Chapter 17. The Relojban letteral system

James Cooke Brown, the founder of the Loglan Project, coined the word “letteral” (by analogy with “numeral”) to mean a letter of the alphabet, such as “f” or “z”. A typical example of its use might be

(Don't forget the one within quotation marks.) Using the word “letteral” avoids confusion with “letter”, the kind you write to someone. Not surprisingly, there is a Relojban gismu for “letteral”, namely lerfu, and this word will be used in the rest of this chapter.

Relojban uses the Latin alphabet, just as English does, right? Then why is there a need for a chapter like this? After all, everyone who can read it already knows the alphabet. The answer is twofold:

First, in English there are a set of words that correspond to and represent the English lerfu. These words are rarely written down in English and have no standard spellings, but if you pronounce the English alphabet to yourself you will hear them: ay, bee, cee, dee ... . They are used in spelling out words and in pronouncing most acronyms. The Relojban equivalents of these words are standardized and must be documented somehow.

Second, English has names only for the lerfu used in writing English. (There are also English names for Greek and Hebrew lerfu: English-speakers usually refer to the Greek lerfu conventionally spelled “phi” as “fye”, whereas “fee” would more nearly represent the name used by Greek-speakers. Still, not all English-speakers know these English names.) Relojban, in order to be culturally neutral, needs a more comprehensive system that can handle, at least potentially, all of the world's alphabets and other writing systems.

Letterals have several uses in Relojban: in forming acronyms and abbreviations, as mathematical symbols, and as sumka'i – the equivalent of English pronouns.

In earlier writings about Relojban, there has been a tendency to use the word lerfu for both the letterals themselves and for the Relojban words which represent them. In this chapter, that tendency will be ruthlessly suppressed, and the term “lerfu word” will invariably be used for the latter. The Relojban equivalent would be lerfu valsi or lervla.

The first requirement of a system of lerfu words for any language is that they must represent the lerfu used to write the language. The lerfu words for English are a motley crew: the relationship between “doubleyou” and “w” is strictly historical in nature; “aitch” represents “h” but has no clear relationship to it at all; and “z” has two distinct lerfu words, “zee” and “zed”, depending on the dialect of English in question.

All of Relojban's basic lerfu words are made by one of three rules:

Therefore, the following table represents the basic Relojban alphabet:

There are several things to note about this table. The consonant lerfu words are a single syllable, whereas the vowel and ' lerfu words are two syllables and must be preceded by pause (since they all begin with a vowel). Another fact, not evident from the table but important nonetheless, is that by and its like are single cmavo of selma'o BY, as is .y'y. The vowel lerfu words, on the other hand, are compound cmavo, made from a single vowel cmavo plus the cmavo bu (which belongs to its own selma'o, BU). All of the vowel cmavo have other meanings in Relojban (logical connectives, sentence separator, hesitation noise), but those meanings are irrelevant when bu follows.

Here are some illustrations of common Relojban words spelled out using the alphabet above:

Spelling out words is less useful in Relojban than in English, for two reasons: Relojban spelling is phonemic, so there can be no real dispute about how a word is spelled; and the Relojban lerfu words sound more alike than the English ones do, since they are made up systematically. The English words “fail” and “vale” sound similar, but just hearing the first lerfu word of either, namely “eff” or “vee”, is enough to discriminate easily between them – and even if the first lerfu word were somehow confused, neither “vail” nor “fale” is a word of ordinary English, so the rest of the spelling determines which word is meant. Still, the capability of spelling out words does exist in Relojban.

Note that the lerfu words ending in y were written (in Example 17.2 and Example 17.3) with pauses after them. It is not always necessary to pause after such lerfu words, but failure to do so can in some cases lead to incorrect parses:

without a pause after cy would be interpreted as:

A safe guideline is to pause after any cmavo ending in y unless the next word is also a cmavo ending in y. The safest and easiest guideline is to pause after all of them.

Relojban doesn't use lower-case (small) letters and upper-case (capital) letters in the same way that English does; sentences do not begin with an upper-case letter, nor do names. However, upper-case letters are used in Relojban to mark irregular stress within names, thus:

It would require far too many cmavo to assign one for each upper-case and one for each lower-case lerfu, so instead we have two special cmavo ga'e and to'a representing upper case and lower case respectively. They belong to the same selma'o as the basic lerfu words, namely BY, and they may be freely interspersed with them.

The effect of ga'e is to change the interpretation of all lerfu words following it to be the upper-case version of the lerfu. An occurrence of to'a causes the interpretation to revert to lower case. Thus, ga'e .abu means not “a” but “A”, and Ivan's name may be spelled out thus:

The cmavo and compound cmavo of this type will be called “shift words”.

How long does a shift word last? Theoretically, until the next shift word that contradicts it or until the end of text. In practice, it is common to presume that a shift word is only in effect until the next word other than a lerfu word is found.

It is often convenient to shift just a single letter to upper case. The cmavo tau, of selma'o LAU, is useful for the purpose. A LAU cmavo must always be immediately followed by a BY cmavo or its equivalent: the combination is grammatically equivalent to a single BY. (See Section 17.14 for details.)

A likely use of tau is in the internationally standardized symbols for the chemical elements. Each element is represented using either a single upper-case lerfu or one upper-case lerfu followed by one lower-case lerfu:

If a shift to upper-case is in effect when tau appears, it shifts the next lerfu word only to lower case, reversing its usual effect.

So far we have seen bu only as a suffix to vowel cmavo to produce vowel lerfu words. Originally, this was the only use of bu. In developing the lerfu word system, however, it proved to be useful to allow bu to be attached to any word whatsoever, in order to allow arbitrary extensions of the basic lerfu word set.

Formally, bu may be attached to any single Relojban word. Compound cmavo do not count as words for this purpose. The special cmavo zei, zo, si, sa, fa'o may not have bu attached, because they are interpreted before bu detection is done; in particular,

is needed when discussing bu in Relojban. It is also illegal to attach bu to itself, but more than one bu may be attached to a word; thus .abubu is legal, if ugly. (Its meaning is not defined, but it is presumably different from .abu.) It does not matter if the word is a cmavo, a cmevla, or a brivla. All such words suffixed by bu are treated grammatically as if they were cmavo belonging to selma'o BY. However, if the word is a cmevla it is always necessary to precede and follow it by a pause, because otherwise the cmevla may absorb preceding or following words.

The ability to attach bu to words has been used primarily to make names for various logograms and other unusual characters. For example, the Relojban name for the “happy face” is uibu, based on the attitudinal ui that means “happiness”. Likewise, the “smiley face”, written “:-)” and used on computer networks to indicate humor, is called zo'obu The existence of these names does not mean that you should insert uibu into running Relojban text to indicate that you are happy, or zo'obu when something is funny; instead, use the appropriate attitudinal directly.

Likewise, joibu represents the ampersand character, “&”, based on the cmavo joi meaning “mixed and”. Many more such lerfu words will probably be invented in future.

The . and , characters used in Relojbanic writing to represent pause and syllable break respectively have been assigned the lerfu words denpa bu (literally, “pause bu”) and slaka bu (literally, “syllable bu”). The written space is mandatory here, because denpa and slaka are normal gismu with normal stress: denpabu would be a fu'ivla (word borrowed from another language into Relojban) stressed denPAbu. No pause is required between denpa (or slaka) and bu, though.

As stated in Section 17.1, Relojban's goal of cultural neutrality demands a standard set of lerfu words for the lerfu of as many other writing systems as possible. When we meet these lerfu in written text (particularly, though not exclusively, mathematical text), we need a standard Relojbanic way to pronounce them.

There are certainly hundreds of alphabets and other writing systems in use around the world, and it is probably an unachievable goal to create a single system which can express all of them, but if perfection is not demanded, a usable system can be created from the raw material which Relojban provides.

One possibility would be to use the lerfu word associated with the language itself, Lojbanized and with bu added. Indeed, an isolated Greek “alpha” in running Relojban text is probably most easily handled by calling it .alfas. bu. Here the Greek lerfu word has been made into a Lojbanized name by adding s and then into a Relojban lerfu word by adding bu. Note that the pause after .alfas. is still needed.

Likewise, the easiest way to handle the Latin letters “h”, “q”, and “w” that are not used in Relojban is by a consonant lerfu word with bu attached. The following assignments have been made:

As an example, the English word “quack” would be spelled in Relojban thus:

Note that the fact that the letter “c” in this word has nothing to do with the sound of the Relojban letter c is irrelevant; we are spelling an English word and English rules control the choice of letters, but we are speaking Relojban and Relojban rules control the pronunciations of those letters.

A few more possibilities for Latin-alphabet letters used in languages other than English:

However, this system is not ideal for all purposes. For one thing, it is verbose. The native lerfu words are often quite long, and with bu added they become even longer: the worst-case Greek lerfu word would be .Omikron. bu, with four syllables and two mandatory pauses. In addition, alphabets that are used by many languages have separate sets of lerfu words for each language, and which set is Relojban to choose?

The alternative plan, therefore, is to use a shift word similar to those introduced in Section 17.3. After the appearance of such a shift word, the regular lerfu words are re-interpreted to represent the lerfu of the alphabet now in use. After a shift to the Greek alphabet, for example, the lerfu word ty would represent not Latin “t” but Greek “tau”. Why “tau”? Because it is, in some sense, the closest counterpart of “t” within the Greek lerfu system. In principle it would be all right to map ty. to “phi” or even “omega”, but such an arbitrary relationship would be extremely hard to remember.

Where no obvious closest counterpart exists, some more or less arbitrary choice must be made. Some alien lerfu may simply not have any shifted equivalent, forcing the speaker to fall back on a bu form. Since a bu form may mean different things in different alphabets, it is safest to employ a shift word even when bu forms are in use.

Shifts for several alphabets have been assigned cmavo of selma'o BY:

The cmavo zai (of selma'o LAU) is used to create shift words to still other alphabets. The BY word which must follow any LAU cmavo would typically be a name representing the alphabet with bu suffixed:

Unlike the cmavo above, these shift words have not been standardized and probably will not be until someone actually has a need for them. (Note the . characters marking leading and following pauses.)

In addition, there may be multiple visible representations within a single alphabet for a given letter: roman vs. italics, handwriting vs. print, Bodoni vs. Helvetica. These traditional “font and face” distinctions are also represented by shift words, indicated with the cmavo ce'a (of selma'o LAU) and a following BY word:

The cmavo na'a (of selma'o BY) is a universal shift-word cancel: it returns the interpretation of lerfu words to the default of lower-case Relojban with no specific font. It is more general than lo'a, which changes the alphabet only, potentially leaving font and case shifts in place.

Several sections at the end of this chapter contain tables of proposed lerfu word assignments for various languages.

Many languages that make use of the Latin alphabet add special marks to some of the lerfu they use. French, for example, uses three accent marks above vowels, called (in English) “acute”, “grave”, and “circumflex”. Likewise, German uses a mark called “umlaut”; a mark which looks the same is also used in French, but with a different name and meaning.

These marks may be considered lerfu, and each has a corresponding lerfu word in Relojban. So far, no problem. But the marks appear over lerfu, whereas the words must be spoken (or written) either before or after the lerfu word representing the basic lerfu. Typewriters (for mechanical reasons) and the computer programs that emulate them usually require their users to type the accent mark before the basic lerfu, whereas in speech the accent mark is often pronounced afterwards (for example, in German “a umlaut” is preferred to “umlaut a”).

Relojban cannot settle this question by fiat. Either it must be left up to default interpretation depending on the language in question, or the lerfu-word compounding cmavo tei (of selma'o TEI) and foi (of selma'o FOI) must be used. These cmavo are always used in pairs; any number of lerfu words may appear between them, and the whole is treated as a single compound lerfu word. The French word “été”, with acute accent marks on both “e” lerfu, could be spelled as:

and it does not matter whether .akut. bu appears before or after .ebu; the tei…foi grouping guarantees that the acute accent is associated with the correct lerfu. Of course, the level of precision represented by Example 17.18 would rarely be required: it might be needed by a Relojban-speaker when spelling out a French word for exact transcription by another Relojban-speaker who did not know French.

This system breaks down in languages which use more than one accent mark on a single lerfu; some other convention must be used for showing which accent marks are written where in that case. The obvious convention is to represent the mark nearest the basic lerfu by the lerfu word closest to the word representing the basic lerfu. Any remaining ambiguities must be resolved by further conventions not yet established.

Some languages, like Swedish and Finnish, consider certain accented lerfu to be completely distinct from their unaccented equivalents, but Relojban does not make a formal distinction, since the printed characters look the same whether they are reckoned as separate letters or not. In addition, some languages consider certain 2-letter combinations (like “ll” and “ch” in Spanish) to be letters; this may be represented by enclosing the combination in tei…foi.

In addition, when discussing a specific language, it is permissible to make up new lerfu words, as long as they are either explained locally or well understood from context: thus Spanish “ll” or Croatian “lj” could be called .ibu, but that usage would not necessarily be universally understood.

Section 17.19 contains a table of proposed lerfu words for some common accent marks.

Relojban does not have punctuation marks as such: the denpa bu and the slaka bu are really a part of the alphabet. Other languages, however, use punctuation marks extensively. As yet, Relojban does not have any words for these punctuation marks, but a mechanism exists for devising them: the cmavo lau of selma'o LAU. lau must always be followed by a BY word; the interpretation of the BY word is changed from a lerfu to a punctuation mark. Typically, this BY word would be a name or brivla with a bu suffix.

Why is lau necessary at all? Why not just use a bu-marked word and announce that it is always to be interpreted as a punctuation mark? Primarily to avoid ambiguity. The bu mechanism is extremely open-ended, and it is easy for Relojban users to make up bu words without bothering to explain what they mean. Using the lau cmavo flags at least the most important of such nonce lerfu words as having a special function: punctuation. (Exactly the same argument applies to the use of zai to signal an alphabet shift or ce'a to signal a font shift.)

Since different alphabets require different punctuation marks, the interpretation of a lau-marked lerfu word is affected by the current alphabet shift and the current font shift.

Chinese characters (“hànzi” in Chinese, “kanji” in Japanese) represent an entirely different approach to writing from alphabets or syllabaries. (A syllabary, such as Japanese hiragana or Amharic writing, has one lerfu for each syllable of the spoken language.) Very roughly, Chinese characters represent single elements of meaning; also very roughly, they represent single syllables of spoken Chinese. There is in principle no limit to the number of Chinese characters that can exist, and many thousands are in regular use.

It is hopeless for Relojban, with its limited lerfu and shift words, to create an alphabet which will match this diversity. However, there are various possible ways around the problem.

First, both Chinese and Japanese have standard Latin-alphabet representations, known as “pinyin” for Chinese and “romaji” for Japanese, and these can be used. Thus, the word “han⁴zi⁴” (“hànzi”) is conventionally written with two characters, but it may be spelled out as:

The cmavo vo is the Relojban digit “4”. It is grammatical to intersperse digits (of selma'o PA) into a string of lerfu words; as long as the first cmavo is a lerfu word, the whole will be interpreted as a string of lerfu words. In Chinese, the digits can be used to represent tones. Pinyin is more usually written using accent marks, the mechanism for which was explained in Section 17.6.

The Japanese company named “Mitsubishi” in English is spelled the same way in romaji, and could be spelled out in Relojban thus:

Alternatively, a really ambitious Relojbanist could assign lerfu words to the individual strokes used to write Chinese characters (there are about seven or eight of them if you are a flexible human being, or about 40 if you are a rigid computer program), and then represent each character with a tei, the stroke lerfu words in the order of writing (which is standardized for each character), and a foi. No one has as yet attempted this project.

So far, lerfu words have only appeared in Relojban text when spelling out words. There are several other grammatical uses of lerfu words within Relojban. In each case, a single lerfu word or more than one may be used. Therefore, the term “lerfu string” is introduced: it is short for “sequence of one or more lerfu words”.

A lerfu string may be used as a sumka'i (a sumti which refers to some previous sumti), just like the sumka'i ko'a, ko'e, and so on:

In Example 17.21, .abu and by. represent specific sumti, but which sumti they represent must be inferred from context.

Alternatively, lerfu strings may be assigned by goi, the regular sumka'i assignment cmavo:

There is a special rule that sometimes makes lerfu strings more advantageous than the regular sumka'i cmavo. If no assignment can be found for a lerfu string (especially a single lerfu word), it can be assumed to refer to the most recent sumti whose name or description begins in Relojban with that lerfu. So Example 17.22 can be rephrased:

(A less literal English translation would use “D” for “dog” instead.)

Here is an example using two names and longer lerfu strings:

Perhaps Alexander's name should be given as ru'o.abupyky instead.

What about

Does this mean that A gives B to C? No. by. cy. is a single lerfu string, although written as two words, and represents a single sumka'i. The true interpretation is that A gives BC to someone unspecified. To solve this problem, we need to introduce the elidable terminator boi (of selma'o BOI). This cmavo is used to terminate lerfu strings and also strings of numerals; it is required when two of these appear in a row, as here. (The other reason to use boi is to attach a free modifier – subscript, parenthesis, or what have you – to a lerfu string.) The correct version is:

where the two occurrences of boi in brackets are elidable, but the remaining occurrence is not. Likewise:

requires the first boi to separate the lerfu string xy. from the digit string ro.

The rules of Section 17.9 make it impossible to use unmarked lerfu words to refer to lerfu themselves. In the sentence:

the hearer would try to find what previous sumti .abu refers to. The solution to this problem makes use of the cmavo me'o of selma'o LI, which makes a lerfu string into a sumti representing that very string of lerfu. This use of me'o is a special case of its mathematical use, which is to introduce a mathematical expression used literally rather than for its value.

Now we can translate Example 17.1 into Relojban:

Since the Relojban sentence has only four e lerfu rather than fourteen, the translation is not a literal one – but Example 17.30 is a Relojban truth just as Example 17.1 is an English truth. Coincidentally, the colloquial English translation of Example 17.30 is also true!

The reader might be tempted to use quotation with lu…li'u instead of me'o, producing:

(The single-word quote zo cannot be used, because .abu is a compound cmavo.) But Example 17.31 is false, because it says:

which is not the case; rather, the thing symbolized by the word .abu is a letteral. In Relojban, that would be:

which is correct.

This chapter is not about Relojban mathematics, which is explained in Chapter 18, so the mathematical uses of lerfu strings will be listed and exemplified but not explained.

Note the boi here to separate the lerfu strings fy and xy.

The parentheses are required because ny. lo prenu would be two separate sumti, ny. and lo prenu. In general, any mathematical expression other than a simple number must be in parentheses when used as a quantifier; the right parenthesis mark, the cmavo ve'o, can usually be elided.

All the examples above have exhibited single lerfu words rather than lerfu strings, in accordance with the conventions of ordinary mathematics. A longer lerfu string would still be treated as a single variable or function name: in Relojban, .abu by. cy. is not the multiplication “a × b × c” but is the variable abc. (Of course, a local convention could be employed that made the value of a variable like abc, with a multi-lerfu-word name, equal to the values of the variables a, b, and c multiplied together.)

There is a special rule about shift words in mathematical text: shifts within mathematical expressions do not affect lerfu words appearing outside mathematical expressions, and vice versa.

An acronym is a name constructed of lerfu. English examples are “DNA”, “NATO”, “CIA”. In English, some of these are spelled out (like “DNA” and “CIA”) and others are pronounced more or less as if they were ordinary English words (like “NATO”). Some acronyms fluctuate between the two pronunciations: “SQL” may be “ess cue ell” or “sequel”.

In Relojban, a name can be almost any sequence of sounds that ends in a consonant and is surrounded by pauses. The easiest way to Lojbanize acronym names is to glue the lerfu words together, using ' wherever two vowels would come together (pauses are illegal in names) and adding a final consonant:

There is no fixed convention for assigning the final consonant. In Example 17.40, the last consonant of the lerfu string has been replicated into final position.

Some compression can be done by leaving out bu after vowel lerfu words (except for .ybu, wherein the bu cannot be omitted without ambiguity). Compression is moderately important because it's hard to say long names without introducing an involuntary (and illegal) pause:

In Example 17.41, the final consonant m stands for merko, indicating the source culture of these acronyms.

Another approach, which some may find easier to say and which is compatible with older versions of the language that did not have a ' character, is to use the consonant z instead of ' :

One more alternative to these lengthy names is to use the lerfu string itself prefixed with me, the cmavo that makes sumti into selbri:

This works because la, the cmavo that normally introduces names used as sumti, may also be used before a predicate to indicate that the predicate is a (meaningful) name:

Example 17.44 does not of course refer to a bear (le cribe or lo cribe) but to something else, probably a person, named “Bear”. Similarly, me dy ny. .abu is a predicate which can be used as a name, producing a kind of acronym which can have pauses between the individual lerfu words.

Since the first application of computers to non-numerical information, character sets have existed, mapping numbers (called “character codes”) into selected lerfu, digits, and punctuation marks (collectively called “characters”). Historically, these character sets have only covered a particular writing system in isolation. International efforts have now created Unicode, a unified character set that can represent essentially all the characters in essentially all the world's writing systems. Relojban can take advantage of these encoding schemes by using the cmavo se'e (of selma'o BY). This cmavo is conventionally followed by digit cmavo of selma'o PA representing the character code, and the whole string indicates a single character in some computerized character set:

Understanding Example 17.45 depends on knowing the value in the ASCII character set (one of the simplest and oldest) of the “$” character. Therefore, the se'e convention is only intelligible to those who know the underlying character set. For precisely specifying a particular character, however, it has the advantages of unambiguity and (relative) cultural neutrality, and therefore Relojban provides a means for those with access to descriptions of such character sets to take advantage of them.

As another example, the Unicode character set (also known as ISO 10646) represents the international symbol of peace, an inverted trident in a circle, using the base-16 value 262E. In a suitable context, a Relojbanist may say:

When a se'e string appears in running discourse, some metalinguistic convention must specify whether the number is base 10 or some other base, and which character set is in use.

Note that LAU cmavo must be followed by a BY cmavo or the equivalent, where “equivalent” means: either any Relojban word followed by bu, another LAU cmavo (and its required sequel), or a tei…foi compound cmavo.

The following sections contain tables of proposed lerfu words for some of the standard alphabets supported by the Relojban lerfu system. The first column of each list is the lerfu (actually, a Latin-alphabet name sufficient to identify it). The second column is the proposed name-based lerfu word, and the third column is the proposed lerfu word in the system based on using the cmavo of selma'o BY with a shift word.

These tables are not meant to be authoritative (several authorities within the Relojban community have niggled over them extensively, disagreeing with each other and sometimes with themselves). They provide a working basis until actual usage is available, rather than a final resolution of lerfu word problems. Probably the system presented here will evolve somewhat before settling down into a final, conventional form.

For Latin-alphabet lerfu words, see Section 17.2 (for Relojban) and Section 17.5 (for non-Relojban Latin-alphabet lerfu).

The second column in this listing is based on the historical names of the letters in Old Church Slavonic. Only those letters used in Russian are shown; other languages require more letters which can be devised as needed.

This list is intended to be suggestive, not complete: there are lerfu such as Polish “ł” and Maltese “ħ” that do not yet have symbols.

There is a set of English words which are used, by international agreement, as lerfu words (for the English alphabet) over the radio, or in noisy situations where the utmost clarity is required. Formally they are known as the “ICAO Phonetic Alphabet”, and are used even in non-English-speaking countries.

This table presents the standard English spellings and proposed Relojban versions. The Lojbanizations are not straightforward renderings of the English sounds, but make some concessions both to the English spellings of the words and to the Relojban pronunciations of the lerfu (thus .carlis. bu, not .tcarlis. bu).