Count nouns can be quantified by the following suffixes:
The paucal denotes a few things (or ’some, a couple of things’) while the greater plural denotes many things, e.g. táljo ’few words, a couple of words’, tálmi ’many words’. Pluralia tantum or commonly pluralized words like ’mountains’ normally use either of these two in Talmit:
The suffix -mai denotes a group or collection of count nouns:
Compared to the words derived with -mi above, these are used to mark a contrast to another group, cf. English people vs. a/the people; or German Berge vs. collective Gebirge. For man-made objects, they also suggest a purposeful arrangement, hence dágmai, dáχtamai ’wall’ from dat, dáχta ’stone’, while dágmi is just ’many stones’.
There are analagous endings for mass nouns, but the term ’plural’ cannot be applied here, therefore I use:
The mollitive denotes a small volume, the suffective a sufficient volume, while the intensive denotes a large volume. In addition, the mollitve suffix causes e-umlaut (1.1.3):
The following sound changes in combination with -mne are found:
States can be measured on a continuous scale as well, their scale is just not limited to volume. Therefore they take the same suffixes as mass nouns. The mollitive denotes a small value on the given scale, the suffective a sufficient value, while the intensive denotes a large value, e.g.:
The final unstressed -e of the endings -we, -mne was at some point lax in pronunciation, merging with *ə. It then caused the evening-out of vowels (1.2.3), affecting preceding a, i > e and u > o. Hence one would expect to find both tréwe, trémne < *trawə, *tramnə. However, in the standard dialect, the original vowels were restored by analogy to the unsuffixed words, except for the mollitive form. The reason for that has probably to be sought in the sound-symbolic association of i with higher intensity and of a with larger size (see sound-symbolism: http://sindanoorie.net/glp/phonosymb.php).
The notion of an extreme quantity is expressed by adding -t to -mne and -we, e.g.:
With count nouns, this forms a collective plural, denoting all things of a kind in existence, or just in a specific context, e.g. médamit ’human beings as a species’. Hence also the name of the language: tálmit means ’all the words and labels’, colloquially ’language’.
A further emphasis can be given by repeating the suffix as a prefix, e.g.:
Note that a separate word ’very’ is not required in Talmit. With the collective plural, reduplication further emphasizes the completeness:
Size is a natural quantity of any existing object, and separate endings are found for it:
They are frequently used in word derivation, e.g.:
Colloquially, there is some confusion of the endings, so that -we and -mne are used instead, clearly by influence from the mass nouns (hence dékwe ’pebble’, dágne ’boulder’), but this should be strictly avoided if one wants to get an official proficiency certificate in Talmit (not that there is one).
For all non-permanent states, the natural quantity is duration, and so the same endings refer to time:
Zero quantity can be expressed by prefixing mul- (mu- before roots containing l) to both states and nouns, e.g.:
Note that words may belong to both count and mass noun categories, e.g. kju ’part, division’ as a count noun kjúmi ’many parts’, kjúmit ’all parts, the whole’; but as a mass noun: kjúmne ’large part’, kjúmnet ’majority’, kjówet ’minority’. Some words may even be regared either as states or as nouns, e.g. pal as a count noun: ’body/fixed volume of water’ (pálmi, páljo), as a mass noun: ’water as a substance’ (pálne, pélwe) and as a state: ’liquid state’ (immeasurable).
To specify which one is meant, it is common to affix enclitic ta (count noun), ha (mass noun), pa (state):
Tat ’point in time’ becomes táχta ’time in counting (German mal, French fois)’, hence tágmi, táχtami ’many time points’, tágmi-nóllo, táχtami-nóllo ’often’ tákjo, táχtajo ’few time points’, tákjo-nóllo, táχtajo-nóllo ’seldom’. Or it becomes táppa, now grammaticalized as the temporal conjunction ’when, as’:
Note that stone, for example, can be perceived to exist in discrete packages (like stones by the wayside) or as a continuous mass of varying volume (like mountains) and therefore the word dat will be used to describe both. But if one takes something like fruit (kwásta < √kwats), it is only perceived to exist in countable units. In English, ’fruit’ can be a mass noun: ’there is some fruit on the table’, but never so in Talmit – one would simply use paucal kwástajo instead.
Another enclitic modifier is da essentially meaning ’intelligent being’. The commonest usage is its combination with the root √me ’human’ to derive the proper word méda ’human being’ (médata ’individual’, médapa ’state of being human’). Otherwise it can be used in a religious context for the means of personification: kázda and pálda would be ’fire spirit’ and ’water spirit’ respectively, for example.
Finally, professions may be either regarded as states, or may denote people practicing them. The difference can be again made apparent by using -pa or -ta, e.g. pérax, péraxpa ’state of being king’, pérax, péraxta ’a person who is king’.
As mentioned in the introduction, states may have signum, i.e. a positive and a negative part. The former is usually expressed by the vowel a and the latter by i (ι before w, j), more rarely by o and u respectively (mostly for assimilative/dissimilative reasons). With PSV(S) roots, this is usually an infix, e.g. dlon > dalón ’awake state’, dilón ’asleep state’. With PV(S) roots, a prefix is mostly found, e.g. aχál ’alive state’, iχál ’dead state’; although suffixed -a, -i and other idiosyncracies occasionally occur, as okára ’state of good luck’, ikáru ’state of bad luck’ < √kar.
For PSVS-roots compound forms according to the pattern dallo-, dillo- < *dalno-, *dilno-, parzu-, pirzu- < *parsu-, *pirsu- (shifted to Pa/iSSV-), e.g. dillogús ’dream’ are common.
The positive state is, if possible, chosen to be the one beneficial with respect to the speaker; the negative state as harmful, e.g. apéa ’good, proper state’, ipéa ’bad, improper state’, halán ’(morally) good state’, hurán ’bad, evil state’ (for the alternation l|r see 2.6). Therefore, the latter can be often translated by a negative prefix ’un-, in-, dis-’. But often enough, the distribution between a- and i- is purely idiomatic and conventional.
Many states are measured with a signum and vary on a scale on top of it. For instance, vertical position can be measured above or below the speaker (or some point of reference). The positive state is again expressed by a or o in form of a prefix, infix or suffix; and the negative state by i or u. It can be combined with -mne and -we:
Note that mutlé in the latter case would be ’neither cold nor warm’, i.e. ’room temperature’ in the case of weather, which is zero on this given scale; and similarly muϕál ’twilight state’, mulpéa ’neither good nor bad, average/neutral state’ (mulpéa-epéamis-nójo ’comme ci, comme ça’).
All the mentioned states were used on an absolute scale, given by the context. To compare two states, one has to go to a relative scale. It can be done by using e in the same way one uses signum markers (2.1.3) which are not required anymore. Hence perús ’relative vertical position’, telé ’relative temperature’, eϕál ’relative brightness’. The comparative is then expressed by using the postposition láha of the comparative case (2.2) and -mne/-we, which now express the idea of larger/smaller:
Emphasis can be added by reduplicating -mne/-we: A-jar B-láha emneϕálne-nójo ’A is much brighter than B’.
To say that the two values are equal, one uses the suffix -mis (originally meaning ’average value’, etymologically ’middle’) or -nes (an ancient element meaning ’copy’ for nouns, ’repetition’ for verbs) on the relative scale:
In particular, ϕe ’degree of similarity’ is always used on a relative scale (eϕéwe ’dissimilar state’, eϕémis ’same, equal state’) to express an unspecific comparison. Obviously, eϕémne makes no sense, unless one wants to say something like:
An interval or difference between two values on the scale of states can be expressed by prefixing an- or na- (or on-, no-). Sometimes, the interval is also measured with signum (2.1.3) and the negative state is then formed by in-, ni- (or un-, nu-). For example, while prus means ’vertical position’, amprús or naprús means ’difference between two points on the vertical axis’ or simply ’height’. Accordingly, imprús or niprús means ’depth’, measured below the reference point of the speaker.
This was originally formed with the Proto-Tallic element *ṇ- which conveyed the idea of extension in space or time in Proto-Tallic (hence *ṇprus without signum). It was later broken to am-|an-, na- or im-|in-, ni- due to influence of the a/i-classification of the signum states. Forms like *mprus still appear medially in compounds, e.g. kaontát ’year, full cycle’ < káo ’cycle, period’ + *ntat ’time span’.
Intervals of states are still continuous states and can therefore be qualified by -we and -mne, e.g. amprúzne ’great height’, amprózwe ’small height’, imprózwe ’small depth’, imprúzne ’great depth’; or used on a relative scale: amperús ’relative height’, imperús ’relative depth’. But they can also be treated as count nouns, e.g. antát ’time interval, period’, antákjo ’few periods’, antágmi ’many periods’.
An interval of count nouns is the ordinary plural, e.g. antál, natál ’words’, amméda, naméda ’human beings’ and so on. However, it is used much less than in English – it might be translated with ’some, an amount of’ and is mostly used in contexts where the speaker cannot vouch for the amount of the mentioned things (almost like evidentiality). Generalized statements like ’Sheep eat flowers with thorns’ prefer the singular instead – ’Sheep eats flower with thorn’.
In Proto-Tallic, the ordinary plural was actually formed by *-mi, which has since then shifted its meaning to the greater plural (2.1.1) in Talmit, but remains so in Kymna.
Note that before hV- a metathesis *ŋχ > χn takes place, hence héper ’animal’, pl. aχnéper. The variants with an open syllable na-/no-, ni-/nu- are preferred before hS-.
Intervals as count nouns take the prefix anna-/anni- when put into the plural, e.g. annatát ’time intervals, periods’.
Prefixing en- or ne- to count nouns marks a relative scale where amounts can be compared. ’More’ is then expressed by -mi, ’fewer’ by -jo and ’as many as’ by -mis, e.g. emmédami ’more people’, empekwórjo ’fewer flowers’, enkaontágmis ’as many years as’.
Count nouns with signum (2.1.3) are mostly the dual body parts where the right or upper side is positive, e.g.:
The dual is formed by superposing both vowels into a diphthong (a crasis of sorts):
For states this similarly forms a dvandva compound:
Aiχál would theoretically mean ’both dead and alive’ which might be useful in limited contexts...
... but not in too many, so that in such a case the meaning shifts to a disjunction: aiχál ’dead or alive’.
A reduplication of the whole word forms a greater plural for count nouns (similar to Japanese) and carries the additional shade of meaning ’various sorts of’:
Polysyllables reduplicate just the first syllable without its sonant:
Initial spirants of polysyllables turn into unvoiced stops by dissimilation: paϕámreta = ϕamretámi ’many clouds’ (cf. Grassmann’s law in Greek, e.g. perfect active φεύγω> πέφευγα). By analogy this has led to a spirantization of initial stops, so that
are alternative forms.
For states, reduplication conveys the idea of a reserved judgement, the signum marker is used only once:
Some reduplicated words are colloquial names for plants and animals (or used by children), e.g. anderdergáme ’elephant’ (’long-nose’), kinkinkátu ’oak’ (’hard-tree’).
Reduplication of the numeral at, a ’one’ gives *aQa(Q) > áha(t), éa(t) ’pair, both’. This is also used as a prefix ’bi-, di-’, e.g. ahambá, ahabá, eabá, eambá ’biped’. Reduplications of other numerals also appear:
They are usually used where two groups of two/three/four are denoted, e.g.:
A threefold repetition of at gives *aQ(a)Qa(Q) > †áχχa(t), †éahat ’many, a large amount’, an archaic/poetic equivalent to dámme < dan ’number’.
Apart from duality, another important symmetry in Talmit is pentality, that is a grouping in sets of five. It manifests itself in five initial consonant series, five root-final sonorants (at least by a naive analysis) and five vowels. Originally, the numerals themselves seem to have been a quinary, i.e. a base-5 system, and were formed with the five vowels and root-final sonorants:
This system was at some time expanded to a biquinary one, with the numerals 6-10 formed by: modified vowel + nasal n/m + numeral 1-5; and stress on the ultimate syllable:
The initial vowel was originally *ə, later assimilating to the following one according to the evening-out rule (1.2.3). However, unún is probably a reduplicated dual form. Colloquial variants are mat, nil, mes, nor/mor, mun.
Higher numerals are formed by combination: unún es ’13’, enór unún at ’91’ and so on. Alternatively enclitic -mun can be used as a decimal suffix: ágmun ’10’ (rare), ílmun ’20’, ézmun ’30’, órmun ’40’, úmmun ’50’, amágmun ’60’, ιnílmun ’70’, enórmun ’90’.
Furthermore, augmentative -men appended to ’5’ yields †úrmen ’25’ which is archaic; and appended to ’10’ it yields unúrmen ’100’ (cf. French million < mille ’1000’).
A special case outside these patterns is mut ’0’, a variant of the root √mul∼√muj∼√muq ’nothing’.
Numerals precede nouns without any particles and the monosyllabic ones up to five are compounded. The singular is used with all numbers: agméda ’one person’, ilméda ’two persons’, ezméda ’three persons’ and so on.
Cardinals are formed with ordinal + dan ’number, amount’ + gen. mo: aχdán-mo ’1st’ (Jap. ichiban no), ildán-mo ’2nd’ and so on.
Even though natural numbers are discrete, there is an associated scale, and so dan is a mass noun when varying on it: dámme/dénwe ’a large/small number, high/small amount’; otherwise: dámmi/dánjo ’many/few numbers’.
Finally, one can combine numerals with the greater plural -mi (often in form -mni after n by influence of the intensive 2.1.1) and the paucal -jo to express a judgement as to whether this quantity is large or small, e.g. unúrmemni ’100 which is a lot’, unúrmenjo ’just 100’. Hence for example the saying agmédajo, ilmédami ’one-person-pau, two-person-gpl’ meaning that two people can accomplish much more than one.
Small fractions are formed from the root √iq, literally meaning ’cut’, high fractions with the suffix -ki attached to the numeral:
Compare the verb íxun ’cut in two equal parts’, ipíxun ’cut in three equal parts’. Analogy to the latter word explains ípit instead of expected *íkit. It is unclear whether the suffix -ki comes from √iq in the reversed form *√qi, or is a reduced form of kju ’part’.
While iQ- as a prefix means ’half the amount of what is expected’, e.g. ίize-láha tárme-ippá-nójo ’I’m half the man I used to be’ (lit. ’Compared to my past state, [I’m] in a half-man state’), as a suffix it means ’one half out of two’ and has thus naturally come to mean ’(an)other’, e.g. méhit, médahit ’another person’, tá(h)it, táχta(h)it ’another point in time’ and so on.
To emphasize indefiniteness many languages use a construction involving the numeral ’one’, and of course, the indefinite article is very often derived from it. Talmit, although it uses grammaticalized numerals in various constructions, shows no such tendency. The indefiniteness of an object can be stressed by using the attribute agó-mo, lit. ’of existence’ (e.g. agó-mo médata ’a certain individual’), the indefiniteness of a point in time can be stressed by using tát-mo, lit. ’of time’ (e.g. tat-mo aϕéx-nóllo ’on a certain day, once upon a time’).