For some time it has been Dr. Boli’s ambition to build a taxonomy of language cranks. Probably the most difficult part of the enterprise will be in finding appropriate names for the various species of crank; but in the case we examine today, there is no such difficulty. Today’s species is the crank who believes he has found the universal key to all languages, whom we shall call the Panglossian,—referring to all languages, of course, but also reminding us of the incurable optimism of Dr. Pangloss.

Henry Matthews believed he had cracked the linguistic nut. In a little prospectus published in 1821, or possibly 1815, or perhaps some other year (but from the type clearly early nineteenth century), he proposed to create a dictionary that would make it possible for everyone to communicate with everyone else. The title of the prospectus displays his ambition and his optimism clearly enough: “A Plan for Translating Languages, Without Study, or Any Previous Acquaintance Therewith.”

We must give Mr. Matthews credit for articulating the difficulty that keeps an artificial “universal” language from ever succeeding.

An opinion has been entertained by some learned men, in different ages, that the knowledge of overcoming the difficulties of languages would be one day accomplished; and others have thought it possible to contrive, or create, a general or universal language. Several ingenious plans have been suggested for the support of foreign correspondence, by means of a general or universal character; but, in all these, there is much to be acquired and remembered, as well as a thorough acquaintance with the principles of grammar. Except a plan can be devised to be comprehensible by the person who can merely read, as well as the scholar, the sale of such a work would not be sufficiently extensive to justify the expence.

Here he seems to hit the nail on the head. The average intelligent reader will never summon sufficient motivation to learn a constructed language. A few fanatics will do it; but the most successful of all these supposed universal languages is Esperanto, and you may measure its success by going to any large city in the world and attempting, without making prior arrangements with the local Esperanto boosters, to accomplish anything at all in that language.

What, then, is to be done?

A system fully suitable has been conceived, which, by a simple arrangement, will obviate every difficulty.

The dictionary now proposed will enable persons to correspond with foreigners, of whose language they have no knowledge; and to translate freely, every species of their literature.…

By this dictionary, it will be only pastime for children to translate English works into the languages of India and China, and their works into our own. In fact, it will unlock the knowledge of the world, and communicate it to all, the most uninformed as well as the profoundly learned.

You want a copy of this dictionary already! How can you obtain it? Well, it has not been compiled quite yet. This is, after all, a prospectus, designed to raise enough interest to persuade a publisher that it will be worth paying Mr. Matthews to compile the work. But he lays out the system clearly enough. His plan is to create a dictionary of meanings rather than words (anticipating by some decades the work of Peter Mark Roget). The meanings of words must be carefully distinguished, and each listed separately if a word has more than one meaning.

For instance, in the common dictionary the words and significations would sometimes stand thus:

Tiller, s. a ploughman; handle of a rudder.

But, in this dictionary, it must be repeated thus:

  1. Tiller, s. a ploughman.
  2. Tiller, s. a handle of a rudder.

Having made that arrangement of all the meanings in English, he will assign a number to each meaning. This number is the key to everything. Now we may compile a French dictionary, or a Nahuatl dictionary, or an Igbo dictionary, with the same meanings assigned to the same numbers.

Our author recognizes one difficulty that comes up right away: since the English language is the key, only in English will the numerical order and the alphabetical order of words be anything like the same. “The English interpreting dictionary will be both alphabetical and numerical, in one and the same book. This advantage can only fall to the language that first adopts the plan.” In every other language, the dictionary (like Roget’s Thesaurus) must have two sections: one in numerical order, and one in alphabetical order.

Now see how wonderfully this eliminates the work of translation. If you are writing to someone in China who has the Chinese equivalent of this dictionary, then you have only to write your letter as usual, and then find the number for the meaning of each word and write those numbers down. Your Chinese correspondent receives your numerical letter, looks up the numbers in his own dictionary, and transcribes the matching words, and your letter sits before him in pure Chinese! If, however, you suspect that your Chinese correspondents are not supplied with dictionaries on the Matthews system, then you can obtain a Matthews-system Chinese dictionary yourself, and you can write your letters in Chinese just as if you had been born in Peiping.

At this point you may be thinking to yourself that Mr. Matthews is a little bit naive about the extent of the differences between languages. You are correct. He seems to know a little French, though perhaps not much: he gives a translation of the first Psalm from the French version as an example of how well his system works. Even here it is necessary to make some allowances for the differences in grammar between the languages, but he thinks that anyone of ordinary intelligence can make those allowances. “The very few superfluous words, such as the double negative in the French, &c. can, at first sight, be discovered by a common capacity, and struck out, or supplied, as may be required.”

It seems to us, however, that the system utterly wrecks against the rock of Latin, or German, or (heaven forfend) Russian, or any other declined language, where the form of the word itself conveys much of the meaning. Let us apply the system to Virgil, for example:

Arma virumque—

But here we have already run up against a nearly insuperable difficulty. That little parasitic particle -que, meaning “and,” attaches itself to the end of the second word of a pair, forming a single written word, but with two meanings. We do not need to be experts in Latin, but it will be necessary to have at least so much “previous acquaintance therewith” as will allow us to recognize that “virumque” resolves into the two completely unrelated meanings “man” and “and,” and that it will be necessary to place “and” before “man” to translate the word into English. We might overcome the difficulty by listing the combined form of every Latin word that can possibly take que on the end, effectively doubling the size of our Latin dictionary. But wait: we must then effectively double the size of every dictionary in every language, because each of those Latin meanings must be assigned a number.

And we have not begun to address the problem of inflection, which we can practically ignore in English and French, but which entirely determines who does what to whom in Latin. Let us suppose that we have managed to discover, somehow, that the enclitic -que has a separate meaning, and that we have managed to deduce the dictionary entries from the inflected forms of the words, and go on to translate the first few lines of the Aeneid on the Matthews system, remembering that inflection is not accounted for:

Arm man and sing Troy which first from coast
Italy fate fugitive Lavinia and come
shore much that and land thrown and high
force above fierce memory Juno because anger
many also war step while found city
carry god Latium tribe whence Latin
Alban and father and altar wall Rome.

It would take all the imagination of Barry Fell to make a “smooth translation” from this jumble, and then it would bear little resemblance to what Virgil actually meant.

Now, this is an Indo-European language. In the great scheme of world languages, Latin and English are practically dialects of the same tongue. We cannot imagine that a translation from Cherokee would go as smoothly.

So it seems that Mr. Matthews never produced even his English dictionary, let alone the matching dictionaries for other languages that would break down the walls of incomprehension and open up any language to any intelligent reader “without study, or any previous acquaintance therewith.” But for a moment let us admire the ingenuity and optimism of a man who was ready to undo the eleventh chapter of Genesis. He was wrong, but he was gloriously wrong.


  1. The Shadow says:

    Wow. I have seen some cranks in my day, but this guy is one of the crankiest.


    1. an eccentric person, especially one who is obsessed by a particular subject or theory.

  2. Big Brother says:

    I wonder if this is basically how Google Translate and such work under the hood. That is, if they assign a unique (incredibly long, but that’s not a problem for 64-bit computers) number to every word or phrase they can find already translated from one language into others, then match with words or phrases in other language that have already been translated one way or the other between the same two languages. And thus they can go from one language to another with a phrase never before translated between those particular two languages, so long as it’s gone through at least one other language-in-common for both languages in question, either directly or even through still other languages.

    So, say you want to go from Klingon to Tlingit, but no one has previously translated the phrase in question that direction. But someone HAS translated it from Klingon to French, and someone else has translated that French phrase to English, and yet someone else has translated that English phrase to Tlingit. Google Translate has assigned that phrase’s meaning a unique phrase and can thus recognize that they’re all the same meaning.

    At least, once Google Translate updates to include Tlingit.

  3. RepubAnon says:

    Is this the system used to create English translations of installation instructions written in the manufacturer’s native tongue?

Leave a Reply

Your email address will not be published. Required fields are marked *