lessons from them. How do you tell them? Easy: round up all the Pidgins for which data is available. Beach-la-mar, New-Guinea Pidgin, Police Motu (an Austronesian-based Pidgin of New Guinea), Chinook Jargon, Sabir... Let’s round them up and ask ourselves: what have they got in common? And what is it that they don’t have? So far, I haven’t found one with tones, I haven’t found one with cases, I can’t think of one with inflected verbs.” {3}
In summary, there is much testimony available from language teachers, linguists and amateur polyglots indicating that some languages are more difficult for most humans to learn than other languages, and that many of the characteristics which cause such difficulty can be specified.
1.4 the matter of cultural neutrality
Most of the a posteriori artificial languages that we know of are based entirely on European natural languages, not only in vocabulary selection but also in the realm of grammatical functions. Often the supporters of such purely European projects attempt to justify their design by offering chauvinistic statements such as “all persons engaged in international commerce know a European language” or “all educated persons in the world have some familiarity with a Romance language.” The sheer bigotry – not to mention the inaccuracy – of such statements is remarkable. (In view of the importance of Chinese cuisine, philosophy and science in the world, perhaps it would be more accurate to say “all educated persons in the world have some familiarity with the Chinese language.”) The nations of East Asia and the Middle East have become increasingly important players in world commerce, science, and politics. As African and mid-Asian countries become more developed, their participation in global activities will also increase. Europe has never really been the center of civilization, and is even less so now. There seems to be no objective basis to support the Eurocentric IAL design practices of the past.
The quest for ease of learning sometimes seems to conflict with the quest for cultural neutrality. Certain choices in the elements of orthography, syntax, and grammar will inevitably be familiar to some groups of people and unfamiliar to others. In view of our design goals, it seems most reasonable to make such choices in a way that will provide maximum ease of learning (not necessarily the same thing as familiarity) to the largest number of people.
2.1 phoneme selection
As Sapir et al. pointed out, an optimal interlanguage should not use any sounds that would cause serious difficulties to large groups of speakers. {4} The IAL should only use phonemes which most people in the world already know or can easily learn to pronounce. A demonstration of how this selection process can be performed follows.
If we assemble a chart that shows which phonemes occur in which of the world’s languages {5}, we see that some sounds are nearly universal while others are quite rare. Morneau surveyed data on 25 major languages and indicated that the following phonemes are used in at least 22 of the 25: /a, e, i, o, u, b, d, k, l, m, n, p, s, t, y/ (“y” represents the semi-vowel [j] as in “yo-yo”). If we limit ourselves to CV (consonant-vowel) syllables and classify “y” as a consonant, this inventory of phonemes is sufficient to produce 50 one-syllable words, 2500 two-syllable words, etc. If we allow a greater variety of syllable shapes, more words can be made. If we include some phonemes which are not quite as widely distributed as the ones above (but still quite common), even more words become possible.
Sapir et al. recommended an even smaller array of phonemes: /a, i, u, p, t, k, s, l, m, n, v/.
Jason Johnston commented, “My proposal is that the inventory should encompass phonemes that are considered phonologically unmarked. I realise that phonologists do not agree on a definition of markedness, but yet they seem by and large to come up with the same classifications for particular sounds, and these classifications tend by and large to agree with research on the order of acquisition in children, the common errors of foreign learners, the loss of distinctions in various sorts of aphasia, etc. From this very large body of work we can get a pretty clear picture of which sounds are unmarked (and therefore should be easy for the learner even if his/her native language doesn–t contain them). The most unmarked phonemes would be these: /a, i, u, p, t, k, m, n, s, l/. A second rank of slightly more marked, but still generally manageable phonemes would be: /e, o, b, d, g, f, h, y, w/. A third rank of dubious but possible phonemes would be: /v, z, r, ch, sh/.” {3}
2.2 phonotactics
For many of the world’s people, consonant clusters are difficult or impossible to pronounce. (Examples that have caused difficulty for existing IALs: the consonant followed by ‘s’ plurals of Volapük, the pt- and ks- of Glosa ptero and xeno, and the gn-, pt-, sts- of Esperanto gnomo, pterido, scii. ) Therefore the structure of syllables in the IAL must be controlled. Consonant clusters should be totally eliminated, or limited to those which are almost universally easy to pronounce. Linguists have varying opinions as to which consonant clusters are easy for most humans to pronounce, but many consider the following guideline acceptable: A syllable may begin with a vowel, or with any single consonant, or with a consonant followed by a semi-vowel or a liquid (w, y, l). A syllable may end in a vowel or diphthong, or a nasal consonant (such as n).
The morphology should also prevent sequences of more than two vowels from occurring; words such as “ouokivuia” can be tongue-twisters to most speakers of the world’s predominant languages. {6}
2.3 phonemic distinctions
Some languages use tones phonemically (i.e. to distinguish syllables which are otherwise identical; for example, in Mandarin Chinese ren means “a person” when pronounced with a rising tone, but means “to recognize / identify” when it has a falling tone). This sort of tonal distinction does not occur in many of the world’s predominant languages, and learning to use tones in this way is difficult for many people; therefore, the IAL should not use tones in this manner. For similar reasons, vowel length, consonant gemination, and aspiration also should not be used phonemically.
Because an optimal IAL will have a relatively small inventory of phonemes, it can allow some variation in the pronunciation of each item; for example, “f” can be pronounced either as a labiodental or as a bilabial.
Another important factor in the pronunciation of a language is its prosodics – the system of stress and intonation. English and some other languages have very complex systems of prosodics; not only is it difficult to predict which syllable of a polysyllabic word will be stressed, it is also difficult to predict how a certain pitch contour might affect the interpretation of a phrase. Written English does not contain much information about its intonation; punctuation marks, italics and boldface type only convey a small fraction of the underlying complexity. Most people who study English as a second language never completely master the prosodics. Intonation patterns vary from one natural language to another. An intonation that might seem friendly or humorous to speakers of one language could seem vulgar or insulting to others.
An optimal IAL will not contain such complications. Shifts in semantic emphasis and indications of the speaker’s mood should be indicated explicitly by affixes, particles and/or interjections; such items are less likely to be misinterpreted in international usage, and can be written as easily as they can be spoken. If a system of prosodics is specified it must be simple. Alternatively, the IAL designers can simply declare that each speaker can use whatever patterns of stress and intonation he desires. Considering the extreme difficulty that people have in learning the prosodics of languages other than their native tongues, the latter option may be the best choice.
3.1 writing system
Simplicity and regularity are desirable in the orthography of an IAL. Therefore each grapheme (i.e. each letter or special combination of letters) should represent only one phoneme, and each phoneme should be represented by only one grapheme.
The Latin/Roman alphabet is the most widely distributed writing system; it is used with some modifications by many languages all over the world (not only by Indo-European languages). Therefore it is probably the best choice for an optimized IAL.
Some language inventors have believed that creating a unique new alphabet would put all students on an equal footing. However, the resulting difficulties (in producing books and magazines, in transmitting text via computer networks, etc.) would probably out-weigh the gain in “equality.”
Diacritical marks that are widely used (such as acute and grave accent marks over vowels) do not introduce any great difficulty in typesetting or in the computerized transmission and processing of text. However, the use of unique diacritical marks in a constructed language would cause unnecessary difficulties, and could not be considered acceptable in an optimized IAL. (Consider the difficulties caused by the circumflexed consonants in the Esperanto alphabet. Esperantists have spent thousands of man-hours modifying their typewriters, locating and installing special fonts and utilities in their computers, and adding circumflexes by hand to the “masters” of their newsletters.)
There are many reasonable ways to design the structure of morphemes in a planned language. Should there be something about the arrangement of consonants and vowels in a word to indicate its part of speech? Should bound affixes be different in some obvious way from free roots? Reasonable persons might disagree on these matters. However, the following guidelines seem necessary, in view of our goals.
4.1 compounding
The meaning of any compound word must be deducible, on the basis of logic and pragmatics, from the meanings of the individual formatives. Compounds such as “cranberry,” in which one formative has no independent meaning, should not exist. {7} Idiomatic combinations of elements, such as “put up with” as a synonym for “tolerate,” are not acceptable in an optimal IAL.
The language should be designed so that all compounds can be unambiguously divided into their constituent morphemes, and so that no single morpheme can be mistaken for a combination of several morphemes. Confusion can occur if such auto-analysis is not designed into the language; for example, the Dutch word kwartslagen can mean “quarter beats” (kwart + slagen) or “quartz layers” (kwarts + lagen); the Esperanto word sukero might mean “sugar” (suker + o) or “a drop of juice” (suk + er + o). While it is true that context can usually indicate which meaning is intended, there is no reason for such morphemic ambiguity to exist in an optimal language design.
In the past, IAL designers have had a tendency to assign very specific meanings to very vague compounds, e.g. asserting that religion-place can only be interpreted as “church.” This has not worked well in practice, and has resulted in unnecessary bloating of IAL vocabularies as users of the IAL bring in “foreign” words to replace the overly broad and idiomatic compounds (e.g. the notion of “cannon” is often expressed in Esperanto by kanono rather than pafilego. ) Future IAL designers should avoid repeating this mistake and should try to make their compounds specific enough to indicate the intended meaning. For example, the IAL equivalent of public-pray-building might be a more satisfactory way to express the notion of “church” than religion-place.
4.2 allomorphy
Allomorphy refers to one morpheme having two or more variant forms. Examples of allomorphy: the plural indicator in English is -s in “books,” -es in “boxes,” -en in “oxen,” and is represented by a vowel shift in “men”; in Esperanto, possession is indicated by the suffix -a with pronouns, by the suffix -es with the correlatives, and by the preposition de with nouns. A related phenomenon is suppletion, in which an inflected form of a word has a completely different origin than the root; an English example is went, the past tense of go. Allomorphy and suppletion increase the difficulty of memorizing a language’s vocabulary; therefore they could not be considered acceptable in an optimal IAL.
5.1 vocabulary sources
Most proposed IALs have vocabularies that are almost entirely a priori (artificial) or a posteriori (borrowed from natural languages). Compromise or mixed-type vocabularies are relatively rare.
In a priori “philosophical” languages, words are arranged in categories according to their meanings. For example, in the language Ro, lugalab means “broccoli,” lugalap means “parsley,” and lugalat means “lettuce.” It is relatively difficult to memorize a vocabulary of this type; even the inventors of these languages cannot use them without frequently consulting their dictionaries. Words with similar meanings have similar sounds, which apparently contributes to the difficulty of memorization and increases the probability of misunderstanding if there should be a typographical error in a piece of text or a loud bit of background noise during a conversation.
In section 1.4, we established that the most common a posteriori approach to vocabulary design – using a mish-mash of words taken from European languages – is unacceptable for an optimal IAL. Fortunately, there are other ways to create a vocabulary. One is an a priori, non-classificational approach; morphemes can be created randomly, by writing letters on pieces of paper and drawing them out of a hat, or by programming a computer to randomly generate words having the desired syllable structure. This method has the advantage of cultural neutrality; no student of the language will have an advantage over others when it comes to memorizing the lexicon.
It is also possible to create a vocabulary having a mixture of a priori and a posteriori qualities. (Volapük and Loglan are early – and some would say inelegant – examples of this mixed approach.) Although the words are not “recognizable at first sight,” many of them have similarities to words in the world’s predominant languages; if implemented properly, this approach to word creation would make memorization easier than the memorization of a randomly-generated vocabulary, without the disadvantages of chauvinism, pseudo-cognates (“false friends”), and idiomatic baggage that come with a purely Eurocentric a posteriori vocabulary.
5.2 semantics
One factor which complicates the learning of many natural languages is their large supply of synonyms. A newcomer to the English language can have trouble deciding whether to say “start,” “begin” or “commence,” whether to refer to a small publication as a “leaflet,” “brochure,” “flyer,” “handbill,” “circular” or “broad-side”; whether to say “quick,” “fast” or “rapid,” and so forth. Making the wrong choice will produce a phrase that sounds strange to native speakers of the language.
On the other hand, some common words have so many different meanings that they can be hard to interpret; for example, “get” can mean “acquire” (get some money), “move to” (get off the ship), “cause to move to” (get a letter to Chicago), “become” (get ready), “cause to become” (get our feet wet), and several other meanings.
Obviously, an easy-to-learn IAL will not put such obstacles in a student–s path. Anyone who has learned the grammar should be able to pick up a bilingual dictionary and begin confidently assembling sentences immediately. Therefore, an optimal IAL must have a general rule of “one morpheme per meaning” (no synonyms within the inventory of root-words) and “one meaning, or one tightly-constrained and clearly-defined range of closely related meanings, per morpheme.”
It is also desirable that the IAL permit people to communicate using a relatively small vocabulary of root-words. This reduces the burden of memorizing the lexicon. In view of projects such as Ogden’s Basic English and the Voice of America–s Special English, it seems that a vocabulary of fewer than 1500 carefully chosen root-words is sufficient for news broadcasts, personal correspondence and mundane (non-technical) conversation. The German and Chinese languages illustrate that a few thousand basic roots can be amplified through derivation and the creation of compound words to express myriads of concepts.
It seems desirable that the designers of an optimal IAL should study word frequency lists from many languages and attempt to identify the most frequent underlying concepts, and create morphemes representing those concepts as a basis for the vocabulary. For example, the English preposition “about” (when it means “pertaining to”), the noun “topic” (that which is pertained to), the adjectives “relevant” and “irrelevant,” and the verb “to pertain to (something)” could all be expressed as derivatives of a single basic morpheme in an optimal IAL. The semantic space covered by each morpheme should be defined as clearly as possible, and the definitions should be simultaneously written in several natural languages, as well as in the language which is being created. This approach seems the best way to create a vocabulary that is truly international, rather than a simple encoding of the designer’s native language along with all of its quirks.
As far as is humanly possible, the designers of an optimal IAL will avoid the embodiment of sexism or other prejudices in the design of the IAL’s vocabulary. The practice of treating female nouns as secondary derivatives of their male counterparts, seen in many artificial languages that were designed in the 19th century, can hardly be viewed as optimal in today’s world.
6.1 simplicity
As the great linguist Otto Jespersen observed, “That language ranks highest, which goes farthest in the art of accomplishing much with little means, or in other words, which is able to express the greatest amount of meaning with the simplest apparatus.”
6.2 syntax
6.2.1 the need for a planned syntax
Much of the complexity in natural languages exists in the realm of syntax (the rules that govern the arrangement of words and phrases within a sentence). The English sentence “our dog’s barking indicated the presence of an intruder” contains 9 words which could, theoretically, be arranged in 362,880 different combinations, but only one of those combinations is grammatical and conveys the intended meaning.
Unfortunately the designers of many proposed IALs have not considered the difficulties that can be caused by a poorly-designed syntax, or by the lack of any plan at all. In some cases, they have refused to spell out their rules of syntax, claiming that the arrangement of words in their languages would be “natural” or “intuitive.” This invariably means that it is “natural” to those who have the same native language as the IAL designer, and not to others. The syntax that comes ”intuitively” to a speaker of English is quite different from that which seems “intuitive” to a speaker of Bengali or Japanese.
There are formal (objective) techniques for specifying the number of ways in which words can be combined to create grammatical phrases in any given language. A description of a typical natural language’s syntax, complete with all of its irregularities and variations, might contain thousands of production rules. In order to conform to our design goals, the syntax of an optimal IAL must have a simple, predictable, easily learned and thoroughly described syntax.
6.2.2 the SVO approach
Linguists have attempted to classify natural languages based on the arrangement of subject (S), verb (V) and object (O) in simple declarative sentences. (Granted, this approach has some flaws: many languages are difficult to classify or appear to be in a state of transition from one category to another.) There are six possible orderings: VSO, SVO, SOV, VOS, OVS, OSV. Most of the world’s languages fit within the first three categories, placing the subject before the object. SVO is the most common word order in the world’s predominant languages, and is notably common in natural languages which are used as interlinguas between different language groups (e.g. English, French, Swahili, Indonesian).
A language with a relatively strict SVO ordering does not require the use of inflections or marker words to distinguish subject from object. The listener or reader can begin to interpret a sentence before reaching its end, which is not true of languages that have free word order, in which the recipient must wait for the entire sentence to arrive and then must mentally unscramble the various elements before comprehension can begin. Some linguists feel there is a link between grammatical structures and the ease with which the brain can interpret sentences. “The perceptual advantage of SVO languages is the ready identification of subjects and objects, which are separated (by verbs) in SVO but not SOV or VSO languages. It might also be mentioned that English tends to have topics in sentence-initial position... Subject and topic will often coincide, a coincidence that apparently enhances processibility, especially when the subject is also the semantic agent.” {8} Therefore it is reasonable to advocate an SVO syntax for an IAL, or the more flexible option of allowing any sequence in which subject precedes object (SVO/SOV/VSO). However, this still leaves many questions unanswered, such as: Should adjectives precede or follow the nouns which they modify? Prepositions or postpositions (or neither)? And so forth.
6.2.3 configurational or inflectional?
There is a larger question to be considered: should the roles of the verb’s arguments be indicated by word order or by overt markers (such as noun declension or “case tags”)?
Jason Johnston states there is much evidence to indicate that word order is the most natural and efficient choice, “including the fact that the words have to come in some specific order anyway, so that the order may as well signal something useful (a point due, I believe, to Jespersen), and the fact that all pidgins and creoles are strictly configurational (as is still, for what it’s worth, the underlying logical structure of all languages according to Noam Chomsky).” {3}
Sometimes we hear the assertion that an inflectional case-marking system and free word order are necessary to facilitate poetry and literature. This claim seems to be contradicted by the fact that Chinese has no inflections to mark cases but is nevertheless powerfully expressive, as demonstrated by its 2000-year history of unsurpassed poetry and literature.
There is evidence to indicate that the declension of nouns and adjectives is difficult for many people to master. Studies discussed in the newsletter of the Esperantic Studies Foundation indicate that American students of Esperanto at the intermediate level have difficulty using the accusative case flexion correctly. There is corroborating testimony from Chinese Esperantists {9} and those who have taught Esperanto in East Asia. {10}
Because inflectional case-marking is not strictly necessary, and because it provides questionable benefit in return for the significant difficulties it causes, we must conclude that it is not desirable in an optimal IAL. (The same principle can be applied to mandatory marking of tense, definiteness, and number; but these are not syntactic issues.)
6.2.4 a minimalist approach
If simplicity and consistency are viewed as the main objectives in IAL syntax design, the best available choice might be a purely right-branching VSO syntax in which heads always precede their modifiers. Morneau {11} demonstrates that it is possible to design a syntax of this type that can perform all the communicative tasks needed in a human language with fewer than two dozen context-free production rules.
6.2.5 attachment ambiguities
Consider the phrase “little girls’ school.” It is not clear whether the adjective “little” modifies ”girl” or “school” or, possibly, both items. Now consider the sentence “Men don’t talk about their relationships with each other.” Does this mean that men don’t talk to one another about relationships, or does it mean they don’t talk to anyone about the relationships that they have with their fellow-men? These are examples of attachment ambiguities, situations in which there is no formal mechanism available to determine which item a modifier (such as an adjective or prepositional phrase) is meant to modify. Sometimes the context of the entire conversation or text provides enough information to indicate which meaning was intended, but very often it does not.
Attachment ambiguities often go unnoticed in casual conversation. However, they can be a significant cause of misunderstandings in contracts, technical instruction manuals, laws, and other types of communication that require precision. People translating material from one language to another – a scenario in which an IAL might frequently be involved – often must expend a lot of time and mental energy trying to interpret such ambiguous phrases. Many computerized text analysis systems cannot cope with attachment ambiguities; they must pause and ask for human assistance in interpreting the material. For these reasons, an optimal IAL’s syntax should be designed in a way that reduces or eliminates such ambiguities.
6.3 gender
One factor which complicates the learning of languages such as German is the need to memorize the arbitrary gender of every noun. Such an unpredictable feature would not be acceptable in an optimal IAL.
6.4 transitivity
In some constructed languages, it is necessary to memorize whether a verb is inherently transitive or intransitive, and affixes are used to convert transitive verbs to intransitive and vice versa. (And when transitive verbs are derived from inherently intransitive ones, it is not always clear whether the resulting word means “do activity X to Y” or “cause Y to engage in activity X.”) Having to memorize a verb’s arbitrary transitivity is an unnecessary burden, just as having to learn the arbitrary grammatical gender of a noun would be. Therefore an optimal IAL should form transitive, intransitive and causative verbs in some completely regular, predictable and well-defined manner.
To what degree should an IAL be designed to facilitate computerized processing? There is some disagreement about this issue. Fortunately many of the qualities enumerated above which make a language easy for humans to learn and use also facilitate computerized processing. An IAL designed in accord with the criteria specified in this essay will be more computer-tractable than any natural language.
The needs of computers should not be an over-riding factor in the design of an IAL, for many reasons. The whole purpose of an IAL is to facilitate human-to-human communication without the expensive, invasive and unreliable mediation of human or electronic translators. Computerized processing of text and speech is still in its infancy, and a language designed to accommodate current technology would seem absurdly constrained 20 or 30 years from now.
This essay is not meant to be a detailed specification of a particular language design. Instead, it is a call for discussion, a first step toward establishing a set of criteria by which interlinguists could possibly design a language which would be nearly optimal for the role of IAL.
In addition to language design criteria, we should also consider means of gathering support for an IAL and persuading significant numbers of people to use such a language. When I wrote the first draft of this essay, I believed that designing an IAL with optimal features would be an important factor in gathering support. However, when you look at the languages that actually are being used for international communication, it becomes clear that the features of a language don’t matter as much as the benefits of knowing the language.
People generally learn a second language for practical reasons such as qualifying for a better job, being able to communicate while traveling in a new area, or having access to information in a particular field of study.{12} Until auxlang designers find ways to provide some practical benefits to potential students, artificial auxiliary languages will continue to be unpopular, regardless of how well or how poorly they are designed.
footnotes:
1 Sapir, Edward: Wanted – a world language. in: The American Mercury, February 1931.
2 Slobin, D. I.: The acquisition of Russian as a native language. in: The genesis of language – a psycholinguistic approach, edited by Smith et al. (1966)
3 From discussions in the Conlang mailing list (Internet), late 1992 - early 1993.
4 Sapir, Edward, et al. Memorandum on the problem of an international auxiliary language. in: Romanic Review 16: 244-256 (1925)
5 Morneau, Rick: Designing an artificial language - Phonology. September, 1991. Currently located at this website.
6 The phrase “the world’s predominant languages” in this essay refers to those languages which have the largest numbers of speakers and which, taken together, are spoken by approximately half of the world’s people. Estimates vary widely but it seems safe to say that this set contains at least the following members: Mandarin, English, Hindi, Spanish, Russian, Arabic, Bengali, Portuguese and Japanese.
7 Maxwell, Dan: Principles for constructing planned languages. in: Interlinguistics edited by Klaus Schubert; (New York & Berlin: Mouton de Gruyter, 1989)
8 Edward Finegan in: The World’s Major Languages edited by Bernard Comrie (Oxford University Press, 1990)
9 For example, Lee Sau Dan in the newsgroup sci.lang, 1996.01.25: “However, there are still traces of Indo-European grammar: the constrast between accusative and nominative cases, and the constrast between singular and plural forms. I wonder if it justisfies to retain these two features. I can live without problems even if the language I use everyday does not have these two features. I cannot manage with these features well, because they are alien to my native language, Chinese.”
10 For example, Barry Crown in the newsgroup soc.culture.esperanto, October 1992: “It’s true, however, that the grammar of Esperanto is an absolute nightmare for most Asians. Many Asian languages have a much simpler grammatical structure than Esperanto.”
11 Morneau, Rick: Designing an artificial language - Syntax. 1992, 1994. Currently located at www.eskimo.com/~ram/essays.html, but if it moves you can use a search engine to find it.
12 After he left the International Auxiliary Language Association (IALA), Prof. André Martinet wrote in Word (viii, 2), “The common mistake of nearly all language-makers is to assume a demand where there is practically none. Well publicized and perfectly utilizable artificial languages have now been in existence for more than seventy years. Everybody knows about Esperanto, but no one, except a handful of idealists, ever bothers to learn it. The reason for this is not hard to find: except for professional linguists and a few amateurs, a language is never an end in itself. It is a tool with which one expects to achieve definite aims. An American planning a trip to Europe may decide to learn some French because he expects a knowledge of that language to make his tour more pleasant or intellectually and artistically more profitable. A physician may decide to learn German, a language through which he may hope to increase his professional competence. But who will take the trouble of learning even the simplest language if it does not give him access to new values? Would it not therefore be sound, before launching into the shaping of a new language, to investigate whether new values can be created which could be reached only through the proposed linguistic medium. If none can be thought of, it would certainly be wise to desist from the start.”
afterword:
One more thing… Some people think it’s paradoxical that I continue to maintain this essay after writing ”Farewell to Auxiliary Languages.” Maybe so, but this is one of the most frequently viewed pages on my web site, so I continue to maintain it, like the gardener who has lost interest in roses but continues to cultivate a few for the friends and neighbors who “ooh” and “ah” over them.