From the Logical Languages Wiki
Revision as of 13:15, 23 April 2020 by Maiku (talk | contribs)
Jump to navigation Jump to search
created in: 1995
by: Claudio Gnoli

Liva is a highly regular and schematic logical language featuring an experimental syntax created by Claudio T. Gnoli in 1995 and last publicly updated in 2002.


Liva is a loglang whose description was once hosted on Geocities, during the Internet's early years.


Liva is the creation of Claudio T. Gnoli, an Italian information scientist. Gnoli is also known for a constructed language classification scheme called the Gnoli Triangle.


Liva was created mainly for fun, but also as an experiment in combining logic and aesthetics, and as a means to learn about the structure of language.


The name "Liva" has no meaning; it was chosen on phonaesthetic grounds.


A number of people contributed to the development of the language through discussion and feedback with the author. Other logical languages, especially Loglan and Lojban, were influences.


Liva retains some ideas from the author's earlier constructed language, called Aarbane and later Aarbad. The first version of Liva was developed in 1995 and published in 1996. The most recent published version is numbered 4.2 and dated February 2002.


In this section, we look at version 4.2 of Liva.


Liva is written using the basic Latin alphabet and six extra letters:

grapheme a b c d e f g h i j k l m n o p q r s t u v w x y z ð þ ñ £ î ø
phoneme ä b c d ɛ f ɡ ʁ i ɟ k l m n ɔ p ŋ ʟ ç t u v χ y ʝ ð θ ɲ ʎ ə œ

The correspondence between letters and phonemes, though unusual in some respects, is exactly one-to-one. For convenience, the six extra letters, which are not in the ASCII set, may be replaced with ASCII characters:

grapheme ð þ ñ £ î ø
alternate 6 5 4 2 @ or 1 & or 0

It is conventional to write Liva in all lowercase characters, though all uppercase is allowed. Spaces are used between written words as a visual aid, though not required for segregation.


Liva has thirty-two phonemes, divided into twenty-four consonants and eight vowels. The IPA symbols used in this article were selected to indicate the preferred allophone of each phoneme (at least at a fair guess).

In Liva's defining document, the phonemes are arranged conceptually in an 8-high, 4-wide grid, with each row assigned a symbol. The rows and columns of this original grid are closely reflected in the tables of consonants and vowels presented below, except that the original grid has the voiceless and voiced obstruents placed on separate rows. On the other hand, because the row symbols and other labels employed in Liva's document seem somewhat peculiar and not very intuitive, we've opted not to use them in this article; for example, instead of "Os" we will simply write "voiceless plosive".


labial dental or alveolar palatal velar or uvular
nasal m n ɲ ŋ
plosive p b t d c ɟ k g
fricative f v θ ð ç ʝ χ ʁ
lateral l ʎ ʟ

Some alternate realizations are sanctioned: /θ ð/ as sibilants [s z]; /ç ʝ/ also as sibilants, either alveolo-palatals [ɕ ʑ] or palato-alveolars [ʃ ʒ]; /χ ʁ/ as velars [x ɣ].

Notably, there are four laterals. One of these, the "labial lateral" (spelled with ⟨w⟩), appears to be virtually unattested in the world's languages. For the purpose of representing this phoneme, we've adopted the unusual linguolabial lateral symbol, which is formed by placing the "seagull" diacritic below the standard alveolar IPA letter. For reference, the actual description given for this sound is quoted as follows:

"[T]he easiest way to realize it seems to be by the tongue's tip between lips, either in the middle or at one side: in this position imagine to pronounce a common [l]; a more rigorous alternative, as it does not involve the tongue and is hence truely "labial", would be to put lips near only in their central part and leaving two spaces on the sides, though it seems a difficult position, or to put them near on a side and leave a space on the other side."


front central back
rounded unrounded
close y i ə u
open œ ɛ ä ɔ

The schwa, i.e. /ə/, is close-mid, or perhaps closer, and unrounded. The back vowels are allowed to be rounded.


A syllable is composed simply of a required consonantal onset followed by a required vocalic nucleus.

An onset consists of any of the single consonants, or any of the four nasals followed by one of two (voiceless or voiced) homorganic plosives (but not fricatives or laterals): /mp mb nt .../. (Although the author presents these nasal-plosive pairs as the onsets of morphemic elements rather than of syllables, it will simplify our current analysis to view these pairs as syllable onsets.) Since these clusters seem to occur in syllables that usually follow other syllables, it is presumable that the nasals in these clusters are allowed or preferred to be realized as the coda of the preceding syllable. No other onset is allowed.

A nucleus consists of any single vowel, or any vowel pair consisting of one open vowel and one close vowel but not two of the same (in either order, so either a rising or a falling diphthong). In some cases, as with the numerals, longer strings of vowels occur. These strings contain alternating open and close vowels. It is not clear how these strings are to be syllabified.

There are thus 32 possible onsets; with the eight single-vowel nuclei, 256 syllables are possible and with the 32 possible diphthongs, 1024 syllables are possible.


The only clusters that occur in Liva are nasals followed by homorganic plosives, as already seen. On the other hand, Liva is exceptionally rich in diphthongs.

The only complication to this otherwise simple phonotactic picture is that syllables composed of a fricative or a lateral followed by /ə/ (i.e. ⟨î⟩) are allowed to drop the /ə/ "where this does not make pronunciation and perception too difficult". In writing, an elided ⟨î⟩ is dropped; it is optional whether to write an apostrophe in its place.


Stress, tone and quantity are not contrastive. As a convention it is suggested to place stress on the first syllable of a word or phrase.


Liva is a morphologically isolating language; every morpheme is a word, and every word is a single morpheme.

Word shapes

Words consist of one or more syllables. The onset of the first syllable of a word may not contain a voiced obstruent. Specifically, a word may start with one of the following 20 onsets: ⟨m, n, ñ, q; p, t, c, k; f, þ, s, x; w, l, £, r; mp, nt, ñc, qk⟩; a word may not start with one of the following 12 onsets: ⟨b, d, j, g; v, ð, z, h; mb, nd, ñj, qg⟩.

There are thus 160 possible monosyllabic words containing monophthongs (20 onsets × 8 vowels), and 640 possible monosyllabic words containing diphthongs (20 onsets × 32 diphthongs).

Polysyllabic words are formed by combining two or more monosyllabic words. The non-initial monosyllabic words must be those that contain voiceless obstruents as independent words; these obstruents are replaced with their voiced counterparts to obtain compounding forms; for example, ⟨qou⟩ plus ⟨pa⟩ yields ⟨qouba⟩ and ⟨pa⟩ plus ⟨ntou⟩ yields ⟨pandou⟩. Since this compounding process does not reliably result in predictable meanings, these polysyllabic words are regarded as single morphemes, with an exact meaning to be defined in the dictionary.

Word Classes

The initial onset of a word indicates its class as follows.

  • Atomic syntagms (called deictics) begin with a nasal ⟨m, n, ñ, q⟩.
  • Nouns begin with a voiceless plosive ⟨p, t, c, k⟩.
  • Relators begin with a voiceless fricative:
    • symmetric relators begin with ⟨f⟩;
    • dyadic relators begin with ⟨þ, s⟩;
    • triadic relators begin with ⟨x⟩.
  • Quantifiers begin with a non-labial nasal plus voiceless plosive ⟨nt, ñc, qk⟩.
  • Qualifiers (called intensionals) begin with ⟨mp⟩.
  • Markers begin with a lateral:
    • word markers begin with ⟨w⟩;
    • predication markers begin with ⟨l⟩;
    • relation markers begin with ⟨£⟩;
    • quantification markers begin with ⟨r⟩.


The basic unit of Liva's syntax is called a syntagm. Syntagms are something like the sentences or clauses of linguistics, or the formulae and start symbol of a formal language, but more encompassing and displaying a wider range of semantics than these conventional categories. Syntagms may occur either independently and as nested constituents. As an overview, here are the (incomplete) production rules of sytagms (adopted here with some minor changes for clarity and to eliminate a few intermediate symbols):

syntagm := atomic-syntagm
syntagm := syntagm quantifier
syntagm := syntagm qualifier
syntagm := syntagm noun
syntagm := syntagm relation

Relations have the following syntax:

relation := symmetric-relator syntagm syntagm+ 
relation := dyadic-relator syntagm syntagm
relation := triadic-relator syntagm syntagm syntagm 

The symbols not defined here (atomic-syntagm, noun, etc.) are exactly our word classes, already given in the preceding section. The above rules are not complete; the role of "marker" words is not indicated here, but is explained later.

Atomic syntagms

Liva contains a heterogeneous class of particle-like words allowed to stand as one-word syntagms. The author calls these one-word syntagms "deictics", but they might have just as well been called "primitive syntagms"; we've opted to call them "atomic syntagms". It can be observed from the production rules that every syntagm must start with an atomic syntagm. There are several subclasses, listed as follows with a few examples:

  • Personal: ⟨me⟩, "I, me"; ⟨ma⟩, "you and me"; ⟨mo⟩, "you", ⟨mø⟩, "he, she, it, they, them".
  • Attitudinal: ⟨my⟩, generic exclamation; ⟨mu⟩, exhortative, "please, come on".
  • Demonstrative: ⟨nø⟩, "the"; ⟨ne⟩, "this"; ⟨na⟩, "that"; ⟨no⟩, "that yonder".
  • Interrogative: ⟨ny⟩, "what?".
  • Sequential: ⟨nî⟩, "present (here, now, etc.)"; ⟨ni⟩, "what precedes the present"; ⟨nu⟩, "what follows the present" (in a generic linear sequence, or in temporal sense).
  • Generic: ⟨ñî⟩, "something".
  • Modal/polarity: ⟨ñe⟩, "irreality, not"; ⟨ña⟩, "possibly, perhaps, maybe"; ⟨ño⟩, indicative, "reality, existence, yes".
  • Phoric: ⟨qî⟩, "the mentioned"; ⟨qi⟩, "the above"; ⟨qu⟩, "the following".
  • Relative: ⟨qø, qe, qa, qo⟩, argument x0, x1, x2, x3 of the following relator.
  • Variables: ⟨qy⟩, generic variable, "x".

As can be seen, the word class of atomic syntagms encompass both pronominal and adverbial functions.


When paired with a syntagm to its left, a noun describes a property of the thing referenced by that syntagm. For example, combining the syntagm ⟨na⟩, "that", with the noun ⟨ta⟩, "food", yields:

⟨na ta⟩, "that food", "that is food".

The resulting syntagm can be translated either as a noun phrase or a clause; Liva makes no distinction.

Noun-syntagms can be nested, which leads to several possible translations:

⟨na ta tu⟩, "(that food) good", "that good food", "that is food and such food is good".


Unlike nouns, relators are not paired directly with a preceding syntagm, but rather indirectly through an intermediate structure called a relation, which is then combined with a preceding syntagm. The relation is composed of the relator followed by arguments which are syntagms. For example, the relator ⟨xa⟩, "go", takes three arguments, forming the following relation:

⟨xa ñîkî ñîte ñîka⟩, "go [by] some-person [from] some-city [towards] some-building".

A relation is combined with a syntagm, often the positive polarity particle ⟨ño⟩, meaning indicative mood or "yes", to form a complete syntagm:

⟨ñoxa ñîkî ñîte ñîka⟩, "Someone goes from some city to some building".

There are several types of relators: dyadic, triadic, and symmetric, the latter of which allows for an indefinite number of arguments.

Quantifiers and qualifiers

Quantifiers and qualifiers (called "intentionals") have a syntax similar to nouns; they modify a preceding syntagm. There is a variety of kinds, listed here with some examples:

  • Quantifiers: ⟨nta⟩, "two"; ⟨qko⟩, "all".
  • "Typologicals": ⟨qku⟩, "the typical, the paradigmatic"; ⟨qki⟩, name, "the individual called".
  • Generic quantities: ⟨ñcø⟩, "a few"; ⟨ñca⟩, "several", plural.
  • "Intensionals": ⟨mpi⟩, "almost, about, some kind of"; ⟨mpu⟩, "very, extremely".

Numerals, which are reckoned in base 16, start with ⟨nt⟩ and add the following digits (the small digits have two variants):

nucleus -ø-
-ey- -ei- -eî- -eu- -ay- -ai- -aî- -au- -oy- -oi- -oî- -ou-
value 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Larger numerals are expressed with ⟨ntu-⟩, followed by a digit indicating a power, followed by ⟨-nd-⟩ (i.e. the compounding form of the numeral starter ⟨nt⟩), followed by a string of digits (diphthong variants). For example:

⟨ntuendøiøî⟩ = ⟨ntu-e-nd-øi-øî⟩ = 1216 = 1×161 + 2×160 = 18;
⟨ntuandøi⟩ = ⟨ntu-a-nd-øi⟩ = 10016 = 1×162 + 0×161 0×160 = 256 (note the trailing zeros can be omitted).


Markers are particles that perform a number of tasks. They come in several sorts: word markers, predication markers, relation markers, and quantification markers.

Word markers are used to delimit foreign expressions such as personal and geographical names, e.g. ⟨wa helsinki w'⟩, "Helsinki". There are several variants, some of which are used to indicate the syntactic category of the delimited expression: ⟨wa⟩ precedes a noun, ⟨we⟩ precedes a quantifier, etc.; ⟨wî⟩ = ⟨w'⟩ delimits the end of a word. Note that ⟨w⟩ was selected as the onset for word markers precisely because the sound was expected not to occur in foreign expressions, so that foreign expressions might be delimited unambiguously.

Predication markers are used to change the way predication (nouns and relations) works: ⟨li⟩ indicates that the preceding syntagm is modified by the following predicate, e.g. ⟨na [li] ta⟩, "that food" (this is the default behavior, so ⟨li⟩ is rarely used; ⟨lî⟩ = ⟨l'⟩ blocks predication and seems to be functionally equivalent to the generic atomic syntagm ⟨ñî⟩, e.g. ⟨lî ta⟩ = ⟨ñî ta⟩, "(some) food"; finally ⟨ly⟩ reverses the order of predicate and syntagm, e.g. ⟨ta ly na⟩, "It's food, that".

Relation markers are used to reorder, skip, and otherwise manipulate the arguments of relators. They are placed in front of arguments. The markers ⟨£e, £a, £o⟩ indicate arguments 1, 2, & 3 respectively; ⟨£ø⟩ is used to mark arguments of symmetic relators (which have an indefinite number of arguments) after the first two (which are unmarked); ⟨£î⟩ marks the end of a relation. In addition, there are relation markers to mark reciprocals and perform other functions.

Quantification markers are used to indicate how quantifiers are to be applied: ⟨ry⟩ indicates that quantifiers are applied collectively and ⟨rø⟩ indicates that quantifiers are applied distributively.

Self-Segregation Strategy

The self-segregation strategy is relatively simple and straightforward.

Words start with certain onsets, and those same onsets can only appear at the start of words. Therefore the boundaries of words are always clear.

Syntagms consist of an atomic syntagm followed by zero or more nouns, relations, qualtifiers , and qualifiers.

The default valency of relators is indicated by the onset of the word. This valency can be overtly overridden using relation markers.

Foreign expressions can be incorporated by bracketing them with explicit word markers.


Explain the origin of words and morphemes (apriori, aposteriori, etc.). How are proper names handled?

Analysis(?) Critical assessment(?) Impact(?)

  • Novel ideas?
  • The impact this language has had on the LL world?


Language Descriptions
Logical Languages

Loglan (1955)[stub]Lojban (1987)[stub]Voksigid (1991)[stub]Gua\spi (1991)[stub]Liva (1995)[stub]Lojsk (2002)[stub]X-1 (2005)[stub]Trari (2009)[stub]Xorban (2012)[stub]Toaq (2013)[stub]Tanbau (2014)[stub]Eberban (2020)[stub] Xextan (2022)[stub]

Other Languages

Ceqli (1996)[stub]Latejami (1995)[stub]Ithkuil (?)[stub]CycL (?)[stub]IKL (?)[stub]