From Lojban
Jump to navigation Jump to search

tanbau is a logical language developed by la tsani.


Grammatical description of the constructed language. Version 0.1a

Tanbau is a logical language constructed by tsani (Jacob Thomas Erringon) with ideas from Lojban, Gua\spi, and Toan Dzu. By using the more powerful parsing expression grammar formalization, Tanbau uses an extremely rich system of suffixes for grammatical purposes. It tries to solve some of the problems that were discovered with regards to Lojban, but above all else, it tries to build a new kind of formalization for logical languages.

Chapter 1: Phonological and Morphological Properties

Section 1: Phonology

The phonology was originally simply based on Lojban's, but certain difficult consonants were removed, namely the velar fricative and the glottal stop. Vowels, on the other hand were added, bringing the vowel count up to nine.

Having not studied phonology, I offer the following description of Tanbau sounds, based on English, French, and Spanish sounds.

1.1.1 Consonants

Consonants are divided into three categories: voiced, unvoiced, and liquid:

(1) the voiced consonants: b, d, g, j, v, z (2) the unvoiced consonants: p, t, k, c, f, s, h (3) the liquid consonants: l, m, n, r

Their sounds are essentially as in English, but with the following major differences:

(4) 'g' is pronounced /g/ as in the English word "gaunt" not as in "giant". (5) 'j' is pronounced like the 's' in the English word "measure". (6) 'c' is pronounced like the 'sh' sound found in the words "shine" "bash" and "shame". (7) 'r' can be realized as any rhotic sound.

1.1.2 Consonant Clusters

All clusters in Tanbau are binary; there are no consonant triples. The only invalid clusters are voiced-unvoiced/unvoiced-voiced pairs as well as consonant duplicates. 'h' is a non-clustering consonant and cannot cluster with anything. The liquids can cluster with anything, except 'h'.

1.1.3 Vowels

The tanbau vowels consist of the five Lojban vowels {a, e, i, o, u}, as well as the vowels denoted {y, w, ee, aa}. There are several typographical variants for the two last vowels, and they are called "epsilon" and "schwa" respectively.

(1) 'a' is the sound in English "father", Spanish "casa". (2) 'e' is the final sound in French "aimé". (3) 'i' is the sound in English "machine". (4) 'o' is the sound in French "beau". (5) 'u' is the sound in English "boot". (6) 'y' is the sound in French "pu", "tu". (7) 'w' is the sound in French "peut". (8) Epsilon is the sound in English "bet". (9) Schwa is the initial vowel sound in English "american".

Section 2: Morphology

Tanbau's morphology is self-segregating--that is, for a given valid stream of sounds, there exists exactly one possible rendition in text. Likewise, from a given text, there are only certain valid pronunciations, all of which would be recognized only as that text. This is called audio-visual isomorphism (AVI), and Tanbau arguably does it better than Lojban. However, Tanbau's *mechanism* for achieving AVI is roughly the same as Lojban's, using a combination of stress and consonant clusters. To understand how words can thus be made to self-segregate as such, we must first examine the structure of words, of which there exist only two types.

1.2.1 Predicate Words

This class is a bit of a misnomer, since it encompasses many things that are not predicates (a concept we have yet to define, but bear with me--it is important to cover the machinery of the language before describing how to put that machinery to use in order to produce meanings). Let Ci be a predicate-initial consonant, Cf be a predicate-final consonant, Cs be a suffix-initial consonant, and V be any vowel. Predicate words are composed of at least one stem and at least one suffix, where a stem has the form CiVCf and a suffix has the form CsV(l)?. As their name indicates, suffixes are suffixes: they come after one or more stems. The following expression matches all predicate words.

(1) predicate word pattern: ((stem+)suffix+)+

The minimal predicate word is therefore composed of one stem and one suffix. However, we have yet to define what these special character classes described above contain.

(2) contents of Ci: all consonants except {h, l, r}. (3) contents of Cf: all of Ci, plus {r}. (4) contents of Cs: all of Ci.

Recall: Ci is composed precisely of {t, p, s, d, f, g, j, k, z, c, v, b, n, m}.

(5) sample stems: ban-, jan-, taz-, gyz-, dwm- (6) sample suffixes: -ji, -dw, -jo, -kal

Note: It is customary to write bare stems with a hyphen attached on the end to denote the incompleteness of the word, and that the word is expecting something to be attached at that location. Likewise for the suffixes, we typically write a hyphen preceding the suffix to indicate that there should be something attached there.

Stress is important in the determination of word boundaries, i.e. it is crucial in the self-segregating morphological system. In a predicate word, the penultimate (second-to-last) syllable is stressed. This can be marked with a grave accent {`}, although this is optional when writing with spaces.

Note: writing with spaces between words is unnecessary when using the grave accent writing system.

(7) sample predicate words with stress indicated: tèdjo ("the head"), taspà jo ("the one named 'sky'"), temnomìljo ("the little boy")

Since {l} is unused in the consonants of stems, it is allocated as a possible final consonant for suffixes. It does not produce a consonant cluster in the sense that we will be using to discuss self-segregating morphology. It would be more useful to consider it a vowel in this case.

Predicate words have some intrinsic morphological properties:

Definition: the _size_ of a predicate word is the number of stems in the word. The _raw_complexity_ of a predicate word is the number of stems plus the number of suffixes in the word.

We say "a predicate word of size 1" or a "1-part predicate word" to talk about words such as "taspajo" and "peenfalgabu" ("to want to cause friendship"). We say "a predicate word of raw complexity 6" or a "6-raw-complex predicate word" to talk about words such as "taspataknobu" ("to want to talk to the one named 'sky'") and "jibnaasevnotunmo" ("to be binoculars/telescope"). The raw complexity of a predicate word is equal to its syllable count.

1.2.2 Operators

An operator is composed of an operator-initial consonant, any vowel, and possibly the letter 'l'. Operator-initial consonants are denoted by the character class Co, which is very similar to the stem-final consonant class.

(1) contents of Co: all of Cf, plus {r, h}.

Therefore, the following expression matches all operators:

(2) operator pattern: CoV(l?)

(3) sample operators: ba, jil, kw, bee, jaal, ha, ry

Operators are typically unstressed, although stressing them is optional and has no effect on the determination of word boundaries, unlike in Lojban where a misplaced stress on a cmavo can have undesirable results.

Chapter 2: The Sentence

A text in Tanbau is composed of a sequence of sentences, which come in two kinds: answers to questions and predications. The latter kind, being more interesting, will be the focus of this chapter, where we will explore how to construct and manipulate sentences.

The general structure of any sentence is the following: [separator] [terms...] <predicate> [terms...] [illocutionary operator]

The type of brackets used is meaningful: the predicate is the only required part of the predication, and being so crucial will be the first object of our discussion.

The predicate is the most meaningful object in all of Tanbau grammar, but is likely equally the most complex. It is the basis for many other structures, namely most kinds of terms. Semantically, the predicate governs the relationship between the terms: it is a function that relates them and produces the meaning of the sentence. The terms following the predicate form a sequence--the order of the terms is meaningful. We consider a predicate definition of the form "kan- = x1 moves to x2 from x3," and the sequence of terms in the predication fills the sequence of terms in the definition. The term sequence of the definition is called _the_place_structure_ of the predicate.

Section 1: The Predicate

As mentioned earlier, not all predicate words turn out to be predicates. Indeed, to understand how to actually make predicate words that are predicates, we have to consider that a suffix is a monadic function whose argument is the predicate immediately on the left. We say "immediately on the left" since suffixes are myopic: in the case of "zygnobanmo", the suffix -mo is modifying only the stem zyg-, and the suffix -no is modifying only the stem ban-.

Indeed, suffixes having multiple functions leads us to group suffixes into two broad categories: transformative suffixes and semantic suffixes. Transformative suffixes produce a predicate word that is no longer a predicate, whereas semantic suffixes produce a new predicate with an altered meaning (via connotation or an altered place structure.)

Note: We simply call transformative suffixes "transformatives" for the sake of brevity.

Therefore, since semantic suffixes produce predicates, if the final suffix of a predicate word is a semantic suffix, then the predicate word is a predicate.

1.1.1 The Syntax of Terminal, Nonterminal, and Compound Predicates

Let's briefly look at different mechanisms of compounding to have a better understanding of predicates and the functions of suffixes. We won't discuss the semantics of compounds until a much later chapter dedicated to that more complex topic.

The conclusion of the above section was slightly misleading: if the final suffix of a predicate word is a semantic suffix, then it is a _terminal_ predicate. This has to do with how phrasal compounds, also called serial predicates ({tanru} in Lojban), are constructed.

Definition: a _serial_predicate_ or _phrasal_compound_ is composed of a sequence of nonterminal predicates and a single terminal predicate. A _terminal_predicate_ is a predicate word whose final suffix is a semantic suffix. A _nonterminal_predicate_ is a predicate word whose final suffix produces a nonterminal predicate.

Consider the following suffixes: -ne (transformative) = return a nonterminal predicate. -mo, -no (semantic) = null suffix.

The null suffix exists in order to satisfy the rule that predicate words be at least of raw complexity 2, a requirement of the self-segregating morphology. Thus, the null suffix, despite being among the semantic suffixes, has no real semantic value. Likewise, the suffix -ne, a transformative, has no semantic value either; the function of -ne is purely syntactic. We are thus led to define another kind of complexity.

Definition: the _semantic_complexity_ of a predicate word is the number of semantically productive components in that word. Stems and semantic suffixes at the exception of the null suffixes -no and -mo are semantically productive. The _semantic_efficiency_ of a predicate word is the ratio of its semantic complexity to its raw complexity. Numbers closer to 1 are better.

Can a semantic suffix come after a transformative suffix? No. Semantic suffixes are allowed only after a stem or another semantic suffix. Equally, we have to consider what type of predicate word such a beast would be; the presence of the nonterminal transformative makes us think that the word should be a nonterminal predicate, but then the semantic suffix recalls the definition of the terminal predicate. This kind of ambiguity is to be avoided, and so a semantic suffix cannot appear after a transformative.

1.1.2 The Syntax of Lexical Compounds

Whereas the above section concerned itself with the description of compounds composed of more than one word, this section will discuss compounds that are within a single word.

When a predicate word has a size greater than one, then it can be referred to as a lexical compound. We call the individual stem+suffixes groups "sub-predicates". Non-last sub-predicates can be unsuffixed, in which case we consider them to be equivalent in meaning to their null-suffixed analogue, e.g. "banzygno" = "banmozygno".

Rule: the non-last sub-predicates of a lexical compound predicate word must each hypothetically stand as terminal predicates.

Therefore, compounds such as "tamnebanmo" are illegal, since in isolation, "tamne" is a nonterminal predicate, not a terminal predicate.

1.1.3 Constructed Predicates

So far, we have only seen predicates that stem from... stems. Tanbau lets us build predicates from more abstract things such as numbers and even other sentences, expanding its expressiveness. These predicates are all called constructed predicates. Unfortunately, the syntactic description of these predicates requires knowledge of how to form numbers and use terms, which is only covered later.

Definition: a _constructed_predicate_ is a predicate that is built using a mechanism that could involve no stems. All constructed predicates are nonterminal.

Numeric predicates carry meanings of cardinality, ordinality, and rank, which tend to be difficult to express without numbers. All numeric predicates follow the pattern <number> <numeric predicate operator>.

See also: Chapter ???? for a discussion on numbers.

Nonce predicates are built using a kind of subordination operator followed by an entire sentence. The sentence must contain at least one overt or covert "slot" anaphora. The illocutionary operator of the sentence is generally required to be the function-terminator operator (if omitted, this is the operator that is assumed.)

See also: Chapter ???? for a discussion on anaphora, and Chapter ???? for a discussion on illocutionary operators.

Argument-based predicates have a definition that derives from the argument used to build them. The operator "vy" followed by one argument (a kind of term) constitutes a nonterminal predicate with the meaning "x1 is among the referents of [argument]".

Finally, a predicate can be boxed into a group, which itself is nonterminal. Such a predicate is called a group predicate. The group's boundaries are the bracket-operators "kee" and "keel", and the interior must be a full predicate, i.e. zero or more nonterminal predicates followed by a terminal predicate.

See also: Chapter ???? for a discussion on the semantics of serial predicates.

Section 2: Terms

Now that we can talk about relationships between things, it would be good to learn how to refer to things are relate them. Not all terms are related by the predicate, however. Those terms are the most important for the sake of being able to *say* things, and will be the first topic of discussion. We will then explore how to construct adverbs and prepositions.

2.1.1 Arguments

The simplest argument consists of a single operator; these are the anaphora.

gi argument anaphora = I/me the speaker/author. ji argument anaphora = You the listener/reader.