Alan Reed Libert
University of Newcastle, New South Wales
Mainstream theoretical linguists have generally ignored artificial languages, apparently considering them unworthy of attention. This is true not only of “fictional languages” such as Klingon, but also of “serious” languages such as Esperanto. Much of the linguistic work which has been done on artificial languages has been carried out by Esperantists and is not concerned with structural aspects of these languages (see e.g. Schubert, ed. 1989). Serious languages have usually been designed to facilitate international communication and so are generally intended to be optimally designed for easy learning and use. Chomskyan linguists assume that there is a human language faculty, which sets limits on a possible human language. If this is the case, one might think that there would be limits even on languages which have been consciously created — unless one were trying to be perverse, which one presumably would not do in creating a language for international communication — or at least there would be limits on a usable artificial language.
The strangest, and perhaps the most interesting, artificial languages are the a priori languages, that is, those (supposedly) designed from scratch without reference to existing languages. These will seem exotic for various reasons, one being the total unfamiliarity of most of the vocabulary to speakers of any natural language. This is in contrast to another main type of artificial languages, a posteriori languages, which are based on one or more natural languages. Still other languages are called mixed languages, as they contain substantial material of both the a priori and a posteriori types (this way of classifying artificial languages has long been used, e.g. by Couturat and Leau 1903). Esperanto, the best known and most successful artificial language, is of the a posteriori type, and much of Esperanto will be familiar to someone who knows several major European languages (for a recent grammar of Esperanto in English, see Gledhill 2000). Volapük, a mixed language, was the most successful artificial language before Esperanto (see e.g. Post 1890 for more information on it).
By looking at artificial languages, in particular at the a priori ones, we may be able to explore the limits of language universals: perhaps universals will constrain even language creation. In fact, there may not be any truly a priori artificial languages, since we cannot create — or at least cannot fluently use — a language which lacks properties that natural languages must have.
In my book on a priori languages (Libert 2000), after descriptions of some properties of these languages, I compared these with universals that had been posited in the literature (see e.g. ibid. :25-27, 38, 52, 60), mainly using the Universals Archive of Konstanz University as a source. (This is an excellent resource, with more than 2000 putative universals gathered from work by various authors.) I did the same thing in my (2003) book on mixed artificial languages; since these languages have many a priori components one might expect to find some violations of universals in them as well (see e.g. ibid. :20-21, 23).
For example, I looked at some universals concerning phonemic inventories of languages. Universal 775 (778 at the time I was writing the book): “Every language must have at least one Primary Nasal Consonant in its inventory”. This turned out to be true of the a priori languages that I examined at the time, although there is no reason why someone could not design a language without nasals if he so chose. On the other hand, still about nasals, Universal 788 (formerly 791) states that in languages with a single nasal, that nasal will be /n/. This did not hold of all of the languages described in my book. The sole nasal of the a priori language Fitusa, for example, is /m/ (Rosenblum 1935:1).
Some artificial languages have quite complex systems of verbal morphology, and we can thus see whether they follow univerals involving this part of the grammar. The relevant universals that I brought up in Libert (2003:76) were implicational universals, e.g. Universal 433 (formerly 434): “If there is a certain distinction as to aspect or tense in the forms of a non-indicative mood, the same distinction is also represented in the given language in the forms of the indicative mood.” This held in all of the mixed languages which I examined, but the following one did not, Universal 481 (formerly 483): “If progressive is signaled by a suffix, any tense, mood, or person suffixes will follow the progressive suffix.” Qôsmianî violates this, as the progressive suffix (-l-) appears after the mood suffix (Beatty 1922:26).
However, one might argue that the most interesting properties from the point of view of universals would not involve the phonemic inventory or the vocabulary, as these are consciously created, but would be those which the language designer has not reflected on much or at all, or might be completely unaware of. Some aspects of syntax are perhaps like this, such as the possibilities anaphoric reference within discourse: generative grammarians have articulated several putative universal constraints in this area in Binding Theory. Many language designers do not give instructions on how to use the reflexive pronouns of their language, and would not have much of an idea about the principles of Binding Theory, especially since most of the a priori languages were created before the articulation of these principles. Nevertheless, if these principles are indeed part of universal grammar, we would not expect sentences of artificial languages to contain violations of Binding Theory, nor would we expect language designers to give rules for the use of pronouns which would violate these principles. Consider in this regard the following sentence from the a priori language Oz (a serious language despite its name):
hEv az ansAsiUt iftlEplezais kek anpAtpyaup ek iftEgtOg adpad astlaup.
“Because the wicked do not receive their just deserts immediately, they grow bold in transgression.”
The word of interest here is ek, which Elam translated as “they”. In fact, it is a reflexive pronoun, and should be translated as “themselves”. However, the only plausible antecedent of it is az ansAsiUt “the wicked”, which is in a subordinate clause and thus does not “c-command” ek. That is, the two words are not in the right structural relation. This sentence therefore does not follow Binding Theory, in particular Principle A: “An anaphor is bound in its local domain” (Chomsky 1986:166). In other words, a reflexive pronoun must refer back to something which is not too structurally distant from it. This explains, for example, why *John said that himself was going out is ungrammatical.
Either Elam made a mistake in his own language, which could well be the case — it is not hard to see how this could happen — or Oz allows violations of Binding Theory and thus does not conform to a (supposed) universal of human language. Unfortunately, the “corpus” of Oz sentences is not huge, and I don’t think it contains any similar examples, which might allow us to tell whether the above sentence involves a mistake.
I would argue that one can even look at pragmatic universals and artificial languages. This might seem counter-intuitive, since most artificial languages have seen little or no use, and there are those who deny that there can be a serious pragmatics of an artificial language, even of Esperanto. Thus Galdia (2009:334) speaks of languages “such as Esperanto which cannot be fully determined in terms of pragmatics simply because they are artificial” and continues, “An artificial pragmatics, in turn, would not make much sense.” However, in my (2013) paper I did look at pragmatic aspects of some artificial languages.
Consider, for example, illocutionary force. Brown and Levinson (1987:142) say, “[I]ndirect speech acts are universal and for the most part are probably constructed in essentially similar ways in all languages”. There is an artificial language, Konya, whose designer seems at least to frown on the use indirect speech acts (Sulky 2005). In another artificial language, Seaspeak (an English-based controlled language designed for nautical purposes), there are required markers for different kinds of speech acts, which might rule out the possibility of indirect speech acts (see Weeks et al. 1988:96). Seaspeak would only be used in a limited set of circumstances, and so the mandatory use of speech act markers might hold up, but it would be interesting to see what would happen with Konya under conditions of frequent use; one might suspect that indirect speech acts would occur, whether or not the designer approved of them.
These are just some of the potential violations of putative universals I have found during my research into artificial languages. Identifying these may say something about the language faculty and the limits of what a potential human language can be.
Beatty, W. M. L. (1922) Qôsmianî. The Fraternity Press, Washington, DC.
Brown, P. and S. C. Levinson (1987) Politeness. Cambridge University Press: Cambridge.
Chomsky, N. (1986) Knowledge of Language. Praeger, New York.
Couturat, L. and L. Leau (1903) Histoire de la langue universelle. Librairie Hachette, Paris.
Elam, C. M. (1932) The Case for an A Priori Language. The Open Sesame Press, Cincinnati, OH.
Galdia, M. (2009) Legal Linguistics. Peter Lang, Frankfurt.
Gledhill, C. (2000) The Grammar of Esperanto: A Corpus-Based Description (2nd edition). Lincom Europa, Munich.
Libert, A. R. (2000) A Priori Artificial Languages. Lincom Europa, Munich.
——–. (2003) Mixed Artificial Languages. Lincom Europa, Munich.
——–. (2013) “What can Pragmaticists Learn from Studying Artificial Languages?” in A. Capone, F. Lo Piparo, and M. Carapezza, eds., Perspectives on Linguistic Pragmatics. Springer, Berlin.
Post, A. A. (1890) Comprehensive Volapük Grammar. Self-published, Mattapan, MA.
Rosenblum, B. (1935) Fitusa: Die Geburt einer Sprache. No publisher given, Basel.
Schubert, K., ed. (1989) Interlinguistics. Mouton de Gruyter, Berlin.
Sulky, L. (2005) “Konya Language – Dialogues”. At URL http://larrysulky.webs.com/konya/konya-dialogue.html (Accessed 18 July 2010).
Weeks, F., A. Glover, E. Johnson, and P. Strevens (1988) Seaspeak Training Manual. Pergammon Press, Oxford.
How to cite this post:
Libert, Alan Reed. 2013. ‘Theoretical linguistics and artificial languages’. History and Philosophy of the Language Sciences. https://hiphilangsci.net/2013/05/29/theoretical-linguistics-and-artificial-languages
Thanks for that interesting post.
I believe that Esperanto has been used in communities where kids have learned the language in the natural acquisition way. And I’ve always been curious whether there are changes that have been made by second generation (or more) speakers that might be “correcting” design flaws in the original. Esperanto is evidently usable as designed, but if there any universals that it’s violating, those should create conflicts for the natural learners (as opposed to adult learners), and one might find evidence of that. That’s one reason I’ve always wished there were more real linguistic work on Esperanto. Do you know of any studies on first language acquisition of Esperanto?
Much linguistic work on Esperanto, including my own (http://sunflower.com/~miner/) is in Esperanto. The same holds for work on children who learn the language at home as a first language. But check out Jouko Lindstedt “Esperanto as First Language” at
http://www.linguistics.fi/julkaisut/SKY2006_1/1FK60.1.5.LINDSTEDT.pdf “Esperanto as First Language”
[…] This is true not only of “fictional languages” such as Klingon, but also of “serious” languages such as Esperanto. Much of the linguistic work which has been done on artificial languages has been carried out by Esperantists … […]
I wonder why you didn’t look at newer artlangs. I know that nowadays people deliberately try to violate universals.
[…] Another area developed only in recent times is pragmatics, including factivity: it is not clear in Esperanto which verbs and head nouns are factive in cases where their meanings do not make it obvious. Factivity is not clear in, for example, /indigni ke/ ‘resent (it) that’, /lamenti ke/ ‘lament that’, /plendi ke/ ‘complain that’, /priplori/ ‘bemoan’, /grumbli ke/ ‘grumble that’, /koleri ke/ ‘be angry that’, /montri ke/ ‘show that’. The entire area of pragmatics in relation to Esperanto is in need of attention. Presumably speakers simply apply the pragmatics of their native languages to Esperanto. (Alan Reed Libert has already, in this blog, broached the subject.) […]