|This article is about the Tlaliolz language. For the people, see Talo.|
The Tlaliolz "language" was in fact a group of diverse but related dialects, which together constituted the Talo branch of the Talo-Edastean language family. The dialect for which we have the greatest attestation is the variety the Fáralo came into contact with in the classical period, here referred to as Eastern Tlaliolz. The following description is of this dialect. Some basic information about the western dialects is available as well.
| Tlaliolz |
|Period||c. 100-300 YP|
|Spoken in||Lu Tal|
|Classification|| Talo-Edastean languages |
|Basic word order||VSO|
- 1 Genealogy
- 2 Attested words
- 3 Phonology
- 4 Morphosyntax
- 4.1 Nouns
- 4.2 Verbs
- 4.3 Modifiers
- 4.4 Basic Clauses
- 4.5 Complex Clauses
Tlaliolz is the only known language of the Talo branch of Talo-Edastean, which in turn is part of the larger Macro-Edastean family. It is most closely related to Ndak Ta, the language of the ancient Ndak Empire, and Antagg, spoken on the upper Bwimbai. Another, more remote relative of Tlaliolz is Proto-Xoronic, the ancestor of the Habeo languages and Damak.
- Proto-Macro-Edastean (c. -3500 YP)
The following Tlaliolz words from around 300 YP are currently known:
- aqe - to move, walk, go
- ākʷ’awi - to push
- ean - to die
- ēq’u - to kill
- falzam - birch
- Fāzzin - Fáralo
- featli - unique
- fiekʷi - to yell, cry out
- isaf - man
- kieq’o - legendary
- kʷēkā - game (hunted animals)
- k’ān - to win
- k’aefī - stag (a male deer)
- k’eomun - daisy
- lēts - sword
- liamaqa - to reveal, to give away a secret
- līk’litsa - ash (tree)
- loa - to take
- lōfleatsa - ring, circle, loop
- lūk’im - agile; competent
- mak’āt - hat
- meotlo - to freeze
- meonzi - spear
- mēnaya - purposefully
- mēnu - reason
- mētlatso - shoe
- milūyo - serene, composed, calm
- milūyoya - reassuring
- moyanāta - ??? (definition is missing)
- naokʷi - fur
- nazwā - wren
- nāsqak - pain
- nilīk’ - honest
- nitli - woman
- non - to go
- ōtla - to stop
- pāzi - to have
- pisfi - to injure
- saya - to hold, carry
- selantla - wild cat
- setiek - longbow
- sōnata - firefly
- sūzatsa - press
- tāmepa - to imitate
- tayolam - bronze
- tastsefa - tin
- tatstlik’ - shirt, tunic
- teatsi - arrow
- tetsēle - to confirm
- tiliozo - principle, conviction
- toezo - to pack
- tlaliolz - the Tlaliolz people
- tlemezu - tribe; nation
- tleyōk’ā - blue jay
- tsateonna - to recommend
- tsit - to put, place
- tuq’antis - gemstone
- qoa - therefore
- qōfena - creek, small river
- qūtl - to run
- qʷotsfeloz - flight, exodus
- qʷotsfi - to flee
- q’om - to gore, to stab
- q’utla - sack
- ūq’ - egg
- wekʷ’aza - aroma
- yaltlēk - kettle, pot
- yayo - to laugh
- zaomōya - passionate, enthusiastic
- zātl - to hear
- Ejective t’ was a marginal phoneme occuring only in a few words borrowed from the western dialects, and may have been present in several ideophones as well.
|high||i · iː||u · uː|
|mid||e · eː||o · oː|
|low||a · aː|
- There are two mononucleic diphthongs, /ae/ and /ao/.
The spelling convention used here is designed to be a straightforward, phonemic spelling, derived mainly from the IPA but influenced by Fáralo transcription habits (especially for z).
- Most consonants are spelled with IPA (e.g. kʷ) or the text equivalent thereof (e.g. ts).
- The exceptions are: /j/ is spelled y, /ɸ/ is spelled f, /tɬ/ is spelled tl, and /ɹ̝/ is spelled z ([z] is an allophone).
- There are only two digraphs: ts and tl, both of which represent affricates.
- Apostrophes represent the ejective diacritic, and when applicable, follow the superscript ʷ; e.g. kʷ’.
- Long vowels are indicated with macrons.
- Glides [j] and [w] go unwritten wherever their appearance is predictable.
Our evidence for Tlaliolz allophony is insufficient to reconstruct more than a few details and speculations. To summarize them:
- z was probably a coronal approximant when clustered with plosives, or perhaps with all obstruents. In other environments it was probably weakly assibilated or may even have been a full sibilant [z].
- The behavior of z was very likely true of two other approximants as well; l was probably a partially or fully spirantized [ɮ] whenever not in contact with an obstruent, and y may have been a voiced palatal fricative [ʝ] in the same environments.
- l may also have been velarized in some or all positions.
- The nasals, m and n, probably assimilated to the point of articulation of any following plosive.
- Based on evidence from loanwords in dialects of Ndok Aisô, the ejective plosives were probably reduced to aspirated plosives in word-final position.
- The dorsal plosives k and q (and the ejective and/or labialized variants thereof) may have reduced to equivalent fricatives, e.g. [x] and [χ], whenever immediately before another plosive.
- Short vowels were probably significantly laxened or lowered from the near-cardinal values of the long vowels. Most Fáralo scribes transcribed long ē and ō with Fáralo ei and ou but short e o with Fáralo e o. Short a was sometimes transcribed as Fáralo ə in unstressed positions, suggesting a vowel reduction process may have been at work in Tlaliolz as well.
- Two-vowel sequences were almost certainly pronounced as two full syllables, the two vowels being "held apart" by an intervening glide: [j] after i and e, and [w] after o. The exceptions were ae and ao, which probably each comprised only a single syllable (this is why they are listed as independent phonemes).
The great majority of Tlaliolz morphemes follow a syllable pattern of (C)V(ː)(C); an optional onset consonant, a required vowel nucleus, an optional offglide or lengthening of the nuclear vowel, and an optional coda consonant. However, there are sufficient exceptions in the corpus to demonstrate this pattern was not a strict rule, a conspicuous example being the name of the language itself which ends in a cluster. Medial clusters of three components can be found in about a dozen words as well, and in one case, a four-consonant cluster - although this one is not beyond doubt, as we know of it from only one source. There appear to have been few if any restrictions on which consonants could cluster, or in which directions, though some types were certainly more common than others.
Two-vowel sequences occur in Tlaliolz with high frequency; sequences of more than two cannot be found in the corpus. Three important distribution rules should be kept in mind:
1. The vowels in a sequence were both short; long vowels finding themselves adjacent to another vowel due to morphology always shortened.
2. Short u is never found in any vowel sequence. A long ū which must shorten because morphology puts another vowel next to it, shortens to o.
3. ae and ao, while historically arising from vowel sequences, were in fact mononucleic diphthongs (diphthongs which comprise a single syllable). Equivalent sequences formed by morphology also reduced to diphthongs.
Tlaliolz nouns show nominative case by taking the prefix li-, and most do not inflect in any other regard - not for number, gender, possessor, or anything else. The absolutive case in unmarked, and verb suffixes are the sole indication of their grammatical number.
- The nominative is used for agentive nouns.
- The absolutive is used for patientive nouns.
- The absolutive is also used for everything else - for objects of prepositions, and for stand-alone utterances such as answers to questions or addressing someone by name.
However, there is a class of nouns which do change form for plural number, by reduplicating the first syllable - the first CV or VC. It is to be noted that few native words exhibit this reduplication; the pattern appears to have been picked up from the source language of the borrowed words, though which language it was may never be known. The very few native words that show reduplication most probably gained it by analogy.
Reduplicating nouns do so only in the absolutive. In the nominative they do not change form for number.
sōnata, "firefly", non-reduplicating:
- absolutive: sōnata
- nominative: lisōnata
ūq’, "egg", non-reduplicating:
- absolutive: ūq’
- nominative: lioq’
teatsi, "arrow", reduplicating:
- absolutive singular: teatsi
- absolutive plural: teteatsi
- nominative singular: liteatsi
- nominative plural: liteatsi
There are three sets of personal pronouns in Tlaliolz: basic, possessive, and reflexive.
All plural forms simply suffix -k to the pronoun's singular form.
Basic pronouns are not used for the syntactic pivot of a clause, which is sufficiently marked already by the verb suffix.
Tlaliolz verbs inflected for aspect, and to agree with the pivot for its person, number, and case; all these categories were somewhat fused into a single suffix, though patterns are still quite apparent in them - observe, just to start with, the close correlation of the suffix vowel with grammatical person.
Every finite verb had a verb suffix.
Using for example the verb ēq’u "to kill", the verb suffixes of Tlaliolz were as follows. The suffixes are shown in bold, and are normally simply appended to the verb stem. However, notice that singular aorist suffixes begin with vowels, causing vowel sequences or long vowels to occur; the usual rules for them apply here, such as u changing to o in vowel sequences. First, here are the suffixes for first person:
The verb "kill" is not actually attested in the aorist aspect, as most verbs of action are not; it usually appears in stative clauses, while action verbs usually appear in active ones. But there appear to be no irregular verbs in Tlaliolz, so we can nevertheless say with confidence what forms "kill" should take in the aorist.
Next let's skip ahead to third person because the third-person suffix set is identical to the first-person set, save that the vowel is a instead of i:
In the second person, the consonants of the suffix change form as well. Observe also that the suffix vowel in the aorist singular, o, combines with the final o of the root into a single long vowel. If the verb's final vowel were anything else, a vowel sequence would be formed instead.
"Modifier" is a mixed grab-bag category of various word types in Tlaliolz (much as "adverb" is in English). The differences between these words and their uses are poorly understood at this time. Nevertheless, we can be fairly sure of a few basic traits of the class:
- Modifier roots are those which never, in the corpus, appear as a verb or as the head of a noun phrase without any derivational morphology.
- It appears that many of them were fluid enough in function to modify nouns, verbs, or whole clauses without any change in form, while others were restricted to just one of those functions.
- Modifiers can be derived from nouns and verbs; but these behave differently from "native" modifier roots, in many regards.
- No inflectional morphology specific to modifiers can be identified.
Predicative uses of modifiers are covered under the section "Predication".
"Native" modifiers - those roots which belong to this class by default - do not usually stand alone when used attributively. Instead they are incorporated into the noun or verb root.
Modifier incorporation exhibits the following strictures and structures:
- A modifier is suffixed to the end of a noun or verb. In theory any number could be attached, but there are very few examples in the corpus of more than three, and most of those are personal names of chiefs.
- A noun plus a suffixed modifier constitutes a new noun, which may then have another modifier suffixed; thus modifier incorporation is nested in nature. For example isatānqʷi should be read as [[[man]southern]big], not [man-southern-big].
- Incorporation was productive: in most cases any modifier could be combined with any noun or verb root, and in fact were required to as a part of the normal machinery of Tlaliolz.
- Incorporation was compositional: again in most cases, the meaning of a word with an incorporated modifier was trasparently derived from its components. However, some instances had acquired non-compositional meanings; some of these are listed separately in the lexicon as compounds.
- The addition of a suffixed modifier very frequently caused the elision of sounds from the end of the word it was suffixed to - often just a consonant or a vowel would be lost, but other times whole syllables would be deleted. There seem to be few if any rules for where this elision occurred, and was likely up to the speaker's judgement of how far he could reduce a word without risking being misunderstood.
- If a second modifier was suffixed, the end of the first was similarly subject to elision, and so forth; beginnings of modifier roots, however, rarely show elision. isatānqʷi also shows elision at work; the root is isaf, "man".
- Assimilation of nasals to a following point of articulation applied at the boundaries between roots and suffixed modifiers. This too is shown in isatānqʷi - the first modifier is otherwise tām.
Evidence of incorporation
One may justifiably ask whether these suffixed modifiers were actually independent words. The evidence that they were not is as follows.
- Phonological: nasal assimilation and especially sound elision are processes much more typical of word-internal morphophonemics than of behavior across a word boundary.
- Morphosyntactic: inflectional verb suffixes always followed the last incorporated modifier. Derivational suffixes on both nouns and verbs tended to as well, although not always. Roots with incorporated modifiers were moved as single units in transformations.
- Philological: in most Fáralo transcriptions, modified nouns and verbs were written as single words.
- Names: incorporated modifiers appear in the personal names of a majority of Tlaliolz people mentioned in the corpus.
Modifiers that have been morphologically derived from nouns or from verbs are usually not incorporated. Instead, they tend to be placed before or after their head. Both word orders (modifier-head or head-modifier) are documented in Tlaliolz for both nominal and verbal heads, but the least-marked position appears most likely to have been after the head. Preposed modifiers may have been a form of left-dislocation employed for contrastive focus or other pragmatic purposes, although data in the corpus is not adequate to be certain of this.
"the manly woman"
In Tlaliolz, a "basic" clause is one in which the complete construction includes at most one verb with a single root, plus an optional auxiliary. Most basic clauses, the verb is the first element to occur. Nouns follow the verb in an order dependent on clause type - active or stative.
In a clause with two nominal arguments (noun phrases, pronouns, etc.), i.e. a transitive clause, the one seen as originating the action (the "agent") will be marked with the nominative case; the other will be marked with the absolutive case, and both will follow the verb.
In a clause with only one nominal argument, or intransitive clause, the argument may be marked as nominative if it is more like an agent (causing or controlling the action), or it may be marked as absolutive, if it is more like a patient (receiving the action). In either case it follows the verb.
Note that this behavior of intransitive clauses is quite unlike what most other languages do. It makes Tlaliolz an active-stative language of "fluid-S" subtype.
The syntactic pivot
The "syntactic pivot" is the nominal argument that a Tlaliolz clause agrees with. In a clause with only one argument, that argument is the pivot. In a clause with two arguments, the pivot is the absolutive argument in a stative clause, or the nominative argument in an active clause - and usually is placed before the argument that is not the pivot. Every clause has exactly one pivot, but in many cases it is not expressed directly, and is referred to only by the verb suffix (Tlaliolz is a pro-drop language).
Active and stative clauses
An active clause is usually used to express an action or event, and has two syntactic characteristics: its syntactic pivot is in the nominative case, and if the clause is transitive, it follows the word order VSO (verb-subject-object). Most often an active transitive clause will cast the nominative argument as the most salient or topical of the two.
A stative clause is normally used to express a state of affairs, or to give a passive-like sense to the clause. It has two syntactic characteristics, in opposition to those of active clauses: the pivot of a stative clause is in the absolutive case, and if it is transitive, the word order followed is VOS (verb-object-subject). In a stative clause, the absolutive argument is usually cast as the most salient, and is most often a patient or an experiencer (the theta roles - look them up if you need to).
The interaction of syntactic pivot with active or stative clause type and which arguments are expressed overtly, is complex and requires numerous examples. This listing will attempt to be comprehensive.
First, let us look at intransitive clauses. When the argument is a fully expressed nominal, everything is easy: the verb agrees with the nominal and its case, absolutive or nominative, determines whether the clause is stative or active respectively.
active, intransitive, full argument
The man ate.
stative, intransitive, full argument
The man was injured.
When the intransitive clause's single argument is not expressed - i.e. it's a dropped pronoun - the verb stands as a whole clause unto itself, its case agreement the sole marker of active vs. stative.
active, intransitive, full argument
stative, intransitive, full argument
He was injured.
Now let us look at the rather more complex case of transitive clauses. In the simplest case, we have two full arguments; the verb agrees with the first, which is the syntactic pivot. Whether the pivot is nominative or absolutive determines whether the clause is active or stative.
active, transitive, two full arguments
q’omna lik’aefī isaf
gore-PFCT.3.SG.NOM NOM-stag ABS-man
The stag gored the man.
stative, transitive, two full arguments
q’omnam isaf lik’aefī
gore-PFCT.3.SG.ABS ABS-man NOM-stag
The man was gored by the stag.
Note that in transitive clauses, if one argument is expressed only by the verb suffix then word order no longer helps to mark the active-stative distinction, which now relies entirely on the verb suffix. If the suffix agrees with a nominative but unexpressed pivot, and the expressed argument is in the absolutive case, the clause is active; if the other way around, stative.
active, transitive, unexpressed pivot
You gored the stag
stative, transitive, unexpressed pivot
You were gored by the stag.
Only the pivot can go unexpressed in this regard. If the pivot needs to be a full nominal, the other argument cannot be indicated by the verb and requires a pronoun for expression; this situation works exactly as with transitive clauses having two full arguments, save that pronouns do not vary for case.
active, transitive, two full arguments
q’omna lik’aefī lito
gore-PFCT.3.SG.NOM NOM-stag 2.SG
"the stag gored you."
Similarly, when neither argument is a full noun phrase, the situation is exactly as for other transitive clauses with unexpressed pivots.
stative, transitive, no full arguments
You were gored by it.
At this juncture it should be clear that the verb suffixes we here describe as "agreeing" with the nominative or absolutive case of the pivot, could just as well be said to directly mark active or stative clause type - the two always coincide, and in this grammatical sketch we have even defined each in terms of the other.
The Auxiliary Construction
While the required category of aspect is inflected for directly on the Tlaliolz verb, the additional distinctions of tense and modality are offered by the auxiliary verb construction. There are three auxiliaries, each of which is simply another verb that, as an auxiliary, is placed before the main verb and left uninflected, with any final vowel dropped.
|AUX||full verb||auxiliary meaning||example|
|ōtl||ōtla, "stop"||past tense||Ōtl k’ānna. - He ate.|
|non||non, "go"||future tense||Non k’ānna. - He will eat.|
|pāz||pāzi, "have"||modal||Pāz k’ānna. - He should eat.|
It is of course possible to use the same verb as both an auxiliary and a main verb in the same clause:
He will go.
- The past and future tenses are optional categories in Tlaliolz, used typically only when it might otherwise be unclear if the clause refers to the past or future. Most often this is already clear enough through context.
- The modal is a required category in Tlaliolz, in that without it, a clause is always realis. The modal simply indicates that the situation expressed by the clause is probable, or to be desired or expected, or depends on the truth of another clause. Tlaliolz makes no morphological distinction between any subtypes of modality.
Conditional situations require no extra morphology or particles, and can be expressed simply by juxtaposition of two modal clauses - the first is the "if" clause, the second is the "then" clause. The word qoa, "therefore", is required only if the two modal clauses are arranged differently from the basic conditional construction, and is placed before the "then"-clause.
Pāz k’ānna lia, pāz eanna.
If he eats it, he dies.
Qoa pāz eanna, pāz k’anna lia.
He'll die if he eats it.
Various types of complex clauses are documented in Tlaliolz, which all share the common trait that the complete construction includes more than one verb or verb root, not counting auxiliaries. As with all languages, the degree of grammatical integration between the verbs (and their associated clausal material) is tighter in some constructions than in others. The tightest integration of two verbs in Tlaliolz is the compound verb construction; less tightly integrated are subclauses, relative clauses, and serial clauses.
Many languages employ compound verbs as a lexical process; two verb roots are compounded together to form a new word stem. In Tlaliolz however, this process is more grammatically productive. Semantically resembling the serial verb constructions of other languages, a Tlaliolz compound verb consists of multiple verb roots compounded together, each expressing one facet of a more complex situation. But unlike prototypical serial verbs, these verb roots are morphologically bound, and share all core arguments and other clause material in common.
The grammatical mechanics of compound verbs are distinct from the formation of prototypical lexical compounds. The most prominent way is that the order in which verb roots appear in compounds is typically chronological, rather than dependent on a head-modifier relationship. Indeed, none of the roots in a Tlaliolz compound verb can properly said to be modifying or restricting another. In addition, though many compounds have become fixed words, many others are constructed only at the time of use:
Qūtlq’omna lik’aefī lito.
run-gore-PFCT.3.SG.NOM NOM-stag 2.SG
"The stag ran-up-and-gored you."
Often, situations expressed with single verbs in other languages will be broken into multiple semantic chunks in Tlaliolz, expressed by a compound verb. For example, the verb aqe "to move, walk, go" is intransitive only, and cannot be used for a situation in which a person moves an object to a new location. Instead, the compound verb take-put is employed:
"I'm moving the kettle."
Such compound verbs can be used to cover a wide range of situations that in other languages might require modifying adverbs, prepositional phrases, or adverbial clauses:
"I pushed it into place."
"I gave it away."
They are also not limited to two roots per compound; three and even four are attested, if with less frequency than two:
"I went over and stabbed it to death."
Subordinate clauses are embedded clauses serving to complete the main clause or add extra information to it. Tlaliolz does not distinguish morphologically between "complement" and "adverbial" clauses, which tend to be formed separately in other Talo-Edastean languages. Instead, a subordinate clause's function is determined by context and its placement. If it is located before the rest of the main clause's material, it is functioning adverbially; if instead it is located medially in the main clause, it is an argument of the verb. If it comes at the end, context is usually sufficient to disambiguate the functions. Subordinate clauses are formed by prefixing zi- to the verb; all other elements of the subclause appear exactly as they would if the clause were independent. That is to say, Tlaliolz subclauses are finite.
Zisayatak lēts, yayofak liFāzzin.
SUB-hold-IMPFCT.3.PL.NOM ABS-sword, laugh-PFCT.3.PL.NOM NOM-Fáralo
"Holding swords, the Fáralo laughed."
"I heard your yell."
Prototypical coordinate clauses are not a well-installed feature of Tlaliolz, though they are attested to occur. Taking their place are the much more frequently occurring serial clauses, which are distinct from coordinate clauses in that 1. no conjunction is used to join them, they are simply concatenated like in clause-chaining languages; and 2. all clauses in the series must share the same syntactic pivot. On the other hand, serial clauses are not wholly like a protoypical clause-chaining system either, in that each clause of the series is formed as an independent clause and could stand as such if the others in the series were removed -- each verb is inflected for participant reference in the usual way, and there is nothing resembling a switch-reference system. One way to view all of this is that Tlaliolz serial clauses lie somewhere in the middle of the continuum between clause chaining and clause coordination.
The first clause in the series sets the syntactic pivot, and subsequent clauses must match it:
Q’omnam Utsintā lik’aefī, eannam.
gore-PFCT.3.SG.ABS NAME NOM-stag, die-PFCT.3.SG.ABS
"Utsintā was gored by the stag, and he died."
However, the clauses in a series can be a mix of active and stative types, with the shared pivot thus switching between nominative and absolutive cases:
Q’omnam Utsintā lik’aefī, fiekʷina.
gore-PFCT.3.SG.ABS ABS-NAME NOM-stag, scream-PFCT.3.SG.NOM
"Utsintā was gored by the stag, and he screamed."'
Note that there is some overlap in the use of serial clauses with the use of compound verbs - particularly since roots in a compound verb cannot be independently modified. The three-verb compound example given above, "Aqq’omēq’uni lia.", might just as often be expressed in serial clauses, like this:
Aqeni, q’omni lia, ēq’uni lia.
go-PFCT.1.SG.NOM, stab-PFCT.1.SG.NOM 3.SG, kill-PFCT.1.SG.NOM 3.SG
"I went over and stabbed it to death." (literally: "I went, I stabbed it, I killed it.")
It is notable that in one style of traditional Tlaliolz storytelling, an entire tale is sometimes constructed as one very long clause series, with the hero of the tale their shared syntactic pivot.