Glossopoeia logo

Glossopoeia Pages

Glossopoeia & Glossopoeic Languages

In some of the secitions below, reference is made to J.R.R. Tolkien's essay, "A Secret Vice". This essay was delivered as a lecture in 1931; it was published posthumously, together with other of Tolkien's essays:
Christopher Tolkien (editor), 1983, "The Monsters and the Critics and Other Essays", George Allen & Unwin, 0-04-809019-0.

Later publications of "The Monsters and the Critics and Other Essays" are:
1984, Houghton Mifflin, 0-395-35635-0; 2000, HarperCollins, 978-0-261-10263-7; 2006, HarperCollins, 0-261-10263-X.

1. Glossopoeia   /glɒsə'piːə/

1.1 Derivation
Glossopoeia is the Latinized version of the ancient Greek word γλωσσοποιΐα (glōssopoiía) = the making of tongues.

This in turn is derived from the Greek words (transliteration given in parentheses):

  • γλῶσσα (glôssa) = tongue; language; anything shaped like a tongue, e.g. the reed or 'tongue' of a woodwind instrument;
  • *ποιϝός (poiwós) = builder, maker ← root ποιϝ- (poiw-).

Intervocalic [w] was lost at an early date, but we do find it written in some early inscriptions; *ποιός is not attested as a separate word in ancient Greek but we do find it as the second element in many compounds, for example:
ἀρτοποιός (artopoiós) = baker, ἐποποιός (epopoiós) = composer of epic verse, κλινοποιός (klinopoiós) = maker of bedsteads, μυθοποιός (mutʰopoiós) = composer of fiction, ξυλοποιός (xulopoiós) = carpenter, τραπεζοποιός (trapezopoiós) = table maker, etc.

From these nouns, one could derive an abstract noun by adding -ία (-ía) to the stem, and an adjective by adding -ικός (-ikós), for example ἀρτοποιΐα (artopoiía) = [the art of] baking; ἀρτοποιικός (artopoiikós) = of or pertaining to baking.

Similarly the ancient γλωσσοποιΐα (glōssopoiía) was derived from γλωσσοποιός (glōssopiós); but the 'tongues' a glōssopiós made were for woodwind instruments!

The associated adjective is not found, but would be γλωσσοποιικός (glōssopoiikós; Latinized as glōssopoeicus). From these words are derived the modern English words glossopoeia [noun], and glossopoeic [adjective] - but we now use those words to denote the making of spoken tongues, i.e. languages.

Latinization refers to the way the Romans normally adopted Greek loan words, e.g. the Greek diphthong οι /oj/ was written oe and word endings were modified to fit the morphology Latin. This is different, of course, from letter by letter transliteration.

1.2 Modern usage
In the 1980s or possibly a little earlier, the ancient word for "the making of tongues" was brought back to life with the meaning "the creation of [fictional] languages". Many sites on the Internet (including Wikipedia) state that J.R.R. Tolkien coined the word; this is almost certainly incorrect and is yet an example of the not uncommon Internet phenomenon of "canon by repetition".

There is no evidence that I know of that Tolkien ever used the word at all. It appears to have been brought back to life when the "secret vice" of language creation became more widely known through the Internet. The earliest certain public instance of its modern use, that I know of, is in the "Glossopoeic Quarterly", the first edition appearing in the summer of 1988. Its editor was Steve Deyo who, according to information I have received from Roman Rausch, "[i]n 1983 [had] coined the term glossopoeia ('language-making') during correspondence with other practicing creative linguists." Indeed, it is likely that the word occurred in print as early as 1983 in "Mythos: Seeking Truth Through Story", a publication of limited circulation that Steve Deyo produced from 1983 to 1985.

But whether he was aware of the ancient Greek word, or whether he re-coined the words on the analogy of mythopoeia and mythopoeic, I do not know. (Tolkien, by the way, did use the word mythopoeia in "A Secret Vice"; but he did not coin it as there can be no doubt he was well aware of the ancient Greek word μυθοποιΐα, i.e. mythopoeia, which was used by Strabo, Philo and Plutarch among others.)

'Glossopoeia', however, is such an obvious formation both for those who know the English word 'mythopoeia' and for those with any knowledge of ancient Greek that it seems the word was and has been independently re-coined or revived by others both before and since the publication of the "Glossopoeic Quarterly" in 1988. Although we can state with some certainty when and who first publicly use gloosopoeia with its modern meaning, it is in my view impossible and meaningless to say who first used the word that way, especially as it had been around for the past two millennia just waiting for a new usage. What can be said with certainty is that we now have:

  • glossopoeia  /glɒsə'piːə/   [noun] "the creation of language(s)"
  • glossopoeic  /glɒsə'piːɪk/   [adj.] "of or pertaining to the creation of language(s)"
  • glossopoeist  /glɒsə'piːɪst/  [noun] "one who creates language(s)"
One sometimes also comes across glossopoesis or glossopoesia instead of glosspoeia, glossopoetic instead of glossopoeic and glossopoet instead of glossopoeist. Where do these formations come from?

In ancient Greek a verb, "to make, create", was derived from the root ποιϝ- (poiw-) by adding /e/, e.g. in inscriptions we find Boiotian ἐποίϝε̄σε (epoíwēse) and Elean ποιϝέοι (poiwéoi). The intervocalic [w] dropped out in most dialects and the verb is known to students of ancient Greek as the Attic and Koine ποιέω (poiéō) "I do, make, create." From this verb were derived the nouns ποίησις (poíēsis) = 'fabrication, creation', ποιητής (poiētḗs) = 'maker, author, composer' and the adjective ποιητικός (poiētikoós) = 'creative, inventive'.

These three also appeared in Classical Latin as loan words poēsis, poēta, poēticus, hence their combining with glosso- to give the three words above.

But, dear reader, I expect you have noticed a problem. Although the meanings given above were found in ancient Greek, as many know ποίησις (poíēsis) often meant "poetry, poetic composition", ποιητής (poiētḗs) "composer of verse, poet" and ποιητικός (poiētikoós) "poetic, poetical." The Latin loan words had only those meanings, hence the English words: poesy (← Fr. poésie ← Late Latin poēsia = C.L. poēsis), poet (← C.L. poēta), poetic (← C.L. poēticus). Whatever the intentions of those who use glossopoesis and related words, the poetic connotations are liable to be strong.

As far as I know the word for myth-making remains the ancient word mythopoeia; but as well as mythopoeic of ancient derivation, one may also find the modern neologisms mythopoetic and mythopoet. I note the dictionary entry: "mythopō´et a myth maker; a writer of poems on mythical subjects" [Chambers English Dictionary, 1988 edition]; in other words, the term is ambiguous.

My personal view is that glossopoesis/ glossopoesia, glossopoetic and glossopoet should be deprecated in favor of glossopoeia, glossopoeic and glossopoeist, since the former set are potentially ambiguous neologisms but the latter set are unambiguous and are derived from compounds with ancient authority.


2. Glossopoeic languages (Conlangs)

2.1 Introduction
We saw above that from the Greek verb ποιέω (poiéō) = 'I make, create' were derived the nouns ποίησις (poíēsis → Lat. poēsis) = 'fabrication, creation' and ποιητής (poiētḗs → Lat. poēta) = 'maker, author, composer'; there was also another derived noun, namely ποίημα (poíēma → Lat. poēma) = 'anything made'. Hence logically a language created as a result of glossopoesis by a glossopoet should be called a glossopoem, but I have never come across that usage. In any case, I have written above that in my opinion glossopoesis and glossopoet should be deprecated in favor of glossopoeia and glossopoeist.

So what do we call the creation of a glossopoeist? What a -ποιός (-poios) made or created was the first part of the compound, e.g. an ἀρτοποιός (artopoiós) made ἄρτος (artos) "bread", an ἐποποιός (epopoiós) composed ἔπος (epos) "epic verse", a κλινοποιός (klinopoiós) made a κλίνη (klinē) "bedstead" etc. Therefore, what a glossopoeist creates is a tongue or language. But what shall we call such a language?

These languages have been called "invented languages", "artificial languages", "planned languages", "fictional languages", "model languages" and possibly other names besides. But the name that has caught on among glossopoeists on the Internet is conlang. This appears to have been coined in 1991 by those who set up the Conlang mailing list and was likely due to J.R.R. Tolkien who in his essay, "A Secret Vice" talks more than once of 'language construction' or of a language being constructed.

In that year John B Ross starting CCing messages from his personal email account to several people who were interested in glossopoeic languages and on the 18th July these exchanges were moved to a proper listserv, hosted on The first posting of the new list was from Ronald Hale-Evans with the subject line "Ding-dong! The list is up!" and beginning with the words "Saluton, amikoj! Glad to see the mailing list is working." July 18th is now kept as "Conlang Day" on the Conlang List.

There had been some preliminary exchanges among about four dozen glossopoeists before this and it is possible that the term conlang was first coined specifically as the name for the new list; in an email of 1st June 2011, And Rosta wrote: "I am certain that the term [conlang] derives from (a -- perhaps unreflective -- reanalysis of) the name of the list, so either it would be the first use in the archives or, if it seems that everybody is using the term from the outset, then it must date from the brief period before the listservization of the list. Given the name of the list, it is only a short and natural step to apply the list name to the concept 'conlang'."

2.2 Conlangs
Reading through other emails for the remainder of 1991, it becomes apparent that conlang is being used as a common noun to mean a construtcted language and not just as the name of the list. The word conlanger also appears at this early date.

Whatever may have been the reason for these glossopoeists to have coined the term in the first place, it has long been understood to be derived from construtcted language in contrast to natlangnatural language, the term natural language being already established as a linguistic term to mean "[a]ny language which is, or once was, the mother tongue of a group of human beings" [R.L. Trask, "A Dictionary of Grammatical Terms in Linguistics", London & New York, 1993; Trask does give a second meaning for 'natural language', but that does not concern us here].

While generally on the Conlang list the term conlang is understood to refer to the types of languages discussed below, i.e. auxlangs, artlangs and engelangs, it has also been used to describe (a) reconstructed languages such as Proto-Indo-Europaean, (b) revived languages such as modern Israeli Hebrew and (c) literary languages such as Classical Latin. In the case of (a), it is argued that these languages are the artificial creations of individuals or committees and only approximations of now extinct natural languages (indeed, the problem gets even more problematic with reconstructions for languages such as Nostratic since there is no consensus what Nostratic was or, indeed, if it ever existed). As regards (b), it is argued that there are are gaps in our knowledge about, say, Cornish or Biblical Hebrew and that modern revivals, often the work of individuals or committees, are conscious human structures and, in the case of (c), that a language such as Classical Latin was not the natural mother-tongue of any group but an artificial written medium that had to be learnt. In other words, the boundary between conlangs and natlangs is fuzzy and overlaps.

In his attempts to classify conlangs, Jan van Steenbergen includes all the languages discussed in the paragraph above. If you look at Jan's classification matrix, however, you will notice that his categories E (Reconstructed languages) and F (Modifications of natural languages) have only 11 rows each, whereas categories A, B, C and D all have 26 rows. So there is a difference on the one hand between categories E and F and categories A, B, C and D. The latter are what we tend more narrowly to think of as conlangs and we may call them glossopoeic conlangs; they have for some time been considered on the Conlang list under three broad groupings: auxlangs, artlangs and engelangs


3. Conlangs 1: Auxlangs

Auxlangs are artificial auxiliary languages, constructed to facilitate communication between speakers of different languages either over a region or, more commonly, on a global scale. The earliest known constructed auxlangs date from the 17th century.

From the very beginning two quite different approaches have been adopted by the authors of these languages. Adopting terms from epistemology, the French logician, mathematician, philosopher, Louis Couturat, and the French mathematician, Léopold Leau, dubbed them a priori and a posteriori (Couturat and Leau, 1903, Histoire de la langue univers, Paris: Librairie Hachette et Cie.)

  • "Il ya a, d'une part, des projets qui, pour des raisons diverses, ne tiennent aucun compte des langues naturelles, et qui sont des langues originales, construites de toutes pièces : nous les appelons systèmes a priori."
    There are, on the one hand, projects which, for various reasons, take no account of natural languages, and which are original languages, built from scratch: we call them a priori systems.
  • "Il y a, d'autre part, des projets qui, prenant pour modèle les langues naturelles (particulièrement les langues européennes), s'efforcent de les imiter et leur empruntent presque tous leurs éléments : nous les appelons systèmes a posteriori."
    There are, on the other hand, projects which, taking natural languages (particularly European languages) as a model, strive to imitate them and borrow nearly all their elements from them: we call them a posteriori systems.

Whether it is in fact humanly possible to build an entire language from scratch without taking any natural language whatever into account is doubtful. No constructed language can, in my opinion, be strictly a priori; indeed, Couturat and Leau may have chosen the term a priori in order to underscore the impracticality, as they saw it, of these languages, for they had already established the "Delegation for the Adoption of an International Auxiliary Language" (Délégation pour l'Adoption d'une Langue Auxiliaire Internationale) two years before they published "Histoire de la langue universelle."

But their terms, for better or worse, were adopted by interlinguists and have stuck ever since (though in recent years they have acquired a somewhat different meaning in some conlang circles; see below).

Couturat and Leau also used a category called "mixed" (systèmes mixtes) for languages that combined elements of both types. Almost all such languages will be found to be basically a posteriori, but with some a priori features. We then get involved in arguments about how many a priori features are needed to make the language "mixed". While most will agree in labeling Volapük "mixed", there will be argument about Esperanto. I have no wish to be involved in such arguments and shall not be using the category "mixed" below.

Kenneth Searight's auxlang, Sona, however, which was published twenty two years after Couturat & Laeu's book, departs substantially from their definition of a posteriori nor is it "mixed" in the sense that they used the term. I deal with the language separately below.

Since 1996, these languages have had their own mailing list known as Auxlang.

3.1 Origin of the term "auxlang" & of the Auxlang list
The name 'Auxlang' appears to have been coined by Lars Mathiesen in an email of 3rd November 1992 as the name of a suggested new list: "I am well aware that there are others on this list whose interests are mainly in IALs, and for whom discussions of artlangs and research langs may be unwelcome distractions. (I sometimes get the impression that they even feel them to be deliberate obstructions.) Perhaps it is time to create an auxlang list?" Up to that time, as Lars' message implies, the most common term for languages in this group was IAL (International Auxiliary Language), though some mailers simply referred to them as "auxiliary lang(uage)s."

It is interesting to see that at this early date we already have a tripartite division into "IALs … artlangs and research langs." Lars' suggestion, however, was not taken up and all three groups continued on the Conlang list. On the 19th March 1993 we find Robin Gaskell, who had been posting quite a lot of mails about Glosa, writing: "Friends, This is a follow up to one or two of my earlier blips on the possibility of a specialist Mail List to bear down on the IAL concept"; but again there was no support for any such specialist mail list. On the next day, by agreement with John Ross, the Conlang mailing list was moved to, Lars became the new list owner and things carried on with all types of conlangs discussed on this list.

Then after a "busy weekend" Lars wrote on 15th January 1996: "This has certainly been a busy weekend. I think I will repeat something that I have said before: I would personally prefer for discussions of the merits of various conlangs as auxlangs to be held elsewhere. And that goes thrice for discussions of the IAL idea/cause and the history of IAL movements. My offer to run an AUXLANG mailing list on this server still stands." The "busy weekend" had seen a veritable deluge of emails either pro-Esperanto or anti-Esperanto and traffic levels far higher than we would tolerate now. Not surprisingly, on the very next day, after the list had received another 50 or so more such Esperantine exchanges, Lars announced: "by administrative fiat and without further ado, the AUXLANG list is born."

3.2 Bâlaybalan: the world's first auxlang?
The earliest auxlang that I know of is Bâlaybalan (Turkish: Bâleybelen; also transcribed in English as Bala-i-Balan, Balaibalan, Balaibelen etc.). According to the Turkish scholar, Midat Sertoğlu, Bâlaybalan was devised by Mehmed Muhiddin and published in 1580 as a common language for all the peoples of the Middle East with the eventual aim of it being used globally.

It does seem to be a mix of a posteriori and a priori elements; but as the author was not only familiar with Arabic (which he considered to the most perfect language), Persian and Turkish, he does seem to have taken them into account in the creation of his language, we should, perhaps, put it in the a posteriori camp. However, the language seems to have a greater a priori elements than the "mixed" systems Couturat and Leau considered in their book.

We find this mix of a postriori and a priori elements in the lexicon. The words are taken either from natural languages and changed (rather as Schleyer later did with Volapük; see below) or are created by the author under 'divine inspiration'. Of the grammar, Sertoğlu says: "The rules of its grammar were completely invented. He adopted the syntax used by the Arabs to organize sentences." By this I understand that the morphology is a priori but the syntax is a posteriori. It would seem, however, that rather than being specifically Arabic, the syntax is right-branching in the manner of Semitic languages rather than left-branching like Persian and Turkish.

3.3 A priori auxlangs
The concept of an international auxiliary language, as I have written above, dates back at least to the 17th century with attempts such as Logopandecteision (Sir Thomas Urquhart, 1652), Ars Signorum (George Dalgarno, 1661), and An Essay towards a Real Character, and a Philosophical Language (Bishop John Wilkins, 1668). The languages were a priori, that is their authors attempted to construct them from first principles and not from natural languages. A priori auxlangs have continued to be created, for example, Solresol "Langue musicale universelle" (François Sudre, 1866), Ro (Rev. Edward Powell Foster, 1906), Babm (Fuishiki Okamoto, 1962). With the exception of Solresol, these languages never attracted any significant following. The interest in Solresol is probably because it is more interesting than the others in that it can be communicated through a variety of media; a small community of Solresol enthusiasts scattered across the world are able to communicate with one another thanks to the Internet.
3.4 A posteriori auxlangs: 17th to 19th century
The one whose name is best known to the general public is Esperanto (Dr. Ludovic Lazarus Zamenhof, 1887). Unlike the languages in the preceding paragraph, Esperanto is a posteriori, that is, it is constructed from elements taken from natural languages, its phonology being essentially Slavic, and its vocabulary being derived primarily from the Romance languages, with a lesser contribution from the Germanic and Slavic languages. Its morphology, although schematic, is clearly derived from western Indo-European patterns. It has attracted a larger following than other auxlangs and still attracts a following in Europe, eastern Asia and North and South America.

Esperanto, however, is by no means the only a posteriori auxlang. Above, we saw that Bâlaybalan, which preceded it by some three centuries, already had a posteriori elements; and in Europe, it was preceded by more than two centuries by Grammatica linguae universalis missionum et commerciorum (Philippe Labbé, 1663). This was largely a posteriori, being an early attempt to simplify Latin. It did, however, contain some a priori elements; for example, instead of trying to regularize the Latin personal pronouns, he had: a = I, me; e = you [sing.]; o = he, him, she, her, it; as = we, us; es = you [pl.]; o = they, them. How far it had departed from its Latin origin may be seen in the opening words of the 'Our Father': "Oh pat asa, u eno ni cels" (cf. Latin: "Pater noster, qui es in caelis").

More recent pre-Esperanto examples are Communicationssprache (Joseph Schipfer, 1839), Universalglot (Jean Pirro, 1868) and Volapük (Johann Martin Schleyer, 1879–1880).

Indeed Volapük, which though mainly a posteriori did have some a priori features, became the first auxlang to gain a sizable following; by 1889 there were some 283 clubs, 25 periodicals in or about Volapük, and 316 textbooks were available in 25 different languages; in that year the third Volapük conference was held in Paris (the first two conferences being in 1884 in Friedrichshafen and 1887 in Munich). It was decided to hold the third conference entirely in Volapük; this showed that although its morphology was regular and had no exceptions, it was far too complicated. All attempts to reform the language, however, were resisted by its founder and the movement quickly fell apart. Some remained faithful to the language, others tried various reformed projects and others turned to the much simpler Esperanto that had appeared two years earlier. In the 1920s, Arie de Jong, with the consent of the leaders of the small remnant of Volapük speakers, made a revision of the language which he published in 1931. This was accepted by the few speakers of the language and the language still retains a few speakers (said to be about two dozen today).

Out of the break-up of Volapük came, amongst others, Idiom neutral (Akademi Internasional de Lingu Universal under the leadership of Waldemar Rosenberger, 1902). The Akademi had originally been the Kadem bevünetik volapüka, i.e. International Academy of Volapük. In 1908 the Akademi abandoned Idiom Neutral in favor of Latino sine flexione [LSF] (Giuseppe Peano, 1903). This was, as the name suggests, a simplified version of Latin, but sticking more closely to its Latin origin than had Philippe Labbé's language; cf. the opening of the 'Our Father': Patre nostro, qui es in celo; On 26th December 1908, Peano was elected director of the Akademi which, in the following year, was renamed Academia pro Interlingua.

3.5 A posteriori auxlangs: 20th till today
In 1907 Couturat and Leau's Délégation pour l'Adoption d'une Langue Auxiliaire Internationale set up a committee to recommend the adoption of an international auxlang. It considered LSF, Idiom Neutral, some other Volapük off-shoots, as well as Esperanto. Towards the end of the conference the committee members received a proposal by an anonymous author identified as "Ido" which proposed to reform Esperanto, removing its "flaws." The committee decided in favor of Esperanto, but with the reforms outlined by "Ido." A rift developed between the Esperanto movement and the Délégation, especially when it was discovered that "Ido" was Louis de Beaufront, who was the Esperanto representative on the committee, and Louis Couturat, one of the co-founders of the Délégation. The Délégation's 'reformed Esperanto' became known as Ido and still retains a small following.

Other proposed reforms of Esperanto have also appeared and still occasionally appear; but none have attracted any following. Sadly, however, bitterness over the Esperanto-Ido schism rumbles on, with accusation and counter-accusation about Esperanto's "flaws", what did and did not happen at the Délégation's conference of 1907, whether Beaufront was ensuring Esperanto got chosen or had betrayed Esperanto and so on, and on and on … . Indeed, such accusations and counter-accusations were the main substance of the voluminous exchange of emails on Conlang that led to the setting up of Auxlang in January 1996.

Two of the committee members moved away from the Esperanto-Ido type of auxlang and developed more 'naturalistic' auxlangs: Occidental (Edgar de Wahl, 1922) and Novial (Otto Jespersen, 1928). Occidental gained many adherents and was the second most popular auxlang before the World War II; in 1949, the following de Wahl's death, its users changed the language's name to Interlingue, but today the language is more commonly known by its older name. There are, I believe, some who still use the language but many adherents switched their allegiance to Interlingua (International Auxiliary Language Association, 1951). Novial attracted a following and during the 1930s was gradually modified towards greater naturalism; Jespersen died in 1943 and the language barely survived WWII, its remaining adherents going over to Interlingua in the 1950s. There was an attempt made in 1998/99 on Auxlang to revive Novial; but the attempt was bedeviled by the sort of factionalism that followed the break up of Volapük and the Esperanto-Ido schism. Meanwhile Interlingua still has a following.

Many other auxlangs have been constructed; two of the more recent ones are Frater (Phạm Xuân Thái, 1957), which has a Graeco-Latin vocabulary with a non-western grammar more reminiscent of Indonesian, and Glosa (Ron Clark and Wendy Ashby, 1972–1992), which also has a Graeco-Latin vocabulary and an isolating grammar. It is, however, impossible to give anything like an exhaustive list and they are still being constructed. The number of auxlangs are most certainly in the hundreds and may, indeed, now be into four figures; the supply vastly exceeds demand.

3.6 Sona: a posteriori masquerading as a priori?
Sona is an auxlang created by Kenneth Searight (Searight, K, 1935, Sona: an auxiliary neutral language, London: K. Paul, Trench, Trubner & Co.). Searight thought the auxlangs of his time were too Eurocentric and not universally neutral. Like the creators of a priori auxlang, he wanted a truly neutral auxlang. However, he thought the approach taken by authors of such languages had not been practical. He, therefore, used inspiration from many diverse natlangs, including English, Arabic, Turkish, Chinese and Japanese, to create an eclectic yet regular and logical language. It has a 375 radicals, based on the categories in Roget's original thesaurus. Further words are derived by compounds, e.g. ra "male" + ko "child" → rako "boy." The name of the language is derived from so "help, auxiliary" + na "neutral [thing]."

The grammar is original and, as in very many auxlangs, is agglutinative with a strong tendency towards being an isolating language. It never caught on, but there is now a small community on the Internet which is interested in reviving the language and using it.

3.7 Regional auxlangs
As well as the languages above which were intended to be international on a global scale, one also finds regional auxlangs, e.g. Tutonish (Elias Molee, 1902) and other inter-Germanic auxlangs; Afrihili (K. A. Kumi Attobrah, 1970) as a lingua franca for the whole of Africa; Eurolang (Phil Hunt, 1995-1998) to serve an an auxlang for the European Union; Slovianski (Ondrej Rečnik, Gabriel Svoboda, Jan van Steenbergen & Igor Polyakov, 2006) and many other Pan-Slavic auxlangs; once again it is not possible to give an exhaustive list.
3.8 My own experience of and attitude towards auxlangs
As for myself, I discovered Esperanto when I was about 10 years old and was intrigued that someone had actually invented a language and it seemed a good idea that a language should be invented to help mankind. At ten I had no idea there was such a thing as the Esperanto movement or even whether anyone actually used the language. Rather, it inspired me to try to do the same!

Right through my teens I churned out two or three auxlangs each year. To begin with they were modifications of Esperanto, but gradually they took on more distinctive characters, usually influenced by whichever natural language I had been studying at the time. All this came to a halt when I was an undergraduate; university was followed by marriage and a family and that more or less put a halt to language construction.

Then sometime in the Spring of 1997 I joined the Conlang and Auxlang lists. Experience on the Auxlang list, however, gradually made me have more and more misgivings about artificial auxiliary languages and I concluded that any successful IAL was more likely to be a development of a natural language. In particular, I became (and still remain) very disenchanted with the bigotry and bitterness of IAL politics; the crunch time came during the 'Novial Wars' when different schools of "revived Novial" traded insults amongst themselves! I left Auxlang and have no intention of ever returning nor of getting involved with the politics or promotion of any IAL (though I am happy to discuss IALs with sane people purely as constructed languages per se).

My feelings towards auxlangs are very similar to those expressed by Rick Harrison in his "farewell to auxiliary languages". Those, however, who wish to find out more about such languages will find much of interest in James Chandler's "International Auxiliary Languages" page.

3.9 WARNING: use of terms a priori and a posteriori.
The the terms a priori and a posteriori, as defined by Couturat and Leau, have been found useful by interlinguists and others when talking about auxlangs; the terms, with those definitions, are not useful and, in my opinion, at best somewhat meaningless and at worst simply confusing when applied to artlangs & engelangs. It is my personal opinion that the terms are best reserved, as they are in interlinguistics, for auxlangs with the meanings given by Couturat and Leau.

One may, however, find these terms being used in some conlang circles thus:

  • a priori = "with an original lexicon [no matter how the author composed it]";
  • a posteriori = "with a lexicon obviously derived from natural language(s)."

This is even further away from their epistemological origin and, as this usage is not intuitive, it is not always understood.


4. Conlangs 2: Artlangs

Artlang is short for art-language (or, if you prefer, art language, without the hyphen), i.e. language which has been constructed as a work of art in itself.

4.1 Origin of the term "artlang"
The earliest use of the term art-language that I am aware of was by J.R.R. Tolkien in his essay "A Secret Vice". Such a language is not concerned with "base considerations of the 'practical', the easiest for the 'modern mind', or for the million - only a question of taste, a satisfaction of personal pleasure, a private sense of fitness." In other words, an art-lang is a personal creation and its criteria of fitness, appropriateness and so on are essentially subjective.

Although we may admire (or not) an artist's skill, their arrangement and composition and so forth, we will not all agree on liking the same art. I once came across someone who raved over the art of Jackson Pollack; I do not share his enthusiasm. Similarly while I find Quenya a great and moving art-language, this view is not shared by all. I recall an Italian remarking that the language sounded silly, like a parody of his native language. So just as the creation of an artlang is essentially an subjective activity, so is its appreciation by others.

Notice that Tolkien explicitly said that artlangs not concerned with "base considerations of the 'practical', the easiest for the 'modern mind'", in other words: artlangs are not auxlangs. There are not, therefore, concerned with questions such as "Do I create a language from scratch so that it is truly neutral?" or "Do I create my language from those that most people know, so it will be easy to learn?" Therefore questions about whether the author took an a priori or an a posteriori approach is meaningless. What concerns their authors is, as Tolkien wrote, "only a question of taste, a satisfaction of personal pleasure, a private sense of fitness.

Nevertheless, you may find an artlang referred to as a priori or an a posteriori; it refers simply to the vocabulary, as explained in the WARNING above. It adds little, in my opinion, as in practically every case the question of whether the vocabulary is original or derived is conditioned by the type of artlang that is being created.

The earliest use of the term artlang I can find on the Conlang list is from John Chalmers 31st October 1992 "Other authors who have at least provided glossaries of artlangs in their novels are Frank Herbert in "Dune," though his language(s) borrow(s) heavily from Arabic, Urdu?, English, etc., and E[dgar ]R[ice ]B[urroghs]." This, however, looks like a reply to some earlier posting. It is clear, however, the term is assumed to be known by list members. The email refers to the fiction of two authors; it will be found that artlangs are almost inevitable associated with some fictitious world or community.

4.2. Some early artlangs?
Often, as in the examples, quoted by John Chalmer aboves, we get only a glimpse of the artlang and cannot be certain to what extent it has been developed. An often quoted example is a page on the "Utopian alphabet" which Peter Giles of Antwerp provided for Thomas More's "Utopia"; the pages contains a tetrastichon (i.e. four-line stanza) in the Utopian language together with a Latin translation:
Vtopos ha Boccas peula chama polta chamaan.
Bargol he maglomi baccan soma gymnosophaon
Agrama gymnosophon labarem bacha bodamilomin.
Voluala barchin heman la lauoluola dramme pagloni.
Utopus me dux ex non insula fecit insulam.
Una ego terrarum omnium absque philosophia
Civitatem philosophicam expressi mortalibus.
Libenter impertio mea, non gravatim accipio meliora.

The general, Utopos, made me an inland from [being] a non-island. I alone of all lands, without philosophy, have portrayed the philosophical city to mortals. Gladly I share my things, not unwillingly I welcome better things.

The page was omitted from the 1517 edition and subsequent editions of the book. The language, however, has enough internal consistency to suggest it had been worked out with some care.

It will be noticed this and those mentioned in John Chalmers' email of 31st October 1992 are languages of fictitious people. Tolkien speaks of Quenya's "hypothetical historical background (a necessary thing as a constructor finds in the end, both for the satisfactory construction of the word-form, and for the giving of an illusion of coherence and unity to the whole)" [J.R.R. Tolkien, op cit.]. Arika Okrent mentions the inhabitants of the moon in Francis Godwin's "Man in the Moone" (1638), Gabriel de Foigny's "La Terre Australe Connue" (1676) and the Jonathan Swift's "Gulliver's Travels" (1726 - in fact Okrent mentions only Lilliputian, but there are samples of other fictitious languages in the book), the Newspeak of George Orwell's "Nineteen Eighty-Four" (1948) and Anthony Burgess's "A Clockwork Orange" (1962 - the language is Nadsat) and the "jabber of countless works of science fiction" [Arika Okrent, 2009, "In the Land of Invented Languages", Spiegiel and Grau, 978-0-8129-8089-9 (paperback), 978-0-385-52788-0 (hardcover)]. One can site earlier examples: Dante Alighieri (c.1317) Papé Satàn, papé Satàn aleppe! (Inferno VII, 1) and Rafél maí amècche zabí almi (Inferno XXXI, 67); Aristophanes ( 425 BCE) iartamàn exárxan apissóna sutra (ἰαρταμὰν ἐξάρξαν ἀπισσόνα σάτρα - Archarnians, line 100). Indeed, this sort of thing has probably been going on ever since people told stories.

Although some of the above examples, such as Utopian, may represent a language that has been worked out with some care, most are just bits of vocabulary and, perhaps, a few phrases and sentences to serve the story. With Tolkien, however, we find quite the reverse: Middle Earth and its pseudohistory and mythology were developed to serve his languages.

4.3 Some modern artlangs
Tolkien had been working on his two most developed languages, Quenya and Sindarin from 1915 at least; early versions of the languages are given as samples in "A Secret Vice" but the world at large met them in "The Lord of the Rings" (LotR) published in three volumes during 1954-55. The two languages are quite distinct but with ingeniously and methodically worked out pseudo-diachronic development both appear to be derived from 'Primitive Elvish' (i.e. Proto-Elven). How distinct the languages were may be seen in that when the Quenya-speaking Noldor returned to Middle Earth after nearly three and a half millennia, the Noldor and the Sindar were not at first able to understand one another (in the end the Noldor took to speaking Sindarin). The LotR contains fragments of several other constructed languages, e.g. other languages in the Elven group; Adûnaic of the men of Númenor; Khuzdul, the secret language of the Dwarves; Orkish and the Black Speech. For details of Tolkien's language's, see the Ardalambion web-site. Many conlangers have said that it was Tolkien's work that inspired them to begin their own creations.

Since Tolkien, we've had Klingon, the language constructed by the linguist, Marc Okrand, for the alien race of Klingons in Star Trek II and subsequent series. The language caught many people's imagination and we now have the Klingon Language Institute which provides all sorts of support for Klingon speakers and would-be Klingon speakers or, indeed, for anybody who is interested. More recently, in 2009 Paul Frommer, a professor at the Marshall School of Business with a doctorate in linguistics, constructed the language of the Na'vi, the sapient humanoid indigenous inhabitants of the fictional moon Pandora in the film Avatar. This has already attracted a following and one can find out more or learn the language at the Learn Na'vi web-sitee.

With the advent of the Internet, many constructors of artlangs have made their languages known. It would be impossible to give a full list, especially as their number constantly increases. I just name a sample here.

Many, as we would expect, are those of fictional peoples in fictional worlds, such as Tsolyáni (late 1940s), M. A. R. Barker's language of the fictional world Tékumel, and Mark Rosenfelder's Verdurian (1995), the language of Verduria on the planet Almea.

With Sally Caves' Teonaht (1962), the "language of the Teonim of Teon … a region that surfaces and submerges most often within the Black Sea, sometimes the Caspian. It is surmised that the Teonim are perhaps from the Caucasus, or--given their vanishing propensities, their scant but bizarre appearances in our history --from somewhere else entirely", we possibly have people from elsewhere and their fictional land of Teon is certainly odd with its propensity to disappear and reappear, sometimes in the Black Sea and other times in the Caspian. The language itself is a highly elaborated one and some consider it comparable in excellence to Tolkien's Elven languages.

Other languages are firmly placed on earth but in a different timeline from our own. One such as Matt Pearson's Tokana (1996), a language "spoken by an imagined people, also called the Tokana, who live somewhere in northwestern North America, but in a time quite remote from our own - either the distant past or the far future, or perhaps some alternate history", and Andrew Smith's Brithenig(1996), "a constructed language based on the premise that Vulgar Latin took root in Britain the same way it did in France, Spain, etc., thus resulting in a Romance language that went through a similar development as Welsh."

Brithenig gave rise to a set of conlangs which are usually known as 'bogolangs'; they apply the model of the diachronic phonology of one natlang to a natlang base from a different language group (usually, though not always, Vulgar Latin). The diachronic phonology is generally known by the authors of these languages as the Grand Master Plan (GMP). Externally, of course, the GMP is imposed upon the base language; but according to the internal history of their alternate world the sound changes are due to substrate influence. Examples of such languages are: Breathanach (VL with Gaelic GMP), Wenedyk (VL with Polish GMP), Þrjótrunn (VL with Icelandic GMP), etc. A "Gothic as Castilian" has been proposed, but as far as I know it has not been developed. One could, in theory, use any natlang as a basis and apply the diachronic changes of another one to it.

Brithenig and these other languages, apart from Þrjótrunn, inhabit Ill Bethisad the alternate time-line in which these and other some other fictional languages exist; here we find also Jovian which is a little different in that its base is Classical Latin subject in a general way to diachronic phonology of the Alemannic German dialects.

The term 'bogolang' has been objected to from time to time because it does not describe the languages and is felt by many to be offensive. But the term has been around since at least 2009 and so far, despite several discussions on the Conlang list, other suggestions such as graftlang, chimeralang, masqlang and so on have not found much favor.

Padraic Brown's Kerno has sometimes found itself regarded as another bogolang, probably because it became known to the conlang community soon after Brithenig appeared and it is one of the languages of Ill Bethisad. The author had, however, begun work on this language before Brithenig came along; it is not a bogolang nor is it a coherently worked alternate history language (e.g. in A Concise Tabulation of the Cornovian Dialect we find that the initial consonant mutations given page 3 do not cohere with the 'Sound Changes from Latin to Modern Cornovian' given on pages 80-83) but rather, as indeed the author himself says, "a fun frolic through the worst aspects of both Romance and Celtic languages ... [which] takes seemingly ill-fitting and incongruous bits of this and that ... ."

I should perhaps, add, that Brehonnecq, which is also listed among the languages of Ill Bethisad, was intended as a dialect of Kerno but, beyond a few preliminary notes, was not further worked out.

Yet other conlangs are set on earth apparently within our own history but have generally, so to speak, "remained beneath the radar" such Dirk Elzinga's Tepa (1995), the language of the "small Tepa community [which] did not survive to the twentieth century; it was severely decimated by a smallpox epidemic … The location of the last Tepa pueblo is now unknown. It was most likely located on one of the tributaries of the Colorado River, but it is presumably submerged beneath the waters of Lake Powell, which was created when the Glenn Canyon Dam was built." Another such language is Jan van Steenbergen's Hattic (1996) is a "member of a fictional branch of the Indo-European language family, spoken in an unidentified republic somewhere in North Russia."

Neither the fictional speakers of Tepa nor of Hattic are claimed to have had any actual effect on the history of our world. However, the fictional history of Jörg Rhiemeier's Old Albic (2010) does claim that traces of the language persist in our world till the present day. The language is a member of a fictitious language family, the Hesperic language family which "reached its climax in the Early Bronze Age" but which was eventually replaced by Indo-European languages, leaving only trace "as loanwords in the Indo-European languages in central and western Europe, and a number of geographical names, especially river names." Old Albic was the language of the Elbi, the "British Elves … humans with an 'Elvish' culture (similar to Tolkien's Elves)" who were the pre-Celtic inhabitants of British Isles; their language was a substratum to Insular Celtic and some of the peculiar features of the modern Insular Celtic languages are considered due to this substratum.

Last but by no means least are the languages of 'micronations'. Unlike the fictitious peoples and nations above, these have some sort of real existence in our world though they are not recognized by any world governments or major international organizations. Such micronations often occupy no more than their creator's house or maybe just a room in the house! The advent of the Internet has allowed such nations to have members scattered all over the world who interact mostly by electronic means; they have become in fact networking groups engaged in role playing. Probably the most well known is the Kingdom of Talossa, founded by the 14-year-old Robert Ben Madison of Milwaukee on December 26, 1979, and consisting just of his bedroom. Over the years this has grown and evolved into a politico-cultural role-playing game, complete with parties, elections, laws and government institutions, and several online newspapers, and even its own fictional history prior to 1979. Indeed, in 2004 there was a split and we now have both a Kingdom of Talossa and a Republic of Talossa. In 1980 Robert Ben Madison constructed the Talossan language (El Glheþ Talossan) for this micronation and The Association of Talossan Language Organisations (ATLO) now maintains the Talossan language web-sitee which gives the history and grammar of the language, provides tutorials and has much else.

4.4 Why artlangs?
Maybe the difference between artlangs and auxlangs is best shown by an amusing but instructive incident that Arika Okrent tells in "In the Land of Invented Languages." Arika had attended a Klingon gathering known as a qep'a' and was speaking with a prominent Esperantist about it; he was confused and could only ask: "But what are they doing? Really, what are they doing?"

Yes indeed, what are artlangers doing? Arika's answer is a good one: "They are doing language for language's sake, art for art's sake. And like all committed artists, they will do their thing, critics be damned."



5. Conlang 3: Engelangs

Engelang /'ɛnʤlæŋ/ denotes a language that is designed to specific objective criteria, and engineered to meet those criteria.

The existence of a third group of conlangs was apparent from the earliest days of the Conlang list. As we saw above (Section 3.1) Lars Mathiesen wrote in November 1992: "there are others on this list whose interests are mainly in IALs, and for whom discussions of artlangs and research langs may be unwelcome distractions." The term artlang has stuck, as we saw above, and the artificial IALs have been known as auxlangs since the establishment of the Auxlang list in January 1996; but it took until 2001 before this group received the name now currently used.

5.1 Origin of the term "engelang"
On the early Conlang list, the 'research langs' that Lars referred to were Loglan and, more often, Lojban; the former had been proposed in 1955 by James Cooke Brown (JCB) to test the Sapir-Whorf hypothesis that the structure of a human language sets limits on the thinking of those who speak it. He decided that a languages whose grammar was based on predicate logic would free its speakers from the constraints imposed by natural languages and he called his language Loglan ← a logical langauge. In 1960 JCB published an article about the language in Scientific American which attracted interest and he eventually set up "The Loglan institute" (TLI). In the late 1980s differences arose between JCB, who wanted to continued to claim legal restrictions on the use of his language, and others who wanted to make the language freely available and encourage its use as a real language. The latter broke with the TLI and set up the Logical Language Group and developed their own version of the language which they called Lojban ← loj + ban, being the rafsi (i.e. abbreviated forms) of logji (logic) and bang (language) respectively. Hence as both languages, in effect, call themselves 'logical language' the term loglang soon became a generic name for 'research langs'.

But it became apparent that some 'loglangs' had little to do with logic; also the adherents of many a posteriori auxlangs, e.g. Esperanto, claim that their chosen auxlang is 'logical'. So 'loglang' came to be felt somewhat vague and imprecise in meaning. According to And Rosta, "in January 2001 Rick Harrison objected that 'loglang' should mean 'conlang based on formal logic'" [And Rosta, 20th July 2007, "Origin of the term 'engelang", email to the Conlang list]. Certainly that is the way the term loglang is now used.

This meant, as And says in the email, that a term was needed for the 'research langs', loglangs and other kindred conlangs. It was in the Summer of 2001, as a result of emails between John Cowan and And Rosta that engelangengineered langage (the silent medial -e- being needed in English spelling to indicate the 'soft-g' [ʤ]), with the pronunciation and meaning given in the first line of this Section.

5.2 Early engelangs?
Some people claim that the a priori auxlangs mentioned above in Section 3.2 are engelangs in that they were engineered by their creators to be 'philosophical' and that, therefore, they are 'logical'. We have seen that in practice 'logical' is a somewhat vague notion and many a posteriori languages have claimed to be 'logical'. Indeed, one does not have to dig too deeply before coming across claims that certain natural languages are logical; certainly such claims have been made about both French and Latin in the past. As for being 'philosophical', one cannot deny that the a priori way in which their vocabularies were developed were based upon philosophical conceptions of their creators, and certainly one can claim their creators did a good deal of linguistic engineering. But it cannot be denied that the principal aim of their creators was to provide a neutral and international means of communication, i.e. they were designed and engineered with the main purpose of being auxlangs; and not all their design criteria were objective and measurable.

Apart from the odd smattering of words and phrases from potential artlangs, some of which may have been more fully worked out, practically all conlangs prior to the 20th century were intended as auxlangs. The only possible exception may have been Lingua Ignota, which I shall consider in Section 6.

5.3 Loglangs and loglans
As I wrote above, the engelangs that got most frequently discussed on the Conlang list were Loglan and Lojban. Many Lojban supporters use the term 'Loglan' as a generic term to refer to both their own language, and JCB's Loglan, which they refer to as "TLI Loglan" when in need of disambiguation. Many supporters and members of The Loglan Institute find this usage offensive, even though the United States Patent and Trademark Office did uphold that Loglan was a generic term and could not be trademarked. Thus we find in the early Conlang period loglang and loglan used interchangeably to mean "engelang." This is all very unsatisfactory. In this section I used Loglan to mean JCB's language as defined by The Loglan Institute, and Lojban to mean the language, originally derived from JCB's early Loglan, and developed by the Logical Language Group.

The term 'loglang' is now properly confined to those languages which are based on formal logic - in practice predicate logic. The term 'loglan' is sometimes used to mean any conlang ultimately descended from JCB's original language. Thus Voksigid (1991-1992) may be termed a loglan, but many feel it has departed too far from formal logic to be termed a loglang. James F. Carter's Gua\spi (1991) is another loglang inspired by JCB's Loglan. Rex May's Ceqli (1996) is also sometimes quoted as loglang inspired by JCB's Loglan; while it is true that this was its inspiration, it has been influenced also by Mandarin and English and is offered by its author as an auxlang. I shall consider it in Section 6 below.

5.4 Some other engelangs
Another conlang concerned with the relationship of language and thought is Láadan (1984). It's author, Suzette Haden Elgin constructed "a language specifically designed to express the perceptions of human women" because "existing natural languages don't do that adequately."

In the words of its author, John Quijada, Ithkuil "exists as an exercise in exploring how human languages could function, not how human languages do function" and its goal is "to achieve the purpose of cognitive exactness and conciseness of communication." The language "slowly and painstakingly evolved" from ideas John had in his teens. Arika Okrent gives Ithkuil the date 1995 in her book; presumably there was a posting about it on the Web that year. But it was not till early 2004 that the actual first version of Ithkuil was posted on the Internet. In mid-2007 an alternative version, called Ilaksh, was posted which gave the language a simpler phonology. But during this work, John realized that "there were many aspects of Ilaksh design that could be incorporated back into Ithkuil without the constraints of Ilaksh on the number of consonants and vowels" and he rethought a few fine points of grammar. Thus in July 2011, John presented the "third incarnation of the language" which replaces the earlier two earlier version. He has retained the name Ithkuil, an anglicized form of iţkuîl, and regards this as the definitive version of the language.

Also concerned with the way language may affect thinking is Jim Henry's gjâ-zym-byn (gzb) which Jim began working on in early 1998 as a psychological experiment. Of the language, Jim wrote: "I aimed for a balance of exoticity and learnability so I could try out new grammatical and semantic structures and see how they affect my thinking." Unlike the authors of some languages, Jim actually uses gzb "pretty much every day, thinking in it, praying in it, sometimes talking to" himself in it, even on the days when he happens not to read or write it.

The translator and linguist, Sonja Elen Kisa of Toronto, designed Toki Pona (2001) to be "a simple language designed to express the most, using the least." It has only 14 phonemes and 123 root words and "is designed to steer us towards positive and constructive thinking" and "can become a sort of 'yoga for the mind'."

Not all engelangs, however, concern themselves with the relationship of language, thought and cognition. Some are more clearly of an experimental nature. On such is Tom Breton's AllNoun (1990; version 0.8 1995) which, as its name implies, was an attempt to see if a language could be constructed with only one part of speech, i.e. noun, and "to support any advantages that can be obtained with a maximally simple grammar." Another language that claims to have dispensed with verbs is Kēlen; I will not, however, discuss it here as it otherwise has much of the feel of an artlang. I reserve discussion of Kēlen to Section 6.3 below.

R. Srikanth's experiment in constructing Lin (1999) was to achieve as spatially compact a language as he could. He achieves this partly by an interesting and innovative system of enneasemy (a word having nine meanings) and "cements" (inter-word space morphology, including white-space, to disambiguate words). Examples of the language are: h v i2b = "the (human) being sees the interesting book"; i5o m = "the important agreement is possible."


6. Conlangs 4: Auxlang, artlang, engelang or what?

In Section 5.2 above I mentioned Lingua Ignota and Balaibalan and in Section 5.3 Ceqli as examples of languages which present certain problems when assigning languages to one of the three groups: auxlang, artlang, engelang. So let us consider these and a few other 'problems.'

6.1 What is Lingua Ignota?
I wrote above that practically all conlangs prior to the 20th century were intended as auxlangs. An exception may be Lingua Ignota. So is it an artlang or an engelang or what?

The twelfth century visionary, poet, composer, naturalist, healer, and theologian, Hildegard von Bingen (also known as St Hildegard) is credited as the author of the earliest known conlang, Lingua Ignota (c. 1150s). It is not clear, however, how far the language was developed; all we have is an a priori vocabulary of some 900 words, mostly nouns but with a few adjectives. In the extant examples we have, these are used with Latin, e.g.

O orzchis Ecclesia,
armis divinis precincta,
et iacinto ornata,
tu es caldemia stigmatum loifolum
et urbs scientarum.
O, o, tu est etiam crizanta
in alto sono
et es chortza gemma
O measureless Church,
girded with divine arms
and adorned with jacinth,
you are the fragrance of the wounds of nations
and the city of sciences.
O, o, you are also anointed
amid lofty sound,
and you are a sparkling jewel.

Not only are the five words inserted into a Latin text, but one word, loifolum, is given a Latin genitive plural ending! The incomplete nature of Lingua Ignota is reminiscent of the scraps of artlangs noted in 4.2 above. What the purpose was of Hildegard's language is not clear; it was certainly not just for use in religious canticles since among her list of nouns are many botanical words and names of body-parts and illnesses. Hildegard was, as I noted above, a naturalist and healer, and this must account for these words in Lingua Ignota. She was also a mystic and visionary and attributed her "unheard music" and her "unknown language" (Lingua Ignota) to divine revelation. The language, such as it is, appears to be a personal artlang but much remains unknown.

There are other conlangs also that are not always easy to place in the auxlang, artlang, engelang groups.

6.2 Auxlang or engelang?
I wrote above that some people claim that the a priori auxlangs mentioned in Section 3.2 are engelangs and certainly it can well be argued that they were engineered to reflect the philosophical world-view of their authors. But those authors intended them as auxlangs, no matter how unsuitable we may judge them to be. But we do not need to go back to earlier centuries to find conlangs that have a foot, so to speak, both in the auxlang and in the engelang camp.

One such language is World Speedwords whose author, Reginald Dutton offered it as a global system of "ordinary writing" (i.e. the modern Roman alphabet) at shorthand speed. Speedwords began its existence as "International Symbolic Script" in 1935; Dutton's friend, the Rev. F.W.G.Foat, suggested the language would be more likely to catch on if it could be pronounced. The result was a thorough revision and the first version of Speedwords in 1936. It was revised ten years later and again in 1951, which was its final form. All this time Dutton was engineering his language to meet the goal of a pronounceable universal briefscript.

Ceqli (1996) was another off-shoot of Loglan; but, unlike the others in Section 5.2 above, its author, Rex May, later publicized it as an auxlang. In 1999 Rex wrote "Ceqli began as an attempt to reform Loglan and make it more user-friendly. Since that beginning, it has departed more and more from Loglan, and has become a language based on Mandarin grammar with roots from everywhere."

The Ygyde web-page modestly claims that it is "a superb international auxiliary language." It is an a priori language and claims to be oligosynthetic, i.e. a language that uses very few morphemes, perhaps only a few hundred, and combines them synthetically to form statements. Its vocabularies are given in a set of tables and we are told that "Andrew Nowicki invented all the tables of the Ygyde language except preposition table. Patrick Hassel-Zein invented most of the grammar and half of the preposition table."

6.3 Artlang or engelang?
In his introduction to Liva (1997), Claudio Gnoli wrote: "Liva is constructed mainly for fun. It has no practical aim". So Liva is an artlang, then? Well, not exactly because Claudio went on to write: "It is conceived as a very unambiguous and logical language, at all its levels." In the 2002 version of the language, Claudio wrote "It can be seen as an experiment, trying to satisfy both the requirements of logic and the aesthetic preferences of its author." In other words, the language is both a loglang and an artlang. Indeed, the description of the language and its grammar (unfortunately, there no longer seems a version on the Internet) show the characteristics of an engineered language.

Kēlen's (1998) author, Sylvia Sotomayor, "decided to violate" the linguistic universal that all languages make a distinction between verbs and nouns and Kēlen "lost all of its verbs and became a language of nouns and particles." However, there is a great deal of "language for language's sake" about Kēlen and it does have very much the feel of a 'naturalistic' artlang; the language has also a rich and detailed culture that goes with it. Indeed, it is questionable whether the language is verbless; to many, Kēlen's "relationals" are a closed set of verbs. Although the design criterion that set the language off is an engelang principle, an artlang has been built upon it; it is at least in the cyan engelang-artland borderland of the Gnoli triangle below and many would put it firmly in green artlang territory.

Henrik Theiling describes Qþyn|gài (2206) as a polysynthetic language with "only one open lexical class, called full words or substantives." The goals of the language are:

  • Beauty
  • Neutrality
  • Good signal-to-noise ratio for speaking
  • Regularity

The last two goals are objective and measurable (though we might want to make more explicit what constitutes "good signal-to-noise ratio") and one could engineer a conlang with those goals in mind. I am not quite sure what is meant by "neutral" in this context; it is a term we often come across in the claims of auxlang authors. But beauty is a subjective value - in the eye of the beholder and the ear of the listener. There is no way it can be measured; we are dealing here with the aesthetic preferences of its author, which is characteristic of artlangs. Henrik writes that he would "describe it as greenish cyan" on the triangle below [email to the Conlang list, 30th July 2011].

6.4 The Gnoli Triangle
We have seen above that Claudio Gnoli described the purpose of Liva in both artlang and engelang terms. At sometime between December 1997 and March 1998 (regrettably Conlang list archive for this period appears to be missing), Claudio observed that many conlangs were of mixed type like his Liva and that one could think of conlangs as being located within a triangle whose vertices were artlang, auxlang and loglang, thus:

/ \
/  \
/    \
/      \
loglang ------ auxlang

This becameknown on the Conlang list as the "Gnoli triangle."

On the original version of this page in March 2006 I displayed a similar triangle with the vertices labeled: artlang, auxlang, englang; and I called it the Conlang triangle. The name never caught on and, with the new labeling, it was still generally referred to as the Gnoli triangle. I did see it once or twice referred to as the Gnoli-Brown triangle; such an appellation is not correct as it was not I but And Rosta who first proposed changing the label of Claudio's 'logging' vertex to 'engelang'. In an email to the Conlang list of the 12th May 2002, after I had proposed a quadrilateral modification of the Gnoli triangle, And wrote:

I think the Gnoli triangle would be better if 'loglang' were replaced by 'engelang' -- at the time Claudio defined the triangle, the distinction was not apparent, and the term 'engelang' had not been created.

On the following day, I replied:

Yes, I like that. It keeps things simple. Tim May has rightly pointed out the pitfalls of a quadrilateral model. Your redefining the Gnoli triangle is neater. I suppose we must now call it the Gnoli-Rosta triangle :)

It was not renamed the Gnoli-Rosta triangle and should certainly not be called the Gnoli-Brown triangle. It has remained quite simply the Gnoli triangle even after the slight modification proposed by And Rosta.

I am proposing a further modification below; I am adding color! In the former model there was apparently some ambiguity as to what it meant if a conlang were located at the center. One contributor to the Conlang list assumed that as that point was the furthest from the three vertices, a conlang in the center was neither an artlang, nor an auxlang, nor an engelang which, within the terms of the model, must surely mean 'not a conlang.' I had always assumed that a language in the center was a third artlang, a third auxlang and a third engelang. Therefore I have adopted the Maxwell color triangle model to show this:


Thus where the blue and green merges into cyan we have those conlangs that have both significant artlang and engelang features, similarly where green and red merge into yellow we have languages with significant artlang and auxlang features and where red and blue merge into magenta we have languages with significant engelang and auxlang features. As we move towards the center, where the colors merge to give white, we will find conlangs with significant artlang, engelang and auxlang features.

You will notice gridlines on the triangle, but I regret the figures are too obscure to read. The vertical lines show the amount of green (i.e. amount of artlangness); the green vertex is numbered 1.0 and the vertical lines are numbered 0.9, 0.8. 07. etc. down to the y edge which is 0 (no artlangness). Similarly, the top-right to bottom-left diagonals are numbered 0.9, 0.8. 07. etc. over to the x edge which is 0 and show the amount auxlangness (red) and the top-left to bottom-right diagonals are numbered 0.9, 0.8. 07. etc. over to the z edge which is 0 and show the amount engelangness (blue). Thus a language which is claimed to be "artlang 0, auxlang 0, engelang 0" cannot be mapped on the triangle - and that, in my opinion, is correct. In the center would be that conlang which is so finely balanced that it is: artlang 0.33, auxlang 0.33, englang 0.33.

As for the choice of colors;

  • Red had to be the color for auxlangs, if only by default (though some may see it also as symbolizing the passion displayed by some authors and many adherents of individual auxlangs). If I had chosen green, I would assuredly have been thought by some to implicitly favor Esperanto; and blue was ruled out not only by Léon Bollack's "La Langue Bleue" but also by its association with Ido (and I certainly do not want to be involved in the Esperanto-Ido controversy!). Red it had to be.
  • Green is suggested for artlangs as representing the artistic creativity of their authors. In this connexion, Jim Henry (author of gzb) has drawn my attention to Latin viriditās (gen. viriditātis) "greenness", especially as used by Hildegard von Bingen; indeed, at least one web-site claims she coined the word. She did not; the word goes back to classical antiquity. We find viriditās used by Cicero both with its literal meaning "greenness, verdure" and with the derived meanings "freshness, briskness, vigor." The word is found in Augustine, Gregory the Great and other Christian writers. But Hildegard used the word constantly; for her it denoted the "blossoming creativity" found not only in the botanical world but also in the human intellect [Sandra Ballif Straubhaar, 2008, "Sarah Higley, Hildegard of Bingen's Unknown Language: An Edition, Translation and Discussion. (The New Middle Ages series.) Palgrave Macmillan, 2007", Medieval Femininst Forum 44, n. 2 (pages 158-161)].
  • Blue seemed to me appropriate for engelangs as representing the purposeful engineering of their authors. As Jörg Rhiemeier (author of Old Albic) observed in an email to the Conlang list (28th July 2011): "The 'cool' blue colour fits the 'cool' 'objective' atmosphere of engelangs."

As for naming the triangle, Gnoli-Brown is hardly appropriate for such a colorful triangle! Nor, of course, would it be accurate. If one is to name all conlangers who contributed to it then it should be the Gnoli-Rosta-Brown triangle. But that, in my opinion, is clumsy and unnecessary. The most important element is the conception of the triangle for mapping conlangs and that is due to Claudio Gnoli, not to And nor to me. So I suggest we retain the now well established name 'Gnoli Triangle'; and if it is necessary to distinguish the version I've shown above, the 'Gnoli Maxwell Triangle' is surely the most appropriate name.

Valid XHTML 1.0 Transitional Valid CSS!
Created March 2006. Last revision:
Copyright © Ray Brown