The Infancy and Adolescence of Artificial Languages By Noah Savage.

The Infancy and Adolescence of Artificial Languages

By Noah Savage

The Search for the Perfect Language

• Umberto Eco’s book The Search for the Perfect Language covers a multitude of proposals, creations, and observations regarding “perfect” languages.

• What the searchers consider “perfect” varies, but what they all have in common is that cognitive linguistics has decided the ultimate realism of what they sought.

• We will use the findings illustrated in Feldman’s book to evaluate the realism of what has been found by those language searchers and others.

Umberto Eco• Umberto Eco is a professor at the

University of Bologna.• His career is all over the place—Eco

has been a scientist, a historian, a novelist, and a lot more. He has also received more academic honors than can be listed on this slide.

• He is the author of the bestselling novel The Name of the Rose.

• Most importantly to us, he is the author of The Search for the Perfect Language.

Jerome Feldman• Jerome Feldman is the author of

From Molecule to Metaphor.• He is Professor of Computer

Science at UC Berkeley.• His focuses include Information

Technology, Computer Science, and Cognitive Science.

• He is also focused on helping underrepresented groups find opportunities in the field of Information Technology.

Infancy

John Dee

• We will begin our story in medias res with the famous English magician John Dee.

John Dee

• In the 16th century, the man discovered a unique primitive unit.

• Dee’s writings have been interpreted to claim that this primitive is primitive for all things, the way they are, independent of personal construal.

• What do you think it is?

Why, It’s this symbol, of course!

Monad

• This symbol, called the Monad and introduced in the essay Monas Hieroglyphica, was touted as means of generating all alphabets, sounds, and words, and as they were, rather than how they appeared to be, to boot.

• In other words, Monad is the ultimate semantic primitive, from which one generate all linguistic units.

• The Monad “restores the language of Adam”, according to Dee.

• (The term “monad” is used in many places and by a variety of philosophical traditions, but we’ll focus only on Dee’s symbol.)

How does it work?

• This is how we get the sound [i]:

How does it work?

• The average first and second frequencies of the sound waves produced by vibrating vocal cords and amplified by the shape of the vocal tract of the vowel [i] are 300 and 2500. If one adds up the number of the most salient shapes in the Monad (4) and multiplies it by the number of days in a week (7), we get 28, the same number as the sum of the aforementioned frequencies divided by 1000.

• The “love” heart can be made by placing the bottom part of the Monad above the top.

• Etc.

Did John Dee really believe that?

• No. Dee’s methods were even more complicated than those in the previous examples.

• Dee’s specific means of generating meaning from the Monad involved lengthy geometric computations, kabbalism, and a fair amount of self-confidence.

All of which begs the question:

• Are children born with the innate faculty to generate sounds, concepts, and letters via differential calculus performed on the Monad?

• More research is necessary to verify this, but even if his creation becomes discredited, Dee deserves credit for finding a way to combine semantic primitives, objective reality, and religious mysticism into a single project.

Ars Magna

• Another project, this one earlier and more like a language, is Ramon Llull’s Ars Magna.

• Ramon Llull was a Spanish priest who believed that he could create a logical language that would convince the heathens to convert.

Ars Magna

• To do this, he drew up a table of 54 words that could be combined in any which way.

• Llull believed that the heathens would see the logic behind these combinations (i.e. goodness is glorious) and convert.

The Problem Is…

• For one, Llull assumed that linguistic expressions are processed logically. Feldman points out that formal logic is heavily insufficient in explaining linguistic processes. Thus, given what we know about language, we have no reason to believe that a project such as Ars Magna would convince anyone to convert to Llull’s faith.

The Problem Is…

• In addition, there’s nothing terribly logical about Llull’s creation. For example, speakers are simply not allowed to make certain combinations. Combination of certain terms (such as “God” and “bad”) is discounted as “unacceptable” on account of their falsehood.

• However, it is certainly possible to think of such a combination. Even the Bible holds numerous records of people’s making this combination.

• Llull’s creation mirrors his individual perspective, which isn’t the same as that of everyone else.

• Feldman points out that each language promotes its own conceptual system, or way of understanding the world. Llull put his own conceptual system into his creation and assumed that the heathens would see his system as the one true way to imagine the world.

So What Happened?

• Llull was murdered.

“Logical” Languages

• In the seventeenth century, a number of British scholars began the push for the creation of a language unburdened by ambiguity.

• In other words, no polysemy, no metonymy, no metaphor. A word for every concept, exactly way said concepts are, independent of human construal.

Why?

• Universal communication and its benefits.• Philosophical and scientific debate would be a

lot easier if the language of philosophers and scientists described objective reality.

• Logic-based language would also clear up issues of logic in people’s statements and result in people’s speaking more logically.

• Thus, these languages are given the title a priori philosophical languages.

A Priori Philosophical Languages

• You already know how successful they were.• They never quite succeeded creating such a

logical language because human cognition does not operate that way.

• There are 100 trillion neural connections in the brain. It is not known exactly how each neuron and neural connection works in the production of language, but it’s different for each person.

And so in order to eradicate polysemy…

• Each person’s experience is different, so the 100 trillion neural connections are arranged differently for everyone. That’s a lot of possible combinations! Assuming that there are no constraints on how connections can be combined (which, to be fair, is a large—and almost certainly false—assumption), there is a possible 100 trillion factorial combinations in which a notion may reside. In addition, individual meaning depends on even more than the arrangement of neural connections in the brain, as what is encoded in each connection depends on environment in addition to combinatorial arrangement; every brain belongs to a unique body. For example, the encodings of a peasant woman in Medieval England will be different from her counterpart in modern England, as the two women will see, hear, and generally experience different things, for reasons related as well as unrelated to neuron combinations in the brain.

• Thus, in order for a language to be free of polysemy, each concept must be represented by well over a googolplex terms, perhaps an infinite number.

And did the “logical” languages have that many?

• No.

• They did, however, have a number of interesting features.

• Here are a couple of highlights:

Highlights

• Taxonomic lexicon was common throughout the a priori languages.

• Consider George Dalgarno’s word for courage, pot.

• The “p” makes the word an issue of sensitivity, the “o” makes it an issue of “primary emotion”, and the “t” identifies its individual place among the rest of the sensitive primary emotions.

• Problematically, it may well be impossible to explain language and meaning. Thus, this taxonomy of “courage” is not bound to clear up any scientific or philosophical issues.

Highlights

• One language, Solresol, had only seven syllables: do, re, mi, fa, sol, la, and ti.

• Thus, its phonology is easier than that of any other language.

• However, just try to imagine the task of understanding spoken Solresol! Finding breaks between the words, among other aspects of understanding speech, would be unduly hard.

Highlights

• One language had the following words for numbers 1 through 10:

• 1 siba

• 2 sibe

• 3 sibi

• 4 sibo

• 5 sibu

• 6 sibra• 7 sibre• 8 sibri• 9 sibro• 10 sibru

Summary

• In short, “infancy” refers to the status of an artificial language as unworkable due to its creator’s misconceptions about language and cognition.

• Many of such artificial languages are languages created by those who presume that their personal conceptual system is universal.

• After a while, artificial languages began to get a bit more sensible.

And yet

• “Logical” languages live on!

• Consider Lojban, the language created by the “Logical Language Group”.

• According to a pamphlet providing an introduction to Lojban, “Lojban is designed to be unambiguous in orthography, phonology, morphology, and grammar.”

Claims about LojbanThe following are claims made by

Lojban supporters:• Lojban is designed to be used by

people in communication with each other, and possibly in the future with computers.

• Lojban grammar is based on the principles of logic.

• Lojban has an unambiguous grammar.

• Lojban's 1300 root words can be easily combined to form a vocabulary of millions of words.

• Lojban attempts to remove restrictions on creative and clear thought and communication.

• Sound familiar?

Adolescence

Adolescence

• I have classified all of the previous linguistic creations as “in their infancy” due to their complete lack of development in terms of cognitive linguistics.

• However, a number of artificial languages were not created under the assumption that human cognition can be so easily violated.

International Auxiliary Languages

• Beginning in 1879 with Johannes Schleyer’s Volapük, a number of artificial languages were developed for the purpose of serving as an international language that everyone could learn.

• These IALs were imperfect, but people could learn and speak them more easily than a natural language.

Esperanto

• Esperanto, the most successful IAL in history, first appeared in 1887 in the form of a book written by its creator, the Polish oculist Ludwig Zamenhof.

• Zamenhof took notice of the language-fueled violence going on around him and thought to create an IAL that everyone could speak ( as did many IAL creators).

Why Esperanto Works

• Simply put, Esperanto works because the language operates the same way as do natural languages—in tune with people’s actual cognition.

• Esperanto does not claim to feature a universal conceptual system.

• Esperanto is full of metaphor, metonymy, polysemy, and every imaginable form of construal.

For Example

• The Esperanto word for “blink” is “palpebri”, literally meaning “to eyelid”. Metonymy!

• Exceptional proof of Esperanto’s cognitive realism is its schematicity, especially for concepts that developed after its creation.

• Consider the frame in which are found the Esperanto “reptile words”.

• Reptile words are the six terms for speaking a certain language when another would be appropriate: reptilumi, krokodili, aligatori, lacerti, kajmani, and gaviali.

• (Reptile, crocodile, alligator, lizard, caiman, and gharial)

The Facts• Because Esperanto is a

cognitive artificial language, its popularity has grown. Esperanto has:

• A worldwide following.• Spiritual significance for

a number of religions, most clearly Oomoto

• Its own martial art, Aikido.

• And Umberto Eco is a supporter!

Summary

• Adolescent languages are the languages, almost all of which are IALs, that are just as cognitive as natural languages, only simpler and easier to learn.

• These languages differ from natural languages without attempting to change the way that language works.

Bibliography• “Ars Magna.” Digital image. Lullianarts.net. 3 November 2007.

<http://lullianarts.net/Ars-Magna/tfig.gif>.• Barbosa de Aquino, Ildete. “Mi, Esperanto, kaj Oomoto.” Oomoto.or.jp. 5

November 2007. <http://www.oomoto.or.jp/Esperanto/eoGast/barbosaparolo.html>.

• “Dee Hieroglyph.” Digital image. Wikipedia.org. 3 November 2007. <http://en.wikipedia.org/wiki/Image:DeeHieroglyph.gif>.

• Eco, Umberto. The Search for the Perfect Language. Trans. James Fentress. Oxford: Blackwell Publishers, 1995.

• “Esperanto: Oftaj Demandoj.” MIT Societo de Esperanto. 6 November 2007. <web.mit.edu/esperanto/www/reading1.doc>.

• Feldman, Jerome. From Molecule to Metaphor: A Neural Theory of Language. Cambridge: MIT Press, 2006.

• Feldman, Jerome. “Jerome A. Feldman.” 2 February 2006. UC Berkeley. 2 November 2007. <http://www.cs.berkeley.edu/~jfeldman/>

• Leite, Milton. “Esperanto Flag.” Digital image. Flickr.com. 5 November 2006. <http://www.flickr.com/photos/moon316/211745032/>.

Bibliography• Lojban.org. The Logical Language Group. 6 November 2007.

<www.lojban.org>.• “Lojban Flag.” Digital image. Lojban.org. 6 November 2007.

<http://www.lojban.org/tiki/tiki-download_wiki_attachment.php?attId=560>.

• Moshe, Pereg. “Esperanto Stamp.” Digital image. Wikipedia.org. 8 November 2007. <http://upload.wikimedia.org/wikipedia/en/5/5e/Esperanto-PM.jpg>.

• “Oomoto.” Digital image. Oomoto.or.jp. 5 November 2007. <http://www.oomoto.or.jp/Images/kapalogo.gif>.

• “Ouroboros.” Digital image. Wikipedia.org. 4 November 2007. http://<upload.wikimedia.org/wikipedia/commons/f/fa/Ouroboros.png>.

• “Overview of Lojban”. 2000. The Logical Language Group. 6 November 2007. <http://www.lojban.org/tiki/tiki-download_wiki_attachment.php?attId=89>.

Bibliography• Ruch, Allen. “A Short Biography of Umberto Eco”. The Modern

Word. 22 March 2004. 2 November 2007. <http://www.themodernword.com/eco/eco_biography.html>.

• “Sky.” Digital image. Flickr.com. 7 November 2007. <http://farm1.static.flickr.com/68/222320192_a3c99087f3.jpg?v=0>.

• “Steven Seagal.” Digital image. Self-defender.net. 5 November 2007. <http://www.self-defender.net/aikido4.jpg>.

• “Umberto Eco.” Digital image. Umberto Eco. Britannica.com. 2 November 2007. <http://www.britannica.com/eb/article-9031922/Umberto-Eco>.

• “World Flags.” Digital image. Solar4power.org. <http://www.solar4power.com/images/world-flags.jpg>.

Maturity

Gjerojŋ Adoltsano

Maturity

• We have seen how artificial languages fail to grow due to their creators’ cognitive misconceptions.

• We have also seen how artificial languages can achieve cognitive realism.

• But let’s take it a step further.

Maturity

• The basis of this project is a theory developed for the better part of a century by neuroscientists, biologists, and cognitive linguists.

• Feldman references this as “linguistic determinism”.

• The following is a look at a project that explores some possibilities, positive and negative, of artificial languages that make use of linguistic determinism.

• First, the negative.

Prodzjekt Oroboros

• 1-grad spraaksajnzksajŋeɱ besoitstaal Oroboros, sjeturov jeststaal dal ksajŋeθ k go zir gzæʎɟ goɟuʃʎ kastjumæt tsabweθ dzjenerativistθa.

• Translation: A language scientist of the highest perceived use to the state created a language, whose genesis has been observed by the government and whose future is likely to be ongoing, that has and will be for the “best people”. The dirty tricks of the generativists belong to a weak and immoral nation that deserves to be crushed by ours on account of its weakness and will continue to do so until its scheduled conquering.

Where’s the Determinism?

• Feldman documents that every language’s lexical structures drive the way its speakers understand the world.

• Thus, lexical items are reduced and expanded for cognitive effect.

Lexical Examples

• The English “nation” or “country” says nothing about the perceived deserved fate of said nation, while Oroboros does.

• Xɔʎɟ, staat, wrjeʎ, cʎenc, and gzæʎɟ.• Meanwhile, there are no separate terms

for maturity, mean-spiritedness, and domination.

• There is a 5-point scale for “good”.• Oroboros vs. spraak.

Grammatical Examples

• Felman invokes several “neo-Whorfian” studies that prove a relation between grammar style and thought style, so Oroboros’s grammar is designed to influence thought.

• For example, each word’s place on the 5-point graded scale determines many aspects of each sentence.

• (1-grad, 2-grad, 3-grad, 4-grad, and 5-grad)

Grammatical Examples

• The most important aspects go first in terms of sentence structure.

• 1-grad subjects appear only in SVO sentences, 5- and 4-grad subjects appear only in OVS sentences, and 3- and 2-grad subjects appear in either order, depending on the status of the object.

• The time system is based on the speaker’s understanding of the government’s knowledge of the action being described. The tense marker is attached to the item highest on the 5-point scale.

• Knjazn gjerjoɱec vs. gjeroɱæc ʎe

Should

• English has one word for the concept “should.” • Oroboros has five, each coinciding with one of

the points on the 5-point scale of “good”: gjero, rtuʃʎ, kurag, rviθ, and taʃju.

• For example, “Knjaznɨk gjerostaaʎ rtjektjuajeʎ ɰar Spraaksajnz 409 k kuragstaaʎ rtjektjuajeʎ ɟan Spraaksajnz 201.”

• “He should have taken LING 409 and should have taken LING 201.”

Projekto Ksilofono

• We’ve seen a example of how artificial language can rearrange people’s conceptualizations in a negative way.

• The following is a positive artificial language.

Lexical Points

• Growth is an embodied concept, so every language has terms for it.

• While the terms for growth in Oroboros are related to mean-spiritedness, those in Ksilofono are related to good-spiritedness.

• Many vices, bad decisions, and general ills are expressed in terms of maturity.

• I.e. drug use is “droga infanputro”, (loosely) “drug childishness”.

• To do drugs is “infaniĝi droge”, “to be a child drug-ly”.• Etc.

Grammatical Examples

• The maturity inherent in the action is encoded into verbs.

• Kiu estas li? (neutral/mature)

• Li estap ŝtelisto. (immature)

Grammatical Examples

• In Ksilofono possessive constructs, the group goes before the individual.

• La organizo de la viro (the club of the guy).

• La organizo ur la celo (the club’s objective).

Concept Harmony

• For concepts that we want people to associate but don’t necessarily want to combine into one word, Ksilofono suffixes (based on basic-level and embodied concepts) adapt to fit the category.

• Esperanto has a number of suffixes that affect word meaning. Gramatikeco (grammaticality) and fieco (evil), vortaro (dictionary) and monstraro (collection of monsters).

• However, Ksilofono concept harmony acknowledges these concepts differently.

• Gramatiksano (grammaticality) and fiputro (evil).• Vortfamo (dictionary) and monstrobando (collection of

monsters).

Controlled Conceptual Metaphors

• It is not possible to prevent someone from describing one thing in terms of something else, but the meanings of the source and target domain can indeed be controlled to point where certain notions are difficult to express.

• For example, the metaphor ANGER IS STRENGTH is difficult when the ideal “strong” figure is cool and level-headed.

Controlled Conceptual Metaphors

• In English, there are also a number of conceptual metaphors that are usable but uncommon.

• For example, ANGER IS POISON is a realizable conceptual metaphor, but it is relatively uncommon. Most often, anger is related to heat.

• The most promoted metaphors in Ksilofono are those that are constructed to produce the most beneficial mental connections.

Should

• Just like in its negative counterpart, there are multiple words for “should”, each tied to a measure of maturity/good-spiritedness.

• Adolta, afabla, agrabla, egoema, and infana.

• Ri adoltiĝus preni Lingvistikon 409 kaj agrabliĝus preni Lingvistikon 201.

Conclusion

• Artificial languages have stayed in their infancy for a long time, mostly due to incorrect theories of language. In fact, there are still artificial language projects that work on false assumptions about human cognition.

• Esperanto is a relatively mature artificial language, but there’s no reason to stop there.

Conclusion

• The first linguistic conceptual systems were those of tribes fighting each other over resources.

• Societies evolved, as did their languages, and modern languages have adapted to fit modern civilization.

• However, there is no compelling reason to believe that the conceptual systems maintained by modern languages are beyond improvement. That many languages retain unnecessary tribal characteristics is evidence for this.

Conclusion

• Artificial languages have a lot of potential in modifying people’s conceptual systems.

• Ksilofono is an artificial language designed for the purpose of promoting an ideal conceptual system.

• I recommend that such a project be implemented in the future.