LOGICAL LANGUAGES:

A Path to Posthuman Rationality?

by Simon! D. Levy

I. Introduction

Most of us have heard of artificial languages with names like Esperanto and Interlang¹, that purport to avoid various shortcomings of ordinary natural languages through intentional design. Arguments in favor of these languages have ranged from the supposed necessity for a culturally neutral communication medium, to the desirability of a language whose grammar has no pesky exceptions to memorize.

Despite the collectivist mindset behind many designed languages, two such languages have emerged that seem worthy of Extropian attention. Though these languages, Lojban and E-prime, represent polar extremes in the degree to which they differ from ordinary languages, both seek to increase the rationality of their users by eliminating various types of ambiguity and generalization available in a language like English.

II. Lojban

During the 1930’s and 40’s a fire prevention engineer turned linguistic anthropologist by the name of Benjamin Lee Whorf conceived a curious and somewhat revolutionary idea about human language and its relation to thought. In Whorf’s own words,

… the forms of a person’s thoughts are controlled by inexorable laws of pattern of which he is unconscious. These patterns are the unperceived intricate systematizations of his own language. [1956, p. 252]

Named the Sapir-Whorf Hypothesis in acknowledgment of the contribution of Whorf’s teacher Edward Sapir, this idea has come to exert a profound influence on philosophy and anthropology during the present century.

Nevertheless, because of its essentially speculative nature, the Sapir-Whorf hypothesis remained largely an attitude toward language, rather than the basis for a scientific research program, until 1960. In that year, Dr. James Cooke Brown wrote an article for Scientific American describing Lojban, a language that he had developed to test the hypothesis put forth by Whorf. Brown hoped to

create a “logical language,” free of exceptions and irregular forms, based on formal properties of mathematics—properties which he believed gave human beings the ability to reason. If Whorf’s idea had any validity, reasoned Brown, then speakers of Lojban would end up with more disciplined and powerful minds than speakers of natural languages.

In attempt to maximize the “target population” of potential Lojban speakers, Brown picked the eight languages with the largest number of speakers (English, Mandarin Chinese, Hindi, Russian, Spanish, Japanese, French, and German) as a basis for the sound structure of Lojban. For each Lojban word, he found the closest corresponding word in each of the eight languages and allowed each language to contribute to some sub-part of its word to the Lojban word. He then computed a “learnability score” for the new word by multiplying the contribution of a given language by its representation in the language set, and summing over the contributions from each language. For example, the Lojban word blanu, meaning “blue,” contains all of the sounds (/b/, /l/, and /u/) from the English word, and one half the sounds from the Hindi word for blue, nila. Therefore, English contributes all of its percentage points, 28%, to the learnability of blanu, and Hindi contributes half of its points (1/2 of 11% = 5.5%) to the word. (With the other six languages’ contributions added in, it turns out that blanu has a learnability score of 76%).

This word-making scheme points out a major design feature of Lojban, viz., its phonemic spelling system[see sidebar], which facilitates its usage in written communication. Unlike most phonemically spelled languages, however, Lojban has complete resolvability between its spoken and written representations: Given a string of Lojban sounds, one can always determine which sounds go together in a word. Contrast this situation with English, in which a given sequence of letters can represent more than one phrase, for example, ANICEHOUSE.

This resolvability rests on two further

design features of the language. First, Lojban builds its words around a regular consonant-vowel “skeletal structure,” which differs for each word class. For example, predicate words, which describe observable real-world things like objects, actions, and qualities, always contain a consonant cluster (like the /bl/ in blanu) and end in a vowel. Simple operators, such as the pronoun da “he,” consist of a consonant followed by a vowel.

Second, Lojban grammar lacks syntactic ambiguity, which means that the role of each word in a Lojban sentence has only one possible interpretation. Again, natural languages do not always behave this way, leading to ambiguities like “Flying planes can be dangerous,” in which you don’t know whether to worry about the flying planes or the act of flying them. More specifically, Lojban’s designer modeled its syntax on predicate calculus, a system invented by logicians to represent propositions about the world in a simple and unambiguous way. Each Lojban sentence contains a predicate (roughly, the verb), which describes an action or condition, and a set of arguments, which describe, for example, who

15

EXTROPY #12 (6:1) First quarter 1994

performed the action (the subject) or suffered its consequences (the object). Like predicate calculus, Loglan makes verbs out of things that speakers of English would not normally consider verbs, such as ‘blue’.$^{2}$ Unlike predicate calculus, in which the predicate comes first, Loglan uses the word order subject-verb-object, because of the rarity of verb-first order in natural languages, and because the syntax ends up simpler if the predicate does not come first.

Though it has received scant attention in mainstream academic linguistics, Loglan maintained a core of hardcore followers and now has an incorporated organization – the Logical Language Group, headed by Bob Le Chevalier (a.k.a. lojbab) – to support and promote the language, now called Lojban.$^{3}$ A variety of on- and off-line documentation, in addition to learning tapes, software, classes, and discussion groups, exist. This brings us to the question of whether Extropians should spend any time learning Lojban.

Despite its possible merits as a means of disciplining our thought, Lojban differs so strongly from English (and any other natural language) that learning it would require a significant investment of time and effort. We might better spend such time and effort spreading Extropians ideas though more traditional channels of communication, like English, which has more speakers than any other language in the world. This problem leads us to wonder whether we can’t fix what’s wrong with English, without giving the language up entirely.

III. E-prime

Proponents of another artificial language called E-prime have argued that we can still speak English, but eliminate a good deal of what’s wrong with that language, if we follow the advice of Dr. Alfred Korzybski, one of the founders of a field known as general semantics. Unlike the traditional academic discipline of semantics, which tends to focus on abstract issues in formal linguistic theory, general semantics concerns itself with the practical implications and consequences of language use and abuse in everyday realms such as teaching, advertising, and news reporting. Dr. Korzybski’s student David Bourland contributed to the field by proposing a derived language, ‘English minus ‘be”, or ‘E-prime’, to remedy the fact that ‘be’ in all its forms (‘am’, ‘are’, ‘is’, ‘was’, ‘were’, ‘being’, ‘been”) allows for tremendous ambiguity.

Consider, for example, the sentence ‘John is a liar’. What exactly does someone mean when they say this? Should we believe that John always lies whenever he says something? Obviously not, or John would have a good deal of trouble getting by in the world. Does the speaker claim that John lies more often than

Phonemic Spelling

Phonemic spelling, an essential feature of Lojban, can be described to a first approximation as a system of one-to-one correspondence between linguistic sounds and the letters that are used to write them. (We do not wish to call such a system ‘phonetic spelling,’ since a phonetic transcription includes details that are not relevant to distinguishing one word from another.) Phonemic spelling, found in languages like Spanish and Korean, contrasts with the confusing muddle that is English orthography. For example, the first sound in the English word ‘she’ can also be spelled with the letters ‘ti,’ as in ‘action,’ the letters ‘si’ as in ‘fusion,’ the letters ‘sci’ as in ‘conscious,’ or the letters ‘su’ as in ‘erasure.’ This proliferation of spellings, caused mostly by the fact that English has borrowed heavily from several different languages during its history, led George Bernard Shaw to observe that the word ‘fish’ could equally well be spelled ‘ghoti,’ with the ‘gh’ derived from words like ‘tough’ and ‘rough.’ Of course, Shaw conveniently ignored the fact that ‘gh’ can only sound like ‘f’ at the end of a word, and ‘ti’ like ‘sh’ in the middle, but his point was well taken.

The most serious linguistic objection to a purely phonemic system is that English is to some degree morphologically spelled; that is, the same letter or group of letters can be used to spell two different sounds that encode the same meaning. For example, the ‘s’ at the end of ‘cats’ and the ‘s’ at the end of ‘dogs’ are phonemically different (one corresponding the /s/ in ‘sue’ and the other to the /z/ in ‘zoo’), but both serve to signify plurality in nouns and singularity in regular verbs. Still, such examples are vastly outnumbered by the situations in which phonemic spelling would represent an improvement.

From a more practical standpoint, it would be tremendously expensive to rebuild keyboards and other such devices to reflect a new spelling system; however, there are straightforward phonemic alphabets (e.g. ARPABET) that use the standard character set of English, so this objection is not all that serious. There is also the problem of what to do with billions of volumes of books and other works spelled in the traditional way. With the increased availability of optical character readers and the proliferation of personal computers, it is not unreasonable to assume that such literature could eventually be converted on-line into ARPABET or a similar system. No matter what language we end up using, it makes sense to push for a phonemic spelling system.

not – say, 60 percent of the time? Probably not, as the speaker would have a great deal of difficulty persuading us that they had recorded everything that John ever said and divided the number of John’s false statements by the total number of statements John made. Instead, when someone says ‘John is a liar’ they probably mean something closer to ‘I have heard John lie enough times that I don’t trust him to tell the truth about matters of importance to me.’ This sentence, though much longer than ‘John is a liar,’ has the advantage of forcing the speaker to clarify their feelings about John, without making a claim that the listener would find difficult or impossible to falsify.

More generally, eliminating ‘be’ helps speakers avoid the trap of attributing a quality or a behavior to a noun (person, place or thing) without specifying the conditions under which they observed the quality or behavior. Specificity costs more words but buys the advantage of falsifiability, which the philosopher of science Karl Popper considered the criterion for judging scientific hypotheses. For this reason, we have reason to expect that speaking E-prime instead of English will make our dis-

discourse more scientific, and hence more rational.

Dropping ‘be’ also prevents the employment of the passive voice. The passive voice (as in ‘He’s been killed.’) allows speakers to avoid the responsibility of naming the person or group who performed an action. Think of all the times you’ve heard ‘I was led to believe that…’ as an excuse, and you’ll see the benefits of such a constraint.

Some critics of ‘pure’ E-prime have pointed out that ‘be’ can also function more harmlessly as an auxiliary verb to indicate the aspect (discrete or continuous) of another verb, as in ‘I was running around all day yesterday.’ One such critic, William Dallmann (1992) has argued for another form of the language, which he calls E-Prime$_{mod}$. According to Dallmann,

The mod version eliminates the is of identification (He is a general semanticist), the is of predication (She is beautiful), but retains is as an auxiliary (She is dancing), the is of existence (To be or not to be), and the denial of identity (The map is not the territory). [1992, p. 134]

EXTROPY #12 (6:1) First quarter 1994

16

1984

A more troublesome objection involves the possibility of getting around the restrictions of “be”-lessness by the common device of presupposition. So, for example, instead of saying “John is an liar: He told me he’d be here at eight, and it’s now nine-thirty,” I can say “John, that liar, told me he’d be here at eight, …” The information conveyed in the former utterance differs little, if at all, from that conveyed in the latter; I have simply chosen to include the claim “John is a liar” as given (pre-supposed) in the latter.

Still, the inability to use any form of “be” can lead to tremendous improvements in our sense of what we actually say (or fail to say) when we speak and write. Furthermore, E-prime has the obvious advantage over languages like Lojban in ease of usage (for English speakers, anyway) and, therefore, in the amount of effort it would take to become a fluent speaker of the language. Though I did not use any form of the verb “be” in the body of this article (except in quotations), I did not find it much harder to compose than other pieces I have written.

IV. What should we do?

On the one hand, both Lojban and E-prime have a strong intuitive appeal on Extropian grounds. The languages represent a deliberate attempt to design a system of communication both logically consistent and – more importantly – potentially capable of improving the rational powers of their users. Since the transhumanist program involves the shedding of old, arbitrary customs and habits in favor of well-thought-out plans of action, combined with active self-enhancement, Lojban and E-prime look like the “languages of choice” for Extropian communication.

On the other hand, artificial languages – especially Lojban – have their origin in philosophical and empirical premises that appear to run counter to extropian values, and perhaps to plain facts. The Sapir-Whorf hypothesis represents an explicit concession to limits on our ability to think, a concession that stands in direct opposition to the principle of Boundless Expansion. Even if such limits exist, must we attribute them to the specific language that we speak? It seems entirely plausible that our ability to reason and our ability to speak come from our genetic endowment, part of which generates the “universal grammar” that forms the basis for all human languages. Under this view of language – promoted by the linguist Noam Chomsky – learning a given language involves the setting of parameters hard-wired into the human species. Of course, this particular form of biological determinism does not rule out the possibility that a given language may influence the way its speakers think, but it does not rule out the “null hypothesis” that your language does not shape your thinking.

In fact, it seems more intuitive to accept the opposite point of view; namely, that your thinking patterns – as determined, for example, by your physical environment – shape your language. For example, most Indo-European languages have two or three genders for nouns, which correlate to some extent with “male”, “female”, and “other”, but some languages spoken in Africa have many more genders, which distinguish among various classes of animate and inanimate things found there. The genders of these languages often embrace categories like “poisonous plant”, suggesting that grammatical systems derive at least in part from the way people think about their environment.

In these sorts of “chicken-or-egg” problems the truth may lie in some uninteresting compromise between the two extremes, or in a new approach. A very exciting possibility for the latter comes from dynamical systems theory, in which circular causality falls out naturally from the underlying assumptions, rather than posing the paradox it has for traditional scientific thought. In other words, our language shapes our thinking, which shapes our language, and so on. This view does not require a Lamarckian version of language evolution; we don’t want to claim that our language or thinking directly affects our genetic endowment. Rather, we acknowledge that given a particular genetic arrangement, we obtain a system in which language and thought can have mutual influence.

This possibility brings out a greater truth about artificial languages. Natural languages serve as an example par excellence of spontaneous orders, so it could very well end up that any artificial language will mutate into a natural language – warts and all – given enough speakers and enough time. Despite the logical unattractiveness of exceptions and ambiguities, they seem to have tremendous appeal to people at some deep level. We might, for example, find future speakers of E-prime committing the sin of the “be of identification” without using that verb or presupposition, as speakers of Indonesian, Hebrew, and many other languages do now: “John a liar.”

This possibility does not bode well for the future of the species, when we consider the thinking of B.L. Whorf himself:

We cannot but suppose that the future developments of thinking are of primary importance to the human species. They may even determine the duration of human existence on the planet earth or in the universe. [1956, p.83]

At this point, many readers – especially those who have followed the Extropian movement for any length of time – will recognize a different possibility. If, as Chomsky argues, the physical makeup of our brains determines the possible form of our language, it seems obvious that we should seek ways to modify our brains so as to lessen the chances for

destructive irrationality. Such efforts, though clearly beyond the scope of this article, have received a good deal of coverage in this journal, and Extropy readers can expect to see much writing on this topic in future issues.

Acknowledgments

The author wishes to thank Bob Le Chevalier (lojbab) and the Logical Language Group for making available a large amount of Lojban documentation, and Exl Director Russell Whitaker for contact information and enthusiasm about Lojban. You can reach the Logical Language group at 2904 Beau Lane, Fairfax, VA 22031, or by calling (703) 385-0273. Anyone interested E-prime and general semantics can subscribe to Etc., the journal of the International Society for General Semantics, P.O. Box 2469, San Francisco, CA 94126.

References

Bourland, D.B. Jr. and Paul D. Johnston, Eds. (1991) To Be or Not: An E-Prime Anthology. San Francisco: International Society for General Semantics. Brown, J.C. (1960) Loglan. Scientific American 202, 6, 53-63. Chomsky, N. (1965) Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Dallmann, W. (1992) “Is Is Not Is Is Not Is” Etc.: A Review of General Semantics 49, 2. San Francisco: International Society for General Semantics. Lakoff, G. (1990) Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: University of Chicago Press. Popper, K. (1968) Conjectures & Refutations: The Growth of Scientific Knowledge. New York: Harper & Row. Whorf, B.L. (1956). Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf. Cambridge, MA: MIT Press.

Notes

1 and my own personal favorite, Volapük 2 This arrangement would not appear so strange to speakers of Japanese, a language in which adjectives can behave like verbs. 3 According to lojbab, Lojban arose as dialect of Loglan, because Dr. Brown maintains a copyright on the original language. 4 American-educated readers may notice that I put the comma outside the quotation marks in this list, and in the rest of the article, breaking with traditional prescribed style in this country. Inspired by my study of Lojban and E-prime, I have decided to take a more pro-active attitude toward my writing habits, and therefore use the more sensible practice favored in Great Britain. 5 I have also decided, after many years of personal linguistic conservatism, to adopt the common usage of the pronoun “they” to mean “he or she” (and hence “their” for “his or hers” and “them” for “him or her.”) I thank Extropian Rob Michels for pointing out the sensibility of this practice to me. Those who object to it must also have difficulty with French, German, and other languages where an originally plural pronoun became both plural and singular.

17

EXTROPY #12 (6:1) First quarter 1994