This monograph provides a detailed grammatical description of a language called Tokana. Tokana is a 'model language' - that is, an invented language (or the outlines of one), constructed for personal amusement and edification. I have been working on Tokana for about six years now, and this sketch represents my attempt to present it in a form which is accessible to a larger audience.
I should begin by mentioning that Tokana was not designed to be a 'philosophical' or 'logical' language, or a vehicle for international communication like Esperanto and other such projects. Instead, it is intended as a purely personal artistic endeavour (some would say a self-indulgent one, but then all art is to some degree self-indulgent, whether it appeals to a mass audience or a select few). Tokana represents my attempts to explore certain aesthetic impulses and theoretical ideas, and to construct a relatively complete linguistic system which is (I hope) believable, internally consistent, and interesting in its details.
My goal throughout the project has been to keep things 'naturalistic'. That is, I have tried to adhere as closely as possible to what we know - or what I think we know - about how naturally evolving human languages work. Thus Tokana is intended to look and feel like a natural language, with many of the complexities and (seemingly) arbitrary features which such systems tend to have. Although I constructed the grammar of Tokana from the ground up, with the idea that it be a unique creation which is not obviously derived from (or related to) any single living language, I have nevertheless borrowed many phonological and syntactic features from existing sources. Among those languages which have contributed to my creation of Tokana are Basque, Choctaw, Cree, Finnish, French, German, Greek, Hebrew, Hindi, Hungarian, Irish, Kwakwala (Kwakiutl), Lakhota, Malagasy, Quechua, Spanish, Tagalog, Warlpiri, and Welsh.
Like all natural languages, Tokana exists within a cultural and historical context - albeit a somewhat sketchy one. Tokana is (or was, or would be) spoken by an imagined people, also called the Tokana, who live somewhere in northwestern North America, but in a time quite remote from our own - either the distant past or the far future, or perhaps some alternate history. The fictional author of this monograph is a fieldworker, a linguist or anthropologist, who has chosen to present his reasearch in the form of a reference grammar for non-linguists. In keeping with this approach, I have tried to be concise yet compre-hensive, illustrating each point with example sentences. I will assume that the reader is familiar with basic descriptive terms of traditional grammar (such as subject, object, preposition, embedded clause). Beyond this, I have tried to avoid technical terminology wherever possible, and to define - or at least illustrate - those terms which I do use.
Note finally that Tokana is an ongoing project, with no fixed endpoint. It has evolved considerably during the time I have worked on it, and will no doubt continue to evolve as long as I remain interested in it. The grammatical description presented here is thus not the final word on Tokana, but merely represents the state of the language as it exists now. Many changes and corrections have been made to this, the first web version of the monograph. Subsequent modifications, if any, will be reflected in future versions.
Many thanks to my illustrious predecessors in what J. R. R. Tolkien called "the Secret Vice". Thanks also to my many talented contemporaries (most of them members of the CONLANG and LANGDEV email lists) for providing me with much needed support and inspiration, and for giving me a forum to present this strange and very personal work of art.
Tokana is the language of the Tokana people, and has approximately 35,000 native speakers. The Tokana live along the coast of Nala, which appears to correspond to northwestern North America in our universe - a fertile, temperate region covered with grassy hills and coniferous rainforests, surrounding a large inlet called Sutu Tohmi.
Tokana is not related to any of the languages spoken by neighbouring peoples. Its closest relatives appear to be the Kman Group, a family of about two dozen languages spoken in the Sakwot region north and east of Nala. Archaeological and linguistic evidence suggest that the ancestors of the Tokana migrated southward from Sakwot about two thousand years ago, as part of a general dispersion of Proto-Kman-speaking groups down the coast. The first Tokana settlers arrived in Nala about eighteen hundred years ago.
The Tokana live mostly in small villages consisting of between three and six compounds or homesteads called patha, each of which is the ancestral home of a kame, or matrilineal clan. (The towns of Tenmothai, Uiluma, and Kemothasi are rather larger, consisting of several dozen patha, as well as markets, ceremonial centres, and public buildings of various kinds.) The Tokana are a non-industrial people whose principal economies are farming, fishing, sheepherding, and timber. They are also skilled artisans, who specialise in woodworking and textiles. The Tokana participate actively in the Ymuthmol Ilal, a vast trading network extending for many hundreds of miles up and down the coastal ranges of North America (other groups involved in this network include the Hbana, the Tzesek, the Auari, the Awamai, and the Kapakwilo).
There are about half a dozen dialects of Tokana, which differ from each other in details of pronunciation and vocabulary, but share essentially the same grammatical features. The major dialectal break occurs between the so-called Coastal and Inland dialect groups: These are easily distinguishable by the fact that the Coastal dialects have lost the phoneme f, replacing it in most cases with h (e.g. "receive" is fimoita or fimuita in the Inland dialects and himoita in the Coastal dialects). In this sketch I have chosen to represent the speech of Tenmothai, the largest Tokana-speaking settlement. This dialect is typical of the Inland dialects spoken along the eastern shores of Sutu Tohmi.
This monograph is divided into five chapters: Chapter 1 deals with the phonology of Tokana. Chapters 2 and 3 deal with the morphosyntax of noun phrases and verb phrases, respectively. In chapter 4 I discuss prepositional phrases, conditional clauses, and other adjuncts, as well as conjunctions and the formation of complex sentences. In chapter 5 I deal with word order in Tokana, and also discuss various sentential particles (that is, words which have a fixed position in the sentence).
Throughout this monograph, Tokana words (or parts of words) are highlighted, while the corresponding English glosses are set off by double quotes, e.g. halma "book". Italics are occasionally used when citing words from English and other languages.
In this chapter I discuss the sound patterns of Tokana. In 1.1 I present the consonant and vowel sounds of Tokana, with notes on their pronunciation. In 1.2 I discuss syllabification and stress assignment. And in 1.3 I list some common sound changes which occur in the language.
A note on the orthography: Tokana is normally written in a syllabic script called ilo. However, in this monograph, Tokana will be represented in transcription using the following fifteen letters of the Roman alphabet: a, e, f, h, i, k, l, m, n, o, p, s, t, u, y. With the exception of three digraphs (lh, th, and ts), each sound is represented by a single letter, and each letter represents a single sound. In addition to the basic symbols, a vowel diacritic (`) is used to indicate stress in certain words, e.g. napè "daughter".
Consonants: The consonant sounds of Tokana are given in the following table (recall that the Coastal dialects lack the labiodental fricative f, replacing it in most cases with the glottal fricative h; e.g., fana "be unsure" is pronounced hana in the Coastal dialects):
| labial | dental | alveolar | postalv. | velar | glottal | |
| oral stops | p | th | t | k | ||
| nasal stops | m | n | ||||
| affricates | ts | |||||
| fricatives | f | s | lh | h | ||
| glides | l |
Of these, p, t, k, m, n, ts, f, s, h, and l have their normal phonetic (IPA) values. Note that the oral stops are unaspirated in all positions, as in Spanish: Thus p, t, and k are pronounced as in spill, still, and skill, not as in pill, till, and kill (where the stop is accompanied by a puff of air).
The digraph th represents a laminodental stop, produced at roughly the same point of articulation as English th. This stop is usually unaspirated like p, t, and k, except before another consonant, when it is released with a certain amount of aspiration or affrication (e.g. the th in uthma "give" and athpa "make music" sounds quite a bit like the t + th combination at the end of the word eighth).
The digraph lh represents a postalveolar lateral fricative. It is quite similar to the lateral fricative of Welsh, Zulu, and other languages, except that it is produced with the body of the tongue instead of the blade or tip of the tongue (the point of articulation for lh is actually quite close to that of English sh).
Note also that the symbol n, in addition to representing the alveolar nasal (as in English net), also represents an assimilated nasal before a non-labial consonant. Thus nk (e.g. tunku "pain") is pronounced as in English think, while nth (e.g. pekunthe "last") is pronounced somewhat as in English tenth.
Vowels: Tokana has a six vowel system, as shown in the following table:
| front | central | back | |
| high | i | u | |
| mid | e | y | o |
| low | a |
Of these, i, e, a, o, and u have their standard 'European' values - i.e. approximately as in feet, fetch, father, for, and food, respectively. The central vowel y is similar to the sound in English but, but with the tongue positioned slightly higher and farther back. In closed syllables (i.e. before a syllable-final consonant), i, u, and o are somewhat more 'lax' than in open syllables. Thus for instance the vowel in his "star" is about halfway between the vowels in hiss and he's. Similarly, the u in tuk "group" is halfway between the vowels in took and toot, and the o in sot "word" is halfway between the vowels in so and sought.
In addition, i and u are pronounced as glides, [j] and [w], when they occur adjacent to another vowel. Thus ie "with" is pronounced [je] (as in English yet), luan "hair" is pronounced [lwan], and ikei "dog" is pronounced [ikej] (rhymes with prey). The combinations iu and ui are pronounced [ju] and [wi], respectively; e.g. uima "love", pronounced [wima]. (Note also the diphthong yi, found in kyina "you (pl)". This is similar to the i sound in fight, as pronounced in Canadian English and many dialects of British and Irish English.)
For most speakers, the i glide causes palatalisation of a preceding consonant: Thus katia "house" is pronounced [kacja] (where [cj] is close to the t sound in future). Similarly, the s in mosie "shoulders" is like the sh in shell, except that it is produced without any rounding of the lips.
The inventory of possible syllables in Tokana may be represented in the form of a template. Except word-initially, syllables have the following structure (where C = consonant, V = vowel, G = an i or u glide, and parentheses indicate optionality):
(C)-(G)-V-(G)-(C)
That is, a syllable consists minimally of a single vowel, and maximally of an initial ('onset') consonant, a vowel flanked by glides, and a final ('coda') consonant or glide. Examples of words illustrating some of the more common syllable types are given below:
V a "oh!" CV ki "and" VC im "when/during" CVG lai "light" VGC aun "if" CGV lue "it is said" CVC tuk "group" GV ie "with" CVGC hauk "smoke" GVC iat "other" CGVC sian "breast" CGVG huoi "twelve"
Word-initially, a small number of consonant clusters are also attested, viz. kl, ks, ps, sl, st, sth, sk:
klohana "go through" sliahte "story, tale" ksohe "darkness" sten "deer" psyta "spit out" sthoka "destroy" skona "look at, watch"
Syllabification (the division of words into syllables) is fairly straightforward in Tokana. The following rules are observed for dividing up consonant and vowel sequences in the middles of words (where a period is used to indicate a syllable break):
(a) VCV syllabifies as V.CV: E.g. pama "top" is pronounced pa.ma.
(b) VGV syllabifies as V.GV: E.g. paua "wash" is pronounced pa.ua.
(c) VGC syllabifies as VG.C: E.g. huaita "like" is pronounced huai.ta.
(d) VCCV syllabifies as VC.CV: E.g. mukta "close" is pronounced muk.ta, and misla "turn" is pronounced mis.la.
Stress is assigned according to the following rules:
(a) If the word ends in a consonant and/or a glide, main stress falls on the final syllable (e.g. totsát "table", mináp "marrow", kamáit "for them", ilohfói "tomorrow", kumelón "slope of a hill").
(b) If the word ends in a (non-glide) vowel, main stress falls on the penultimate syllable (e.g. téne "steep hill", itskána "arrive", inlótka "yesterday", kékua "each", mósie "shoulders").
Note that stress is a property of (phonological) words rather than stems. Adding suffixes to a stem usually causes main stress to shift rightward:
hiéla "see" hielúha "want to see" hieluhóti "not want to see" hieluhotís "not want to see them"
The above stress rules apply to monosyllables as well as polysyllabic words: If a monosyllabic word ends in a consonant or a glide (e.g. lai "light", elh "and") then it receives its own stress; otherwise it is inherently unstressed, and cliticises onto - i.e. forms a phonological unit with - a following stress-bearing word.
The only (apparent) exceptions to the above stress rules are a handful of stems and inflectional forms which end in a non-glide vowel but nevertheless have their main stress on the final syllable. In the orthography used here, these forms are all marked with a diacritic (`) over the final vowel:
kà "worth, value" napè "daughter" ipalà "herb, medicinal plant"
Certain pairs of monosyllabic words are distinguished solely by the presence or absence of underlying stress (together with a slight lengthening of the vowel in the case of the stressed forms):
kì "time" sù "rain" ki "and" su "or"
Stressed final vowels have two sources in the phonology of Tokana:
(a) Noun stems which end in h drop the h word-finally, but retain their word-final stress as if they still ended in a consonant. Since h is only dropped word-finally, we get alternations of the following sort:
underlying form surface form ipalah > ipalà "herbs" ipalah + -ne > ipalahne "herbs-Inst" ipalah + -ma > ipalahma "my herbs"
There is also a suffix -h, which marks third
person singular inanimate
objects on the verb (see section 2.1.2).
As with the stem-final
lima > lima "open" lima + -h > limà "open it" lima + -te > limate "open them"
(Note that in the conservative Outikfe dialect, spoken in the hills to the east of Sutu Tohmi, h is retained in word-final position. Hence, in this dialect, the word for "herbs" is pronounced ipalah even when there is no suffix attached to it.)
(b) In some inflected forms, two adjacent vowels fuse to become a single stressed vowel. For example, the dative noun suffix -e fuses with a stem-final e to form è (see section 2.2.4):
tene > tene "hill" tene + -e > tenè "hill-Dat"
The following are some common sound changes which apply when an affix attaches to a stem. I list them here for reference; additional examples are given in the sections on clitic determiners (2.1.2), noun morphology (2.2), verb morphology (3.1-3.6), et cetera.
(1) h-metathesis: When a morpheme (i.e. a root or affix) which begins with an h follows a morpheme ending in a consonant (C), the h and the preceding consonant undergo metathesis (i.e. inversion of segments) to become hC. Consider the following examples, where a prefix ending in a consonant attaches to a verb beginning with h:
an- + halhka > ahnalhka "be dry" (Rel) uk- + hiela > uhkiela "see each other"
Here n + h becomes hn and k + h becomes hk, respectively.
(2) Degemination: Geminate (double) nasals, mm and nn, are found both morpheme-internally and across morpheme boundaries (e.g. amme "mother"; lon + -ne > lonne "hill-Inst"). However, no other geminate consonants are allowed in Tokana. When two identical non-nasal consonants come together at a morpheme boundary, the first one becomes h, creating an hC cluster (just as with h-metathesis above). Examples:
mok + -ko > mohko "your home" uk- + kuola > uhkuola "meet each other" yities + -sa > yitiehsa "their towers"
Here k + k becomes hk and s + s becomes hs, respectively. Note also the following example:
alats + -sa > alahtsa "their birch trees"
Here ts + s becomes hts. The combination *tss is presumably ruled out because ts and s are too similar phonetically, even though the phonology treats them as separate sounds.
Note that l, lh, and h are exceptions to the degemination rule: l + l becomes l, lh + lh becomes lh, and h + h becomes h.
(3) Nasal assimilation: When a morpheme ending in a nasal is followed by a morpheme beginning with a non-nasal consonant, the nasal will assimilate in place of articulation to the following non-nasal:
konom + -ko > kononko "your hammer" konom + -kima > kononkima "our hammer(s)" an- + kaila > ankaila "be hot" (Rel) an- + pata > ampata "be tall" (Rel) an- + fiha > amfiha "be young" (Rel)
As these examples show, m + k becomes nk (pronounced as in think), n + p becomes mp, n + f becomes mf, et cetera. Note that nasals do not assimilate before other nasals. Thus:
konom + -na > konomna "his/her hammer" an- + meiloksa > anmeiloksa "be sweet" (Rel)
(4) s to ts: When a morpheme beginning with the fricative s is preceded by a morpheme ending in l, lh, or a nasal, the s becomes ts:
pul + -sa > pultsa "their village" silh + -sa > silhtsa "their fingers" huan + -sa > huantsa "their mouths"
In the following example, both nasal assimilation and s > ts apply:
liunan- + -se > liunantse "that they are old"
(5) Vowel raising: Combinations of two consecutive non-high vowels (*ae, *oe, *oa, etc.) are disallowed in Tokana. When such combinations are created through the concatenation of two morphemes, the mid vowel will normally raise to become a high vowel. Consider the following examples, where the dative suffix -e is added to a noun stem ending in a non-high vowel:
kuna + -e > kunai "friend-Dat" talo + -e > taloi "king-Dat"
Here a + e becomes ai and o + e becomes oi. Note also the examples below, where the verb-forming suffix -a is added to a noun stem ending in a non-high vowel:
amme + -a > ammia "be [someone's] mother" talo + -a > talua "be king"
Here e + a becomes ia and o + a becomes ua.
(6) Vowel lowering: Sequences of two identical high vowels (*ii and *uu) are also ruled out. When such sequences are created by suffixation, etc., i + i becomes ie or ei, and u + u becomes uo or ou, as in the following examples:
lali- + -iha > lalieha "is playing" uosu + -u > uosou "smooth stone-Abl" pau- + -un > pauon "washed" pau- + -uha > pauoha "want to wash"