Tirkunan: A Romance Language

[ Static Lexicon | Scripted Lexicon ]

Inside this orange box I am writing as a conlanger. A blue box will be used when our hobby linguist from this universe reports on what information was sent from there about Tirkunan. If there is no special markup, you are reading the original text that the Institute for Parallel Histories (IPH) sent from there. The IPH themselves use light grey boxes to clarify language history in more detail.


Tirkunan is a Romance conlang derived from Vulgar Latin that I modelled to look like what I felt was a nice Romance language, whatever 'nice' may mean.

Due to my preference to create highly regular engelangs, simplicity and regularity is one design goal, but Tirkunan is not an engelang, nor an auxlang, nor a creole, but constructed like a plausible language from a separate Romance branch, probably in sprachbund with a few surrounding other Romance languages.

The morphological simplification is extensive, but considering the Germanic family, there is a wide rage of morphological complexity, from Africaans to Icelandic. So Tirkunan is meant inside the Romance family to be morphologically simple like Africaans, and even simpler. Tirkunan should still feel like a natural Romance language, and some irregulaties exist to avoid an odd feel. I also try to give examples for individual simplifications of morphology in existing Romance languages to indicate plausibility.

Another goal was to have a fully elaborated Grand Master Plan (which you can play with interactively) for the sound shifts instead of making up words in an ad-hoc manner. This will help me prevent inconsistencies and makes the lexicon more plausible.

The nouns, adjectives and verbs will be derived with a semi-automated set of rules from Vulgar Latin words, but some manual adjustments will be made if it feels more plausible that way. The sets of rules will try to produce something that looks plausibly Romance, without making the conhistorical development explicit. The conhistory is currently not my primary goal as long as the result looks plausibly Romance at first and hopefully second glance.

Tirkunan's location in the multiverse is in the parallel universe of Þrjótrunn somewhere on the Iberian Peninsula. The details are still missing from the IPH.

In summary, the design goals were the following.

  • Very simple morphology.
  • Plausibly a Romance language.
  • Isolating morphology, and isolating or agglutinative derivation.
  • A plausible phonological history of the language, i.e., a grand master plan by which words are regularly derived from Vulgar Latin.
  • An interesting vowel system. This was the initial momentum starting this conlang.
  • No palatalisation, except for very few examples.
  • Flavours of Spanish, Catalan, Italian, Sardinian, Sicilian.
  • Should not sound like French, Portuguese, or Romanian.
  • If Afrikaans can through away all(/most) morphology, so can Tirkunan.
  • Tendency to not reborrow words from Classical Latin, but to keep the Proto-Romance derivation.


To get the right sound I had in mind, I considered the following.

Like Logudorese (Sardinian), Tirkunan did not palatalise /k ɡ/ (nor other consonants) before [e i]. This is the reason why I first chose to use 'k' instead of 'c': it is naturally pronounced [k] before [e i] even when thinking Romance (the choice of orthography is a bit unstable in my head, because it looks non-Romance, but at least Walloon and Judeo-Espanyol do use 'k'). However, later I changed it back to 'c' because the look of 'k' just did not feel right.

Like Sicilian, Calabrese and other Italian languages, Tirkunan drops some initial unstressed vowels, most frequently /i/, e.g. mpiri < impērivm 'empire' (compare Sicilian/Calabrese 'mperu').

Like Catalan, Tirkunan does not fear consonantal endings on words, especially on nouns and particles. E.g. ciutat [kjuˈtat] < civitātem 'town' (compare Catalan 'ciutat'), cel [kɛl] < caelvm 'sky; heaven'.

Like many dialects of Catalan, Portuguese, and some Southern Italian dialects like Sicilian/Calabrese, reduces unstressed vowels into a three-vowel system /a i u/, e.g. the participle of solu [ˈsolʊ] is sulut [sʊˈlʊt].

Like Catalan, has new endings in /u/ from /v/ which resurfaces in derivation: nou 'nine' vs. nuvint 'ninety'.

Like Spanish, Tirkunan has five phonemic vowels: /i e a o u/. Some words show similar sound shifts: ombra < hominem 'man' (compare Spanish 'hombre').

Like Italian (and French), Tirkunan has quite a lot of elided vowels and consonants, and many function words end in a vowel.

Like in Modern Tuscan, Romanian, and other languages, the infinitive '-re' on verbs is dropped.

Similar to Sicilian (and to a more limited degree Catalan and Portuguese), many prepositions often drop their final consonants and even vowels, e.g. pi < per can become p' and cun < cvm can become cu.

Like some Catalan dialects, the preposition di to form the genitive is often dropped. The preposition is not dropped in other meanings or usages, like 'from': gen di Tali 'person from Italy' cannot be shortened.

Unlike French, there are no nasal vowels, and no [y ø ʒ ʃ ɥ].

Unlike Portuguese, Old French, Old Spanish, there are no nasal vowels and no [l] > [u] shift, although I like it very much and it was tempting to include it.

Unlike most Romance languages, there is almost no palatalisation, so no [ɲ ʎ ʃ ʒ t͡ʃ d͡ʒ]. Instead, Tirkunan retains the [i] and [u] glides, e.g. ali < alivm and coru < corvvs.

Unlike all existing Romance languages, Tirkunan has lost gender completely. Still, I do not think it is inconceivable in Romance, because Germanic has done it, and Romance and Germanic come from the same source. While Romance is relatively homogeneous, in Germanic, the completely different handling of gender (and case) between Icelandic and Afrikaans is estonishing, so why should there not be a hidden Romance language that lost gender altogether?

With the loss of endings, the number marking on nouns was also lost. This also does not feel implausible to me as French has only weak number marking on nouns, too.

Further, especially the verbal system is untypically analytical for Romance languages, although all the principles and forms are taken from existing Romance languages so none of the isolated verb forms should not appear out of place. I found that when Afrikaans can go so far in Germanic, then so can a Romance language.

I am positively surprised about how making Tirkunan isolating could be done by using selected features from existing Romance languages: analytic verb forms using auxiliaries, dropped endings in Catalan and Tuscan, ordinals with di+number like in Sardinian, replacement of possessive adjectives by di+pronoun in Balearic Catalan, causatives with fi auxilary like Portuguese, plural marking only on article (almost) like French. Sometimes, I found the existing languages with the simplified features only after accidentally doing it in a very similar way. E.g. the singular pronouns of Ligurian, plural marking in French, analytic possessives of Asturian.


Structurally, Tirkunan is more French than I had planned. Some problems were elegantly solved that way, and some things probably just happend by chance:

  • Plural is shown on articles, not on nouns: ciutat 'city/cities': li ciutat 'the city' vs. lur ciutat 'the cities'.
  • Drop of vowels, very short verb forms, qv>c: l'ombra 'the man', cimp 'five'.
  • Some vocab does sounds very French, but even more Catalan in most cases: Lac Liman 'Lake Geneva'.
  • Quite some vocab is actually identical to Catalan: neu 'snow'.
  • Some similarities with Romanian derivation. E.g., '-re' (old infinitive, now noun formation) has inspired -(a)r in Tirkunan for nominalisation, and '-tor,-toare' -(a)tur is also used for adjective formation.


The name 'Tircunan' or 'Tirkunan' means 'Tarragonian', i.e. 'language of (the city of) Tarragona'. This town is called 'Tracunis' in Modern Tirkunan and was the capital of the Roman province of Hispānia Tarracōnēnsis for several centuries. Today, it is the second largest city (~2.300.000 inhabitans) of Lustany (or 'Lustani' in Tirkunan).[1] In texts written before the unified orthography was introduced, Tirkunan is also frequently referred to as 'Trakunan', 'Tracunan', 'Terkunan', 'Tercunan', 'Tircunan', or 'Tirkunan'. Further, some dialects have metathesis (or used to have, as by the influence of the standard language this is no longer the case), in analogy with other words, so we also find 'Tricunan', etc.

Derived regularly from the Latin word, the city is called Tracunis. For the language, the ending -an was eventually preferred for adjectives instead of -is as the latter was more and more interpreted as a locative, today often found in city names.

The vowel change in the language name was either by influence of another language, or maybe because it was taken to be related to ter 'earth'. We don't know the exact reason. Historical documents merely show that the name changed around four-three hundred years ago, and there are various inconsistent spellings in older texts.


Phonology and Spelling


Spelling of Tirkunan is completely phonemic. All phonemes are represented by a single grapheme, i.e., there are no digraphs.

Stress and syllable boundaries are mostly implicit and not specifically marked, but the full stress pattern can be derived regularly from the spelling. The apostrophe is used to indicate a vowel or consonant that was dropped due to a elision, and at the end of a word, this apostrophe also marks that stress is retained on the last syllable.

In the following, the phonemic representation is usually not given except in special cases, because the spelling of Tirkunan is phonemic. When pronunciation is discussed, the phonetics are given in IPA.


The following consonants exist:

grapheme phoneme phones
b /b/ [b]
c /k/ [k]
d /d/ [d]
f /f/ [f]
g /ɡ/ [ɡ]
l /l/ [l]
m /m/ [m ɱ]
n /n/ [n ŋ m ɱ]
p /p/ [p]
r /r/ [r]
s /s/ [s]
t /t/ [t]
v /v/ [v]


Tirkunan allows any sequence of vowels anywhere in a word, and each vowel is counted phonemically as a separate syllable. Some vowel sequences tend to be phonetically realised as diphthongs, as the section on vowels explains. Also, sequences of u + vowel emerging from derivation are usually resolved as v + vowel.

The language allows the following initial consonants or internal clusters:

b bl br mb mbl mbr sb sbl sbr
p pl pr mp mpl mpr sp spl spr nsp nspl nspr
d - dr nd - ndr sd - sdr
t - tr nt - ntr st - str nst - nstr
g gl gr ng ngl ngr sg sgl sgr
k kl kr nk nkl nkr sk skl skr nsk nskl nskr
f fl fr mf mfl mfr sf sfl sfr
l - - nl - - sl
m - - - - - sm
n - - - - - sn
r - - nr - - sr
s - - ns
v vl vr mv mvl mvr sv svl svr

Inner clusters are usually simpler, but prefixes may cause more complex initial clusters, e.g., n(i) < in- and s(a) < -ex. The resulting combinations are already considered above -- many combinations starting with s, n, m are not found in a lemma in the lexicon.

Phonemically, all the clusters are single syllable, e.g. for determining stress, but phonetically a pre-consonantal n or m is pronounced syllabic.

If a consonant cluster is not feasible, the epenthetic vowel surfaces from the prefixes: this is before nasal for n(i) and before s for s(a). In the case of s, before voiced plosives, it causes the plosive to become phonetically unvoiced, e.g., sba [spa]. In sr, an epenthetic t may be heard, and in nr, an epenthetic d.

The language allows the following final consonant clusters:

p mp lp rp sp
t nt lt rt st
k nk lk rk sk
m - - rm
n - - rn
- ng
f - lf rf
s - - - - ts nts ps cs

The final ng represents the phoneme /ŋ/.

To avoid other final consonant clusters, an epenthetic a emerges. Some consonants usually do not occur in stems but are derived from Proto-Romance in different ways, e.g. final d usually appears as r and final b and v usually appear as u.

Vowels and Syllables

Tirkunan has the following vowels, which are pronounced slightly differently in different phonological context, determined by length and stress:

spelling phoneme
a /a/ [] [a] [ɐ]
e /e/ [] [ɛ] -
i /i/ [] [ɪ] [ɪ]
o /o/ [] [ɔ] -
u /u/ [] [ʊ] [ʊ]

Length is not phonemic.

In unstressed position, only vowels a i u may occur: e is pronounced like i and o is pronounced like u. This is always shown in spelling, i.e., e o never occur in unstressed position in spelling nor in pronunciation.

Some dialects do not distinguish between [a] and [ɐ], but only use [a].

Tirkunan does not have phonemic diphthongs. Phonemically, adjacent vowels are separate syllables, and there is no theoretical limit of vowel sequences. Same vowels merge, though.

Spelling does not mark the phonetic syllables intuitively, but is focussed on indicating phonemic syllables and stress. This means that, for example, phones may belong to an adjacent phonological word instead of the one they are written in: comu ni cel [ˈkomʊ ni kɛl] 'like in heaven'.

Syllable breaks are as follows: V-V, V-CV, VC-CV, VC-CCV..., i.e., a single consonant belongs to the next syllable, while of a sequence of at least two consonants, the first one closes the previous syllable. There is one exception: stop + r,l behaves like one consonant.
pan /pan/ [pan] bread
bisti /ˈbis.ti/ [ˈbɪstɪ] wild animal
fimbra /ˈfim.bra/ [ˈfɪmbrɐ] woman
matra /ˈma.tra/ [ˈmatrɐ] mother
reni /ˈre.ni/ [ˈrenɪ] kingdom
fili /ˈfi.li/ [ˈfilɪ] son; daughter

Vowel Length

Vowel length is non-phonemic. Yet vowels have different quantity in pronunciation. In stressed, open syllables, a vowel is long. Otherwise, it is short. The quantity changes only slightly, so we do not indicate it in this document except for stressing the slight difference. The quality, however, changes as the above table shows.

Foreign Names and Loans

Foreign names are usually spelled in Tirkunan-style phonetic spelling, even if the source language uses the Latin alphabet. This includes names of cites, countries, places, but also foreign names of people.

Some Latin consonant graphemes are not used by Tirkunan natively. Some of these are used in transscribing foreign names. There is an educated pronunciation that is closer to the transscribed language, and a common one that just uses the 'closest' Tirkunan phoneme.

Grapheme Educated Common
h /h/ silent
j /j/ /i/
k /k/ /k/
q /q/ /k/
x /ʃ/ /s/
tx /t͡ʃ͡/ /ts/
w /w/ /u/
y /ʎ/ /i/
z /z/ /s/

The graphme k is sometimes used like c if then transscribed language uses that grapheme. There is no difference in pronunciation. The grapheme x is used also in some recent loans into Tirkunan, but is usually quickly replaced by s.

Additional to the plain Latin letters, some letters with diacritics are used in educated transscriptions to more closely render the foreign pronunciation. This is by no means as close as it could get, but only a few sounds are rendered. The following tables gives a list of phonetic and typical educated spelling. It also shows how some sounds are not spelled distinctively.

Phone Educated Spelling Common Pronunciation
[øˌ œ] ë /e/
[yˌ ʏ] ï /i/
[ɤ] o /o/
[ɯ] u /u/
[ɨ] i /i/
[ʉ] u /u/
[ə] ä /a/
[æ] e /e/
[ʃ] š /s/
[ʒ] ž /s/
[t͡ʃ͡] /ts/
[d͡ʒ͡] /ts/
[ɲ] nj /ni/
[c] cj /ki/
[ɟ] gj /ɡi/
[xˌ χ] ch /k/

The spelling x for [ʃ] is preferred if the source language uses it that way and in loans into Tirkunan, otherwise, š is used.


Stress can be determined directly from spelling. Words are stressed

on the last syllable if an apostrophe is put after the word
on the last syllable if the word ends in a consonant
on the last syllable if the word is phonemically monosyllabic
on the penultimate syllable if the word ends in a vowel

Note that for the sake of determining stress, each vowel is counted as a separate syllable, i.e., phonetic diphthongs like ia ie io iu ai au ou still count as two syllables.

A nasal prefixed to a consonant cluster is phonetically articulated as a syllabic nasal, but phonemically, it is not a separate syllable, i.e., mpre is a single phonemic syllable and not stressed on the nasal, but it is two phonetic syllables, with the nasal a separate phonetic syllable.

Exceptionally, if a final consonant is dropped by phonotactic phenomena, or because it is lexicalised that way, a stress on the word final vowel is indicated by an apostrophe. An apostrophe would also be used if the stress stays on the last syllable when adding a prefix, but in Tirkunan, stress may move backwards, so this does not happen.

Stress and Affixes

In Tirkunan, stress often moves towards the end of the word when suffixes are added, as stress is relative to the end of a morphological word.

If prefixes are added, stress may also move towards the beginning of the word, in case the word to which a prefix is added has only one syllable and ends in a vowel: multisyllabic words that end in a vowel are stressed on the penultimate. E.g. di + mpre becomes dimpri [ˈdɪmprɪ].[2]

When a suffix is then added to such a prefixed word and the stress shifts back to the original syllable, the reduced vowel stays, i.e., derivation eliminates knowledge about original vowels, e.g. dimpri [ˈdɪmprɪ] + (a)t becomes dimprit [dɪmˈprɪt].[3]

Some examples

mpir /m.pir/ [m̩pɪr] to rule, to command
mpiri /ˈmpi.ri/ [m̩ˈpirɪ] empire
mpirian /m.pi.ri.ˈan/ [m̩pɪrˈjan] imperial
mpiratur /m.pi.ra.ˈtur/ [m̩pɪrɐˈtʊr] emperor
racun /ra.ˈkun/ [rɐˈkʊn] some(one)
racu' /ra.ˈku/ [rɐˈku] some(one) (before consonant)
aoi /a.ˈo.i/ [ɐˈo͡ɪ̯] today
ariu /a.ˈri.u/ [ɐˈri͡ʊ̯] to arrive
arivi /a.ˈri.vi/ [ɐˈrivɪ] arrival
useru /u.ˈse.ru/ [ʊˈserʊ] to observe
usiruti /u.si.ˈru.ti/ [ʊsɪˈrutɪ] observation
pricau /pri.ˈka.u/ [prɪˈka͡ʊ̯] to pre-dig
scau /ˈska.u/ [ska͡ʊ̯] to excavate
scavati /ska.ˈva.ti/ [skɐˈvatɪ] excavation
ciutat /ki.u.ˈtat/ [kjʊˈtat] city
Iuan /i.u.ˈan/ [jʊˈan] John
Iacou /i.a.ˈko.u/ [jɐˈko͡ʊ̯] Jakob
Iuli /i.ˈu.li/ [ˈjulɪ] Julia
air /a.ˈir/ [ɐˈɪr] yesterday
aoi /a.ˈo.i/ [ɐˈo͡ɪ̯] today
mpre /m.pre/ [m̩pre] fill up
dimpri /ˈdim.pri/ [ˈdɪmprɪ] empty out



In consonant clusters, the point of articulation is sometimes assimilated: n is [ŋ] before c g and [m] before b p. n m are both [ɱ] before v f. These shifts are partially shown in spelling when morphemes join: e.g. n + vrot > mvrot, but this is not shown across word boundaries.


Phonetically, dialects may exhibit diphthongs, although phonemically, there are none. I.e., diphthongs do not influence determination of stress. The following table lists all possible diphthongs that a speaker may have, but dialects and ideolects may differ in what diphthongs are actually used. This text will show phonetic pronunciation with the maximum amount of diphthongs, just to show what is possible, but speakers may have much less, maybe none.

Basically all vowel sequences ending with u or i as well as those starting with i may be diphthongs. Vowels sequences starting with u are very unlikely, maybe only loan words have them, so these are not shown in this section.

As usual, e o cannot occur in unstressed position.
spelling phoneme
au /au/ [a͡ʊ̯] [a͡ʊ̯] [ɐ͡ʊ̯]
eu /eu/ [e͡ʊ̯] [e͡ʊ̯] -
iu /iu/ [i͡ʊ̯] [i͡ʊ̯] [ɪ͡ʊ̯]
ou /ou/ [o͡ʊ̯] [o͡ʊ̯] -
ai /ai/ [a͡ɪ̯] [a͡ɪ̯] [ɐ͡ɪ̯]
ei /ei/ [e͡ɪ̯] [e͡ɪ̯] -
oi /oi/ [o͡ɪ̯] [o͡ɪ̯] -
ui /ui/ [u͡ɪ̯] [u͡ɪ̯] [ʊ͡ɪ̯]
ia /i͡a/ [jaˑ] [ja] []
ie /i͡e/ [jeˑ] [] -
io /i͡o/ [joˑ] [] -
iu /i͡u/ [juˑ] [] []
Note that iu can be two different diphthongs, depending on stress. Unstressed, usually [] is preferred.

Falling diphthongs, ending in i, u, like au ou ai ui ..., are always considered long (if the dialect has them). This means that the main vowel is only reduced in quality in unstressed position, but not due to a closing consonant.

Rising diphthongs, starting with j, may be long or short in speakers that have them, i.e., depending on the closedness of the syllable.

Diphthongs [j͡ɪ], [u͡ʊ̯] do not occur, because they collapsed into i and u, respectively.

No falling diphthongs will be used if stress falls on the second vowel.

No raising diphthongs will be used if stress falls on the first vowel.

spelling syl. & stress pronunciation translation phenomenon
Caitan /ka.i.ˈtan/ [ka͡ɪ̯ˈtan] Caietana a i are unstressed: rising diphthong
Mai /ˈma.i/ [ma͡ɪ̯] May a is stressed: rising diphthong
pais /pa.ˈis/ [pɐˈɪs] country i is stressed: no rising diphthong
cau /ˈka.u/ [ka͡ʊ̯] to dig a is stressed: falling diphthong
ariu /a.ˈri.u/ [ɐˈri͡ʊ̯] to arrive i is stressed: falling diphthong, i.e., not []
glurius /ɡlu.ri.ˈus/ [ɡlʊrˈjʊs] glorious u is stressed: falling diphthong
Iuli /i.ˈu.li/ [ˈjulɪ] Julia u is stressed: falling diphthong
ciutat /ki.u.ˈtat/ [kjʊˈtat] town i u are unstressed: usually rising diphthong

Sequences of Same and Similar Phonemes

Sequences of same vowels collapse into a single vowel, even across word boundaries. I.e., There are no long vowels. This happens after stress and syllable openness is considered, i.e., there is no change in stress or vowel quality.

Similarly, sequences of same phonemic, but reduced and unreduced vowel collapse into the unreduced vowel, e.g. [] > [i]. Stress and vowel qualities do not change.

Sequences of same consonants collapse into a single consonant inside words (no geminates), but are kept and pronounced across word boundaries. Some pronological assimilation will occur, particularly gemination, e.g., two plosives merge such that only one release is audible, but making the merged consonant long. The same happens to most consonants, but the effect is less obvious as the result probably sounds just like pronouncing two of the same consonant. For nasals, if the second nasal is syllabic, it stays that way and the first nasal is pronounced separately as a coda of the previous syllable.

Sequences of two different plosives may merge similarly in that the first plosive loses its audible release. The sequence keeps its two consonant length. This is not universally done but only by some dialects/idolects, so it is not marked in this text.

Similary, nasal release or no audible release occurs with plosive plus homoorganic nasal sequences. For some dialects/ideolects, it may occur with non-homorganic nasals, particularly if the nasal's point of articulation is closer to the lung. Due to the diversity of realisaion, and since this level of detail may confuse readers, this is not marked in this document.

For lexicalised phrases that act as a single word, more assimilation may occur: double consonants and vowels may collapse into single ones, etc. This is not reflected in the phonetic descriptions in this text if the words are written separately, because such entries are usually written as one word.

spelling syl. & stress pronunciation translation phenomenon
posi imant /ˈpo.si i.ˈmant/ [ˈposɪ ɪˈmant] put a magnet vowel + vowel: syl. po is still open
posi butic /ˈpo.si bu.ˈtik/ [ˈposɪ bʊˈtɪk] put a bottle no vowel + vowel: syl. po is open
mbros imant /m.bros i.ˈmant/ [m̩brɔ sɪˈmant] bite a magnet syl. bros is closed
acit talian /a.ˈkit ta.li.ˈan/ [ɐˈkɪt̚ tɐlˈjan] Italian vinegar consonant + consonant: syl. cit is still closed
acit di Tali /a.ˈkit di ˈta.li/ [ɐˈkɪt̚ di ˈtalɪ] vinegar from Italy t + d: syl. cit is still closed
ali i oli /ˈa.li i ˈo.li/ [ˈalɪ i ˈolɪ] garlic and oil written separately here, but usually lexicalised as a single word
aliioli /a.li.ˈo.li/ [ɐlˈjolɪ] aioli functions and is pronounced as one word
ap p'ucel /ap pu.ˈkel/ [ap̚ pʊˈkɛl] water for the bird pi stays audible as pit is pronounced separately
ap ucel /ap u.ˈkel/ [a pʊˈkɛl] bird water without the pi
ap marel /ap ma.ˈrel/ [ap mɐˈrɛl] yellow water probably has nasal/no audible release of p (not shown)
vult nic /vult nik/ [vʊlt nɪk] small vulture probably has nasal/no audible release of t (not shown)

Consonants Cross Syllable Breaks

Consonants at ends of words before vowels in the next word are pronounced as if they belonged to the second word. This does not change the closedness of the last syllable of the first word, i.e., does not change the vowel quality of that syllable.

Similarly, if a word ending in a vowel is followed by a word that starts with a consonants cluster, the first consonant may be pronounced as if part of the previous syllable -- for most consonants, this may be an insignificant pronunciation difference, but it may reduce the number of syllable when a syllabic nasal becomes non-syllabic by this.

spelling syl. & stress pronunciation translation phenomenon
posi imant /ˈpo.si i.ˈmant/ [ˈposɪ ɪˈmant] put a magnet s moved into syllable with following i
mbros imant /m.bros i.ˈmant/ [m̩brɔ sɪˈmant] bite a magnet s moved into syllable with following i
posi sal /ˈpo.si sal/ [ˈposɪ sal] put salt s is not a cluster, will not move back
posi spraga /ˈpo.si ˈspra.ɡa/ [ˈposɪs ˈpraɡɐ] put esparagus s can move back out of cluster: not much difference in pronunciation
vis Mbrac /vis m.brak/ [vɪs m̩brak] see Marcus m is syllabic
ami Mbrac /ˈa.mi m.brak/ [ˈamɪm brak] love Marcus m can move back and becomes non-syllabic
mpiri /ˈmpi.ri/ [m̩ˈpirɪ] empire syllabic nasal
ni mpiri /ni ˈmpi.ri/ [nim ˈpirɪ] in an/the empire nasal is non-syllabic
fimbra /ˈfim.bra/ [ˈfɪmbrɐ] woman nasal is non-syllabic


Vowel Development

Vowel Shift Overview This figure shows an overview of what happened to the vowels from Classical Latin to Modern Tirkunan.

Vowel phones shown in the same colour are allophones of the same phoneme.

In detail, the following is shown. The Classical Latin vowel system had ten phonemic vowels, five long ones and five short ones, /iː i eː e aː a oː o uː u/ which were probably pronounced [iː ɪ eː ɛ aː a oː ɔ uː ʊ].

In Vulgar Latin, phonemic length was lost, but the quality of the vowels was retained, with three mergers: /aː a/ > /a/, /eː ɪ/ > /e/, and /oː ʊ/ > /o/. The result is a seven vowel system in Proto-Romance: /i e ɛ a u o ɔ/.

The vowels in Proto-Romance then came to be pronounced long and short again, depending on whether a syllable was open or closed: open syllables had long vowels, closed syllables short vowels. This is shown by the split of all vowels in the middle of the figure.

Tirkunan developed differently from this point than many other Romance languages. It collapsed the seven vowels with phonetic length into a five vowel system with phonetic length by four mergers: [e i] > [ɪ], [eː ɛː] > [], [o u] > [ʊ], and [oː ɔː] > [].

S25 Vowel Shift This figure shows the final vowel shift that reduced the four vowel heights of Proto-Romance to the three heights of Tirkunan.

The phones of modern Tirkunan are not lengthened much, so the modern phones can be given as [iˑ ɪ eˑ ɛ aˑ a uˑ ʊ oˑ ɔ]. In stressed syllables, the lengthening is slightly stronger.

Note that in this document, vowel length is usually not indicated for Tirkunan pronunciation, even in phonetic descriptions, since the differences are rather small.



The morphology has greatly simplified compared to Latin. Tirkunan is grammatically isolating with agglutinative derivation.

The verb is usually quite complex morphologically in Romance languages, but it is undeclinable in Tirkunan, except maybe for a past participle, which could also be counted as derivation.

All tenses, aspects, and moods are formed analytically with auxiliaries. The analytical forms that have completely replaced the old system, are still quite typical for Romance. There is only one irregular verb (e 'to be') with a few irregular verb forms, the rest of the verbal system is analytical and regular.

Tirkunan has dropped all grammatical gender distinctions.

There is no trace of Latin case. This is just like in many Romance languages. Tirkunan has thoroughly dropped case from pronouns, too.


The derivational system of Tirkunan is quite productive. Stems never change irregularly when affixes are added. When Latin allows several ways to form a word, often only one has survived and was generalised to be used in many more cases. As an example, from the verb am 'to love', we can derive amatur 'lover', ami 'love'.

The pronouns were regularised and some plural forms reanalysed as polite/formal forms. Only first person distinguishes number (i.e., 'I' vs. 'we').


For most words, Old Tirkunan drops a final vowel if the preceding consonant was -l, -r, -n, -s, -t, -c or -p (voiceless plosives or alveolar sonorants or sibilants), and if there was no final cluster except -(n,m,r,l,s)(p,t,c), and -rn. Final vowels that are not dropped collapse into -a (earlier [ə]). Nouns ending in rising diphthongs behaved similarly, dropping the main vowel, leaving the plain glide. This [j] or [w] glide becomes syllabic again and shows up as a final -i or -u in the corresponding words, e.g. glori < glōria, mpiri < impērivm, coru < corvvm.

Fusion and Elision

Tirkunan exhibits most prominently dropping of vowels or consonants when two morphemes are jointed in derivation or between words next to each other in a sentence. Some small and frequent words have alternative forms, depending on the preceding or following word. The following list shows the main rules, roughly from most widespread rule to most restricted/special.

Firstly, final 'a' in words and prefixes and initial 'a' in suffixes is always a very weak vowel that is dropped readily, because it derives from an epenthetic [ə].

A final u in words ending in a two-vowel sequence regularly becomes v when another vowel follows, e.g., when a derivational ending is suffixed.

Adjacent same vowels merge into a single vowel. This is shown in spelling only inside words, but spoken language also exhibits this across words.

Final 'a' is dropped before vowels all words and affixes ombra + umbla > ombr' umbla 'humble man'
Final u becomes v between vowels many words deu + at > divat 'debt'
Initial 'a' is dropped after vowel all suffixes solu + -ati > suluti 'solution'
Vowel is dropped after/before same vowel all affixes neru + -us > nirus 'nervous

The general 'final "a" drops' rule is not shown in the lexicon, because it is universal. Still, as usual, there are exceptions to the rules even if the above table claims they apply to 'all' words. E.g. loan words ending in 'a' may not drop that 'a'. Such exceptions are clarified in the lexicon.

Many small words, particularly those that are often unstressed in a sentence, as well as affixes have additional rules of dropping vowels and consonants. These are not universally applied, but the lexicon lists this with each word and affix. The patterns repeat, i.e., this is not completely arbitrary, but follow the general rules of which consonant clusters and vowel sequences are acceptable -- typical those that occur already in stems. Typical phenomena are shown in the following table:

Final vowel is dropped before vowels small words and affixes li + umbra > l'umbra 'the shadow'
Final vowel is dropped after m, before v,f small words and affixes comu + flur > com' flur 'like (a) flower'
Final cons. is dropped before cons. small words and affixes suv + tabra > su tabra 'under (a/the) table'
Final vowel is dropped after vowel individual words ni + li > ni'l 'in the'
Initial phoneme is dropped special circumstances individual affixes supril + itat > supriltat 'superiority'

The most complex rules are those of ni and li which drop the final vowel not only if a vowel follows, but also when a vowel precedes the word.

The interaction between preposition and article is special. Firstly, the article only fuses backwards with the previous word if that is a monosyllabic preposition -- it does not fuse backwards with any other word: sopra li ter, not *sopra'l ter.

Secondly, the fusion of preposition and article is preferred over fusion with surrounding words: ni'l ap 'in the water' instead of *ni l'ap.

Note again that pronouns to not fuse: Cu le e bon 'with him/her, it is good' instead of *Cu l'e bon. But articles do: cu'l gen 'with the man/woman/person.

Some derivational endings work in special ways, e.g., -ament drops the initial a also after l and the abstraction ending -ati sometimes reduces to i like in visi 'vision'. The lexicon lists this in detail case-by-case.

In compound phrases with (dropped) di, final vowels drop in more words than usually. Because this process is not regular and cannot be generalised into a rule, these cases are all marked in the lexicon. Also, it is not wrong to keep these vowels, it is just that usually, the additional abbreviation is used.

oli + di + uliu > oli d'uliu 'olive oil' long version with di
oli + uliu > ol' uliu 'olive oil' usual version without -i
oli + uliu > oli uliu 'olive oil' uncommon but not wrong

Apostrophe and Stress

The apostrophe is used to indicate vowels and consonants that have been dropped due to elision and sandhi, and also to indicate the case of a word whose last syllable is stressed if it ends in a vowel.

Consonants dropped from single syllable words are not indicated with an apostrophe: cu ti 'with you' instead of *cu' ti. Otherwise, dropped consonants and vowels are indicated by an apostrophe.

For multisyllabic words, e.g., catun: catu', racun: racu', and nisun: nisu', this apostrophe at the end also indicates that stress is still on the last syllable if the n is dropped.

A dropped vowel or consonant is indicated on the side that causes the drop (i.e., not the side that drops the phoneme): racu' can 'some dog' and ni'l reni 'in the kingdom', although the i drops on the right side of li. If both sides could trigger the drop, then the right side carries the apostrophe, because the only way this can happen is if the first word is be a preposition: ni'l ap 'in the water' instead of *ni l'ap.

Theoretically, when prefixing a syllable to a mono-syllabic word that ends in a vowel, to indicate that stress stays on the last syllable, an apostrophe could be written at the end of the resulting word: an imagined *di + xe would become *dixe'. This apostrophe would then be removed if a suffix was added: *dixe' + (a)t would become *dixet. The lexicon can mark such stems with the 'stressed' feature, but there are no such stems in the current lexicon, i.e., what actually happens is that stress moves backwards in such situations in Tirkunan (like already in Classical Latin), so no apostrophe is needed, e.g., di + mpre becomes dimpri [ˈdɪmprɪ].

The drop of a single vowel after a consonant at the end of a stem does not change stress. However, if a ends in two vowels and is at least trisyllabic, dropping the final vowel might, a similar situation could theoretically occur that would require a stress marker at the end, e.g., in an imagined *axia + axa, the result could theoretically be axi' axa. However, Tirkunan does not have stems that behave like that -- the stress would shift backwards in this case, and no apostrophe would be needed. Also, there is no word in the current Tirkunan lexicon that drops a final vowel if the word ends in two vowels.

White space around an apostrophe is handled as follows. Firstly, no additional whitespace is used when adding an apostrophe. If only a single letter remains in the word that dropped a letter, whitespace is removed between an apostrophe on the right side and the following word: l'ap instead of *l' ap. No whitespace is used either between an apostrophe on the left side and a previous monosyllabic word: di'l instead of di 'l. Otherwise, normal whitespace is used: racu' reni instead of *racu'reni. Also note the difference in racu' reni (apostrophe is used to indicate stress) and u reni (monosyllabic: no apostrophe is used).

The following table shows many elision phenomena for prepositions, articles, and verbs:

ni reni [ni ˈrenɪ] in a/the kingdom
n'u reni [nu ˈrenɪ] in a kingdom
ni'l reni [nɪl ˈrenɪ] in the kingdom
ni gradin [ni ɡrɐˈdɪn] in a/the garden
n'u gradin [nu ɡrɐˈdɪn] in a garden
ni'l gradin [nɪl ɡrɐˈdɪn] in the garden
can ni mpiri [kan nim ˈpirɪ] a/the dog in a/the empire
ni mpiri [nim ˈpirɪ] in a/the empire
andra ni mpiri [ˈandrɐ nim ˈpirɪ] walk in a/the empire
n'u mpiri [num ˈpirɪ] in an empire
ni'l mpiri [nɪl m̩ˈpirɪ] in the empire
n'umbra [ˈnʊmbrɐ] in a/the shadow
andra n'umbra [ˈandrɐ ˈnʊmbrɐ] walk in a/the shadow
n'un umbra [nʊ ˈnʊmbrɐ] in a shadow
ni'l umbra [nɪ ˈlʊmbrɐ] in the shadow
cu reni [ku ˈrenɪ] with a/the kingdom
cu gradin [ku ɡrɐˈdɪn] with a/the garden
cun u gradin [kʊ nu ɡrɐˈdɪn] with a garden
cu'l gradin [kʊl ɡrɐˈdɪn] with the garden
cu mpiri [kum ˈpirɪ] with a/the empire
cun u mpiri [kʊ num ˈpirɪ] with an empire
cu'l mpiri [kʊl m̩ˈpirɪ] with the empire
cun ombra [kʊ ˈnɔmbrɐ] with a/the man
cun un ombra [kʊ nʊ ˈnɔmbrɐ] with a man
cu'l ombra [kʊ ˈlɔmbrɐ] with the man

Personal Pronouns

sg. pl.
1. mi no
2.informal ti
2.formal vo
2.representative tro
3. le
impersonal se
reflexive se
relative ce

Morphological case and mostly also number have disappeared, also for pronouns. The pronouns are not distinguished by case, i.e., subject and object pronouns, as well as oblique pronouns and possessive pronouns are the same, and none of the pronouns usually reduces, i.e., no vowel or consonant may drop.[4]

Number has disappeared in the personal pronouns, too, except in the first person. The second person reinterpreted plural to indicate formal vs. informal addressing, based on the usage of the plural form for formal address in earlier stages of the language. The third person pronoun lost its number just like nouns did, mainly because referring back with number information is pointless if no number is conveyed in the preceding noun phrase. This means that Tirkunan has only one third person pronoun left, which is often not enough to disambiguate what it references, so noun phrases or demonstrative determiners with selector nouns are often used instead.

Note that se cannot be in subject position.

The impersonal pronoun se is used if no specific entity is expressed. This pronoun is exclusively used as a subject, complementing the reflexive pronoun.

Note that pronouns are not mandatory, in particular subject pronouns. This can lead to ambiguities that have to be resolved by context, because the verb does not carry any information about person either.

Note that how le behaves differently after a vowel than the article: e.g. di + le (pronoun) is di le while di + li (article) is di'l.

The pronoun tro is used in conversations, often formal, to refer to the organisation, company, or group the interlocutor is part of, often in order to avoid direct addressing of the interlocutor. It derives from tva horda 'your gang'. The plural is used when talking to a group of representatives.

Since tro is 2nd person, for its reflexive, also tro is used, not se.

The 3rd person pronoun se is strictly reflexive. This pronoun is only used to refer to a 3rd person, i.e., reflexive 1st and 2nd persons are referred to using 1st and 2nd person pronouns, resp.

Mi va ni gradin mi. I (am) walk(ing) in my garden.
Vo va ni gradin vo. You (are) walk(ing) in your garden.
Tro va ni gradin tro. Your guy (is) walk(ing) into his garden.
Le va ni gradin se. He/She (is) walk(ing) in his/her (own) garden.
Le va ni gradin le. He/She (is) walk(ing) in his/her (someone else's) garden.
Se va ni gradin se. One (is) walk(ing) in one's (own) garden.
Se va ni gradin le. One (is) walk(ing) in his/her (someone else's) garden.


Possessives follow the noun and are formed with d(i) + noun/pronoun:

patra di fimbra father of a/the woman
patra d'ombra father of a/the man
patra di'l ombra father of the man
patra di le his father

Pronouns follows the same pattern. Also not that di may drop, particularly for lexicalised phrases and also for pronouns, so shorter possessives exist:

patra di mi my father
patra mi my father
patra no our father
patra ti your father
patra tro your guy's father
patra se his/her (refl.) father
patra le his/her (non-refl.) father

Determiners and Indefinite Pronouns

li, l', 'l < ille, illa article the (singular) reduces to l' before vowels and, otherwise, to 'l after vowels
lur < illōrvm article the (plural)
un, u < vna, vnv(m) article a, an coincides with number 1
cul < eccv(m) ille determiner, pronoun this; that; this one; that one
tut < tōtv(m)? determiner, pronoun every, all
catun, catu' < *cata vnvm determiner, pronoun each, each one, everyone
racun, racu' < aliqv(is) vn(vm) determiner, pronoun some(one), some(thing), any(one), any(thing) {singular}
nisun, nisu' < ne ipsv(m) vn(vm) determiner, pronoun no-one

Number is indicated on the article. There are the following articles:

li [li] 'the': singular definite
lur [lʊr] 'the': plural definite
un [ʊn] 'a/an': singular indefinite, also numeral 'one'

The articles are optional and are used only rarely, in constrast to many other Romance languages.

Note that there is no plural indefinite article. 'Some' can be expressed using rapan.

In some situations, Tirkunan does use the definite articles, e.g.:

pronoun-like phrases from numbers lur du [lʊr du] 'the two'

un, catun, racun and nisun drop the final n to become u, catu', racu', and nisu' before consonants.

u bal [u bal] a dance
u cafi [u ˈkafɪ] a cafe; a coffee
u fimbra [u ˈfɪmbrɐ] a woman
u mpiri [um ˈpirɪ] an empire
un ombra [ʊ ˈnɔmbrɐ] a man

Nouns and Articles

Nouns have only one form and are not inflected.

Since number is marked only on pronouns and articles, noun phrases are often also underspecified for number: gen filic 'a/the/some happy man/men/woman/women/person/people. The previous example also shows that the loss of the thematic gender vowel from Latin is continued in the lexicon, which often has gender-neutral nouns where English would be more specific, e.g. the very frequent gen 'man/woman/person' and many kin terms like tiul 'aunt/uncle' and fili 'son/daughter'.


The declension system of Latin was dropped quickly in Tirkunan. More quickly than in other Romance languages. The accusative singular form, pronounced without -m in Vulgar Latin, was the only form that was left over. Most vowels in endings collapsed to only a /ə/ very quickly. In some cases, the thematic vowel of the declension was mistaken to be part of the stem, so some traces of the old -u- declension are still visible today.


To model the history of nouns, I use the following simulation. It takes care of the collapse of declension classes and deterioration of endings.

Nouns are based on the accusative singular form, but a general /-əm/ ending is assumed instead of the original ending, because declensions are assumed to having vanished greatly in late Vulgar Latin and so 1st (-a-), 2nd (-o-), 3rd (mixed), and 5th (-e) declensions are basically treated the same.

However, as some (very few) -i- declension nouns and some -u- (4th) declension nouns retained the -i-/-u- in the form of a final stem glide., assumed /jəm/ and /wəm/ endings are used. (In the alternate timeline, Þrjótrunn also retained the 4th declension for some words, so the Vulgar Latin there may be a little different from ours.)

Adjectives are derived in the same way, starting with the f.acc.sg. form (there should be no difference when starting with other genders' acc.sg. form).

Note that since the GMP is made for shifting words that resemble the Classical Latin written form to make life easier (it is the look of those classical forms the author is more familiar with), input is reconstructed to a classical form, and thus we use the /m/ ending although it is dropped by the GMP and does not cause any change the result. This is just a technical cheat for making it look like an accusative – this conlang is based on Vulgar Latin nevertheless. The GMP handles Vulgar forms, too.

You can experiment with these rules here.

Input for GMP
Tirkunan Pronunciation
porta 1st (-a-) port-a-m *port-əm prot [prɔt]
hortvs 2nd (-o-) hort-v-m *hort-əm rot [rɔt]
corvvs 2nd (-o-) corv-v-m *corv-əm coru [ˈkorʊ]
imperivm 2nd (-o-) imperi-v-m *imperi-əm mpiri [m̩ˈpirɪ]
bonvs,-a,-vm 1st, 2nd (-a-, -o-) bon-a-m *bon-əm bon [bɔn]
nox 3rd (consonantal) noct--em *noct-əm not [nɔt]
tvrris 3rd (-i-) tvrr-i-m *tvrr-jəm tori [ˈtorɪ]
fēlīx 3rd (consonantal) fēlīc-e-m *fēlīc-əm filic [fɪˈlɪk]
portvs 4th (-u-) port-v-m *port-wəm protu [ˈprotʊ]
dīēs 5th (-e-) dī-e-m *di-əm di [di]

There are some exceptions from this construction rule, especially for very short words, e.g. deu < dēvs.


Attributive adjectives always follow the noun they modify.

The comparative and superlative is formed with pru 'more'. The superlative is formed by using masma 'most'. For smaller comparison, min 'less' is used, which has no superlative, so it is modified also using masma if really needed.

ciutat gran a/the large town
ciutat pru gran a/the larger town
ciutat masma gran the largest town
ciutat min gran a/the less large town
ciutat masma min gran the least large town
ciutat trop gran a/the town that is too large

Also note the usage of gran 'big/famous one' and san 'saint': they are adjectives and follow the noun, unlike other Romance languages.

Cral Gran Charlemagne
Ciceli San Saint Cecily

Distinction of Adjective vs. Noun

In Tirkunan, the distinction between adjectives and nouns is weak. E.g., the ending -tur is often used to form adjectives while -tor was used exclusively for noun formation from verbs in Latin and adjectives used participle endings, e.g., muritur 'mortal (both adj. and noun)'.[5] The boundary is further blurred by Tirkunan's tendency to drop the connective di preposition between a noun and a modifying noun phrase, e.g., sup cur bou vs. sup di cur di bou 'oxtail soup'.

There still is a formative ending -an to derive adjectives from nouns, derived from and similar to -ānvs and -ālis, but this gets more and more unimportant, especially in spoken language.

So when the lexicon lists a noun or an adjective, using that word as the respective other word class is likely also possible.


Many lexicalised adverbs end in -u to mark them as adverb, probably from the -u in comu < cōmō < qvōmōdo.[6] This ending is usually dropped before vowels (but not for all words). Not all lexicalised adverbs end in -u.

sut(u) < svbitō 'suddenly'
com(u) < qvōmodo < *qvō modō 'how'
sim(u) < sīc mentem 'so, in this way'
tan < tanqvam 'so, so much'
is < *est eccv̄ illvm qvī Y/N question marker
crau < crās 'tomorrow'
aoi < ad hōdie 'today'
air < ab herī 'yesterday'
acur < hāc hōra 'now'
laur < ad illam hōram 'then'


The -u ending is a remnant of Latin like in svbitō, which was generalised to mark adverbs. Sometimes the -u spread to other word classes, too, e.g. in candu 'when'. Hence, this -u is essentially now a marker for 'small words', although it is not consistently used.


Adjectives can be converted into adverbs by adding the suffix -(a)ment. After vowel and -l (but not after -r), the -a- is dropped.

A few adjectives have irregular adverbs.

custant 'constant' > custantament 'constantly'
grau 'serious' > graument 'seriously'
runali 'ordinary' > runaliment 'ordinarily'
pal 'equal' > palment 'equally'
sul 'single' > sul 'only',
= > sulment 'alone', 'only'
bon 'good' > bin 'well'
mal 'bad > mal 'badly'
sut 'sudden' > sutu 'suddenly'
puc 'few, little' > puc 'a bit'
mbrut 'many, much' > mbrut 'very'


Conjunctions are generally unstressed, so all the vowels collapse to unstressed a, i, u. Conjunctions that are derived from other words (e.g., from relative pronouns or adverbs) also lose stress and, therefore, may have different vowels.

Conjunction Translation Derives From Translation
i and
sinu but
upur or
i ncur and yet
nipi either, neither
c(i) that, than ce who, what, that
pi c(i) because pi ce why, therefore
u(v) where ou where
cumu how comu how
si if
cand(u) when
upur...upur... either...or
nipi...nipi... neither...nor
nu sul...sinu ncui... not only...but also, both...and


cu(n) with
sint(i) without
a(d) at, to, for
di, d' of, off, from
pi, p' for, by, towards
ni, n' in, into, to, on
tra between, inside, within
contr(a) against
sopr(a) over, about
su(v) under
avan before, in front of
dap after, behind
fin until, up to
dis since, starting from, from

Note that usage of prepositions is largely lexicalised, i.e., translating does not mean to map prepositions 1:1 according to the this table. One typical source of confusion is the preposition 'on', which has no direct equivalent in Tirkunan. In other Romance languages, a preposition may have been formed, either by reinterpreting sopra or by using a derivative of sv̄rsvm. In Tirkunan, usually ni is used. E.g. ni tabra 'on the table', ni lun 'on the moon'.

'Weak' consonansts will surface before vowels:

cu lat with milk
cun ap with water
a lat to milk
a cran to meat
ad ap to water

Tirkunan has a tendency in many dialects to drop the di preposition entirely when expressing a genetive construction or in ordinals, particularly in lexicalised and very common phrases. di is never dropped with pronouns (e.g. in di le 'his').

mbricat di pruc mbricat pruc flea market
Pul Iuan Di Du Pul Iuan Du John Paul the Second



The basic digits 0,..,9, as well as 20,..,90 and the larger numbers 100, 200, 1000, 2000, etc., are typical for a Romance language.

In contrast to other Romance language, Tirkunan has very simply constructed numbers for 11,..,19, by using dec+1,...,9. Other Romance language have reversed or even fused forms, usually up to around 16.

0 nul
1 un 11 dec i un, dec un 10 dec
2 du 12 dec i du, dec du 20 vint
3 tre 13 dec i tre, dec tre 30 trint
4 patr(u) 14 dec i patru, dec patru 40 patrint
5 cim(p) 15 dec i cim, dec cim 50 cimpint
6 ses 16 dec i ses, dec ses 60 sisint
7 set 17 dec i set, dec set 70 sitint
8 ot 18 dec i ot, dec ot 80 utint
9 nou 19 dec i nou, dec nou 90 nuvint
21 vint i un, vint un 32 trint i du, trint du 43 patrint i tre, patrint tre
100 cent 200 du cent 120 cent i vint, cent vint
1000 mil 2000 du mil
1000000 milun

As you can see, i can be used to string together the numbers. i is not necessary, however, and often left out. It is most frequently left out before a vowel, i.e., dec un '11', vint ot '28', etc. i.e., instead of using i.


The original synthetic ordinal number of Latin have been lost. Ordinal numbers in Tirkunan are formed analytically by using di + cardinal.[7]

gen di dec a/the tenth man/woman/person
li gen di dec the tenth man/woman/person
u gen di dec a tenth man/woman/person
Pap Iuan Pul Di Du Pope John Paul the Second
Pap Bindic Di Dec Ses Pope Benedict the Sixteenth

Note that this use of di is not considered optional as it is not a possessive. It is still often dropped in colloquial speech, because it sounds like counting and meaning roughly the same, basically like 'third person' vs. 'person no. three': gen di tre vs. gen tre.

There is an irregular ordinal number, marked in the lexicon as word type 'adj.num.':

prim first
gen prim ni lun first person on the moon


Fractions in Tirkunan are similar to ordinals, but they usually use a connecting pi. I.e., the formation of fractionals is completely analytic. Fractions are also used to express 'percent' pi cent, and 'per mille' pi mil, and 'ppm' pi milun, etc. Fractions behave syntactically like normal numbers, i.e., as a quantifier, they precede the noun. If it is a single unit, the un is sometimes dropped in colloquial speech.

u pi tre a third
tre pi patru three quarters
u pi ot litru vin an eighth of a litre of wine
pi ot litru vin an eighth of a litre of wine
patrint ot pi cent gen fu fimbra 48% of the people were women

There is an irregular fraction, marked in the lexicon simply as word type 'num.' like cardinal numerals:

mer half
mer litru vin half a litre of wine


Verbs in Tirkunan only have one morphological form, i.e., they are not inflected, but some derivational forms exist, most notably -at, which forms a past passive participle. This is used in analytic verb forms. It is also often used to enrich the lexicon with adjectives (e.g. sicat 'dry') and nouns (e.g. ripat 'receipt').

When an ending is added to a verb, like in any derivation, a possible final -a drops if the ending starts with a vowel. Likewise, an initial a- of endings drops after a possible i or u vowel on the verb stem.

A verb can be nominaliser, i.e., made into nouns or adjectives, in many ways using suffixes. The gerund or noun of process is formed by suffixing -ar: Mancar e bon 'Eating is good', Mi am mancar 'I like to eat' (lit. 'I love eating'), Mi e cansat di mancar 'I am tired of eating.'

Plain verb forms carry no tense information, i.e., the plain verb can be used for present, past, or future tense, and the meaning will be inferred from context. There are auxiliary verbs for marking tense. The only exception is the verb e, which expresses present tense and which has irregular past and future tense forms fu and eri, resp.


Participles are adjectives that derive from verbs. They can be used as nouns in Tirkunan with the meaning 'the ... one'. The verb e 'to be' has no participles. All other verbs regularly form one participle as follows, which is used in analytic verb forms, too.

da 'to give' lau 'to wash' lega 'to bind solu 'to solve' odi 'to hear'
past passive participle -(a)t dunat 'given' lavat 'washed' ligat 'bound' sulut 'solved' udit 'heard'

Tirkunan likes to replace short ce relative clauses (typically V or V+Obj) by participle constructions: gen amat bin 'the man/woman/person very much loved' for gen ce e amat bin.

The participle present, the gerund, and the gerundive of Latin have merged into a single suffix -an, but this is pure derivation and does not survive in verb forms in Tirkunan.

Analytical Forms

The following table lists the verb forms of Tirkunan.

Plain Mi manc. I eat. / I ate. / I will eat.
Present Tense mo + verb[8] Mi mo manc. I eat.
Past Tense au + verb+(a)t Mi au mancat. I ate. / I have eaten. (Lit. 'I have eaten.')
Mi au tenat frir. I was cold. / I felt cold.
Future Tense veni + verb[9] Mi veni manc. I will eat. (Lit. 'I come eat.')
Mi veni veni. I will come.
Negation nu(n) + verb Mi nu manc. I don't eat.
Positive Emphasis si + verb Mi si manc. I do/did eat.
Present Passive e + verb+(a)t Mi e mancat. I am eaten.
Past Passive fu + verb+(a)t Mi fu mancat. I was eaten.
Future Passive eri + verb+(a)t Mi eri mancat. I will be eaten.
Progressive sta + verb Mi sta manc. I am/was/will be eating. (Lit. 'I stand to eat.')
Prospective sta pi + verb Mi sta pi manc. I am/was/will be about to eat. (Lit. 'I stand for eating.')
Interrogative is + phrase[10] Is ti manc? Do/Did/Will you eat?
Is cul e gen ce am ti? Is this the person who loves you?
Optative 'to be' sa Mar sa crar. Let the ocean be warm.
Optative Active sa + verb Ap sa coc. May the water cook.
No sa va. Let's go!
Optative Passive sa + verb+(a)t Luc sa fit. Let there be light.
Imperative va + verb Vo va am mi! Love me! (formal)
Va manc pan! Eat the bread!
Vo va manc! Eat! (formal)
drop S and O Manc! Eat!

For formal addressing, the optative is prefered over the imperative, as it is felt less direct and less rude. E.g., instead of Vo va manc!, the optative Vo sa manc! is used. Usually, the formal, polite phrases tend to be longer anyway: Vo sa manc pan, pi prac! 'Please, try the bread.'

A corollary is that the passive imperative cannot be formed with a non-empty subject, because the passive never has an object. Passive imperatives seem to have limited usefulness anyway.

The progressive form has evolved from an earlier form with ad, e.g., mi sta a manc. The preposition in this construction was lost and is not used anymore today. In some cases, the old construction may be easier to understand intuitively by speakers of other Romance languages, e.g., Mi sta ad e mancat instead of modern Mi sta e mancat for 'I am being eaten.'

More analytical forms can be combined into longer forms. The following is a selection.

Present Progressive Mi mo sta manc. I am eating.
Past Progressive Mi au stat manc. I was eating. / I have been eating.
Future Progressive Mi veni sta manc. I will be eating. (Lit. 'I come stand eating.')
Present Progressive Passive Mi sta e mancat. I am being eaten.
Present Prospective Passive Mi sta p'e mancat. I am about to be eaten.
Past Progressive Passive Mi au stat e mancat. I was being eaten.
Future Progressive Passive Mi veni sta e mancat. I will be being eaten.
Interrogative Future Progressive Is mi veni sta manc? Will I be eating?
Optative Past Ombra sa au mancat. May the man have eaten.
Optative Progressive Passive Cran sa sta e mancat. May the meat be being eaten.
Negation + Anything Mi nu sta manc. I am not eating.
Mi nun au mancat. I did not eat. / I have not eaten.
Is mi nu veni sta manc? Will I not be eating?
Nu manc! Don't eat (that)!

The general order of the verb phrase is as follows:

Mood Negation, Emphasis, Other Aux. Tense marker] Aspect marker Passive Marker Verb
sa, va nu, si, ro, ariu, fi, pos, deu, ... mo, au, veni sta, sta pi e, fu, eri V, V+at

The order may be different if logic is different, e.g., the negation may be after 'other auxiliaries', as in 'I like not to read.' vs. 'I do not like to read.'. This group also allows multiple auxiliaries and nu to be stringed together.

Also, not the full sequence is possible: a tense marker is not used right before e, fu, eri is used, because that already contains the tense. Combining of analytical forms is further limited as the forms get longer. Especially spoken language does not like long sequences of auxiliaries and may leave things underspecified.

Also, the combination sta sta is usually not used. This will probably not hurt, because sta, as a stative verb, cannot really form the progressive (there is no progress while standing).

The YN-question marker is is not part of the syntactic verb phrase, but starts a sentence.

Note that no participles of e or au are necessary for any verb forms. The irregular verb e and all its irregular derivatives (fu, eri, sa) have no participles.

Irregular Verb

The following irregular synthetic verb forms exist of the verb e 'to be'. e has no past passive participle from (it would likely also be irregular *fut instead of a regular **iat).

sa optative auxiliary (derives from e)
*sa e sa optative passive auxiliary
*au fut fu was, were (past of e)
*veni e eri will be (future of e)

Some Verbs and Auxiliaries

e are, am, is Mi e frunari. I am a baker.
fu was, were Le fu frunari. He/she/they was/were a baker/bakers.
eri will be Ti eri frunari. Your will be a baker.
e, fu, eri, sa form passive + ppp Ti fu amat. You were loved.
sa forms optative + verb Ap sa coc. Let the water cook.
au forms past tense + ppp Mi au vinat. I came.
veni forms future tense Mi veni va. I will go.
va forms imperatives Va manc pan! Eat (the) bread!
pos be able to Mi pos manc. I can eat.
ro to want to + verb Mi ro am ti. I want to love you.
ro to want to Mi ro umbra. I want (some) shadow.
sapi to know how to + verb
sapi to know
manc to eat
am to love
ten to hold, to have
fi to make, to create; to do; forms causative
va to go
veni to come
di to say
deu to owe, to be in debt, must
ariu to manage to do s.t. Mi ariu veni. I manage to come.
ariu to arrive at Mi ariu Tali. I arrive in Italy.

Note that tense and aspect are not mandatory categories. They are often derived from context. This is especially true in narratives, which generally use the plain verb instead of au+VERB-at. An exceptional verb is fu, the only synthetical past tense, which is used in narratives frequently.

Transitivity, Reflexivity

Verbs in Tirkunan can be either transitive (marked v.t. in the lexicon) or intransitive (v.i.), depending on whether they take an object or not. Further, the object may be se in which case the verb is said to be reflexive, marked (v.refl.) in the lexicon if a separate entry is necessary because the meaning is different from the transitive meaning.

Some verbs have neither subject nor object and are called avalent verbs (marked v.a. in the lexicon).

Verbs that make verbal phrases, i.e., stand-ins for intransitive verbs, from a following nouns or adjective are called 'copula' and marked 'v.cop.' in the lexicon. Syntactically, they are identical to transitive verbs if a noun follows, but adjectives may also directly follow, which is not the case for transitive verbs. Also, these verbs have no passive voice and cannot be used intransitively, though an intransitve homonymic verb may exist and will have a separate lexicon entry.

There also exist reflexive verbs that act similar to a normal copula, but do not allow nouns to follow, but only adjectives. These are marked as 'v.refl.cop.'.

Finally, auxiliary verbs are marked 'v.aux.' in the lexicon. These verbs can directly precede another verb to modify it.

scriu [skri͡ʊ̯] v.t. to write Ti scriu lebra. [tis kri͡ʊ̯ ˈlebrɐ] You write a book.
nat [nat] v.i. to swim No nat. [no nat] We swim.
mou [mo͡ʊ̯] v.refl. to move (by itself) Ap mou se. [ap mo͡ʊ̯ se] The water moves.
nivic [nɪˈvɪk] v.a. to snow Aoi nivic. [ɐˈo͡ɪ̯ nɪˈvɪk] It is snowing today.
diveni [dɪˈvenɪ] v.cop. to become Mi diveni frot. [mi dɪˈvenɪ frɔt] I become strong.
senti [ˈsɛntɪ] v.refl.cop. to feel Senti mi mal. [ˈsɛntɪ mi mal] I feel bad.
ro [ro] v.aux. to want Le ro e frau. [le ro e fra͡ʊ̯] He/she wants to be a blacksmith.

In many cases, transitive verbs can be used intransitively, by just dropping the object. E.g. scriu 'to write' can be used both as mi scriu lebra 'I write a book' and as mi scriu 'I write (something'). In this general case, the subject is semantically the same in the transitive and intransitive verbs.

Many reflexive verbs simply express that the object is the same as the subject, i.e., the subject acts upon the object. E.g. Lau mi. 'I wash myself.'

Reflexive verbs may also express an effect on the subject of the verb when the subject is semantically not an agent, but the patient of the homonymic transitive, usually causative, verb. E.g. for a transitive verb that implies causation like mou 'to move s.t.', the reflexive verb mou se means 'to move (by itself)', literally 'to move oneself', i.e., the effect of moving works on the subject, although the subject of mou se is not the semantical agent. These verbs are called uncausative verbs, and they are reflexive in Tirkunan quite consistently.

On the other hand, there are pairs of homonymic transtive and intransitive verbs, where again the subject of the intransitive verb corresponds with the object of the transitive verb, i.e., it is not the agent. If a (causative) effect is missing, then these intransitive verbs are usually not reflexive. E.g. flaga v.t. 'to sense the smell of' vs. flaga v.i. 'to emit smell'. The transitive verb is not causative here, so the intransitve one is not uncausative, so no reflexive se is used here. The reflexive verb, in these cases, corresponds with the transitive verb, e.g., flaga se 'to sense the smell of oneself'.


From the verbs we see today in Tirkunan, it appears that Old Tirkunan often used frequentative verbs instead of the original Classical Latin verbs, e.g. we have vis, probably from vīsāre instead of *vid from vidēre.[11] Many Latin verbs have not survived into Tirkunan, often only the compound verbs survived. Probably for this reason, Tirkunan has a tendency to preserve the compound stem of verbs rather than the original stem.[12] From the compound forms, the isolated verb was often reconstructed by analogy (e.g. *ficere instead of facere), so we even find many original compound stem in isolated verbs in Tirkunan.


To model what has happend during the history of Tirkunan, I use the following simulation in the hope that the result is a plausible selection of verbs. The model has to select verb stems, apply the sound changes and do simplifications of the morphological endings.

Verbs usually derive from either the present stem or the frequentative, so we compare the present stem and the frequentative stem (which is formed by -(i)tāre and often is equal to the supine stem: c.f. vidēre, vīsvmvīsāre but agere, āctvmagitāre). Deponent verbs always derive from the supine stem (from the perfect passive participle).

The construction described seems to yield a set of verbs that looks plausible. The history of Tirkunan verbs should currently be treated like a black box: there is no internal explanation (yet) which verbs derive from the frequentative and which from the original, plain present tense. Maybe it was chaotic in conhistory. The set of rules is used by me, the conlanger, to have an automated method to produce a nice-looking result. For now, let the history of the language be accidentally such that this construction and the historical language development yield the same results. The construction is meant to produce a plausible set of verbs, so if you think the resulting set of verbs looks implausible, please tell me.

Now for the construction: first of all, very short verbs and very irregular verbs will need manual care. By 'short' I mean verbs whose stem does not end in a consonant, e.g. stāre. Often, the plain stem can be used for sound shifting (e.g. to produce sta). The irregular verbs to be handled manually would be esse, ferre, velle, īre, fieri and all compounds thereof. Manual care may also mean that it is decided that the verb did not survive.

For the remaining verbs, if the frequentative verb is a regular extension (+ optional vowel + t) of the original stem, neglecting a possible stem vowel change and neglecting a possible drop of stem-final glides, and if no consonants fuse (for vidēre, use vis-, not vid-), then the original present stem is used including a potential vocalic/glide ending.

Otherwise, the frequentative stem is used. Sometimes, the regular -t ending (if still present and not fused) is stripped off together with a possible vowel in front of it. (This rule is still experimental – a close look at the results will be taken for evaluation.)

Further, we assumed a tendency to reconstruct the isolated verbs from compounds by analogy. So as the stem vowel, the reduced vowel from compound stems is usually used to derive the forms in Tirkunan (for facere, use fic-, not fac-), but we're free to decide that this would be quite implausible. To find a good choice, other Romance languages will have to be checked.

To the resulting stub, an /-iə/ is added for the i-conjugation, and an // otherwise, assuming a 3sg.pres.ind.act. form that dropped its -t -- the vowel is insignificant and will be reduced to schwa anyway, so equally well, the thematic vowel could be used. This is then sound-shifted with the GMP. A possible glide before that eding vowel will surface as /-u/ or /-i/, and a thematic /-i/ will be retained also. Some consonantal endings will retain the schwa, which will surface as /-a/. In total, this will produce many recognisable base forms and lead to the perfect participle often having the same vowel as the stem ending in Latin and so to 'sound right' (the simplification will take its toll, though).

Often, prefixes are sound-shifted separately so that compound verbs are compounded in Tirkunan, too, if it is plausible that it was clear to the speaker of (Old) Tirkunan that the prefix was indeed a prefix not part of the stem. Such prefixes are regularised by analogy with other verbs.

You can experiment with these rules here.

Some examples:

present stem Tirkunan
infinitive isolated compound supine stem input for GMP base part.perf.pass.
amāre am am-āt *am-ə am amat
solvere solv sol-v̄t *solv-ə solu sulut
avdīre avd avd-īt *avd-iə odi udit
avdēre avd avs *avs-ə os usat
vidēre vid vīs *vīs-ə vis visat
facere fac fic fec-t *fic-ə fi fit
discvtere qvat cvt cvss *cvss-ə discus discusat
inqvirere qvaer qvir qvis-it *qvis-ə ncis ncisat


some indef.
single ce nisun catun racun cium cul cul ca cul la
can tut rapan canum puc mbrut tan tan comu ca tan comu la
quality cal nisu' tal catu' tal racu' tal calum tal tal comu ca tal comu la
manner comu ni nisu' manir ni catu' manir ni racu' manir cumum simu comu ca comu la
place ou ni nisu' loc ni catu' loc ni racu' loc uvum la ca la
source d'ou di nisu' loc di catu' loc di racu' loc d'uvum di la di ca di la
dest. ad ou a nisu' loc a catu' loc a racu' loc ad uvum a la a ca a la
time candu mai tutur a racun ur laur acur laur
event cal vic nisu' vic catu' vic racu' vic calum vic cul vic cul ca vic cul la vic
repetition tut vic
reason pi nisu' rati pi catu' rati pi racu' rati pi cium pi puc rati pi mbrut rati
order li di can li di tan

The table conflates the interrogative and the relative usages, because they are always the same words in Tirkunan. This is the same as in the translations in that column.

The table also conflates the indefinite relative and plain versions, because Tirkunan uses the same words for them. This requires two translations, which are usually given and separated with a semicolon. The relative one contains a question word and is usually given first, e.g., 'whoever; someone'.

Indefinite words (pronouns and adverbs) are regularly formed by suffixing -um to the plain relative word, corresponding to appending the word um 'ever', the indefinite adverb of time.[13] This is counted as derivation, because the result is one word with single stress, e.g., cium 'whoever; someone' and uvum 'whereever; somewhere'. Note that indefinite relative and plain versions need two translations, but are a single lexicon entry in Tirkunan.

Indefinite words are often used instead of a phrase from the some column to replace a more complex prepositional phrase when there is no single generic word. Note that infinite word can also be used as relative word, but words from the 'some' column cannot.

Words expressing 'other' are regularly formed by prepending the word alt, most frequently to indefinite words, although to can also be used with the plain relatives, e.g., alt candu 'when else; at another time' or more often alt candum 'when ever else; at some other time'.[14] Linked to the near complete loss of number (except in articles), Tirkunan usually does not distinguish between amount and count, hence 'number' and 'quantity' are on the same row. E.g. a countable noun tan vic is translated as plural 'so many times' while an uncountable tan ap is 'so much water'. These words can also be applied to adjectives and adverbs: tan gran 'so large' and tan suvin 'so often'.

Tirkunan usually does distinguish amount vs. count in the 'single' column, where all words except nisun imply counting. When used with an uncountable noun, portions are counted: racun ap 'some glas/bottle/portion/body of water' vs rapan ap 'some amount of water'. Also ce crivis 'which glas/bottle of beer' vs. cal crivis 'which kind/type/brand of beer'. nisun does not enforce countability: nisun ap 'no water' is perfectly fine for amounts of water and equivalent to nint ap or nul ap (which is somewhat rarer).

The 'reason' row has prepositional, analytical expressions for each column, and it can be used as a template for other noun based constructions when a single word or special term is missing.

The 'source' and 'destination' rows also show how prepositions are used. This can be applied to other prepositions, too, like dis 'starting from', fin 'up to', tra 'between, inside'.

There is no special set of words for handling two things or persons. E.g., 'both' is lur du. If this is to be expressed, then some version of di lur du is used, e.g., ce di lur du 'who of the two'.

Tirkunan can specialise questions by using nouns, e.g., cal cus 'what, which thing' vs. cal gen 'who, which person'. This works with question words and also with nisun, catun, racun, rapan, cal, tal, cul, ..., e.g., tal cus 'such a thing' or di cul uring 'from this origin'.

There is a difference between ce 'who, what' vs. cal 'which' and correspondingly between cul 'this/that (one)' vs. tal 'such a': ce, cul are generic while cal, tal select from a specific group or imply some selection criterion. This is reflected in the translations 'who' vs. 'which'. Often, this distinction is blurry, especially if a word of selection follows, in which case usually cal, tal is preferred although ce, cul may be correct, too: di cal uring 'from which origin' would strictly ask to select from some group of origins already mentioned or ask to give a specification of the kind of origin, while di ce uring 'from what origin' would just generically ask for some origin. In practice, the former is used regularly in both semantic cases.

Related Small Words

Tirkunan has a few more words that do not fit the categories in the previous section, but should be mentioned here.

singra [ˈsɪŋrɐ] < singvlōs 'one each' Rei cu spus sir ni singra tron. The king and his wife each sit on a throne.
singra implies plural meaning, so singra tron is more than one throne, which may be confusing as the word is cognate to 'single'. But it derived from the Latin plural form which could already be used in the modern meaning, and the word was probably retained because Tirkunan needed some expressiveness for singular/plural distinctions in a few places when the plural endings disappeared. Compare Spanish 'sendos'.


Derivation in Tirkunan is a regular agglutination process. The following is an overview of some of the derivational suffixes.

noun > adj. -an cresi church > crisian ecclisial
adj. > emphatic -isma gran large > granisma huge
adj. > abstraction -(i)tat pusabra posible > pusabritat posibility
verb > agent -(a)tur am to love > amatur lover
verb > patient -(a)t posi to put > pusit put
verb > ability -abra pos be able > pusabra possibly
verb > abstraction -((a)t)i vis to see > visi vision
verb > event -i cumint to begin > cuminti beginning

The rules of elision and fusion are applied according to the lexicon.

Here, -(at)i drops the (at) after single -s (not in clusters).

Initial a of verb endings drops after -u and -i, e.g. solu+at > sulut, udi+at > udit, so the resulting participles look similar to the Latin ones, keeping the thematic vowel. This is not always the case, as pusabra shows, where the original consonantal stem becomes pos without -i.

Some endings are often used in city, country, mountain, river, etc. names:

noun > noun -is often found in cities Tircunis Tarragona

Note that final -a is dropped when adding a vocalic ending, but final -i or -u are not. However, two identical vowels are collapsed into one. Some endings drop the initial vowel after -i and -u.

miracra + us > miracrus miraculous
istra + an > istran insular
mpiri + an > mpirian imperial
os + at > usat dared
odi + at > udit heard; listened
solu + at > sulut solved
solu + ati > suluti solution
solu + atur > sulutur solver, solvent
selu + ac > silvac wild, savage


General Word Order

Tirkunan is mainly SVO except in special cases discussed below.

In contrast to other Romance languages, pronouns have no special role in ordering, i.e., object and oblique pronouns are put after the verb where also a noun would go: Mi am ti. 'I love you'.

Tirkunan is pro-drop for subjects, so the subject pronoun is often dropped. Object pronouns cannot usually be dropped except maybe in very short answers to questions or other special circumstances.

Some verbs do not have a subject at all, so none surfaces: Prou. 'It rains.'.

Oblique objects are placed after the direct objects: Mi da lebr' a ti. 'I give the book to you.'. For stylistic reasons or for stressing, they can be fronted: A ti mi da lebra! 'To you I gave the book!'. Oblique objects generally behave like adjuncts (see following section).

Auxilaries precede the verb: Mi ro manc 'I want to eat'.


Most modifiers follow the noun: prepositional constructions: mpiri ni Lustani, adjectives: mpiri gran, participles: pan mancat, relative clauses: mpiri ce e ni Lustani, possesives: patra mi, just like prepositional possessives: patra di mi.

In contrast to other Romance languages, adjectives cannot be fronted.

Adverbs to a verb can go at the beginning of the sentence, or after the verb: Aoi mi va a Lustani 'Today, I go to Lustany' or Mi va aoi a Lustani 'I go to Lustany today'.

Adverbs to adjectives and nouns precede the modified phrase: comu cat 'like a cat', mbrut gran 'very big'.

Relative Clauses

Relative clauses are introduced with the relative pronoun ce or with a prepositional phrase with that pronoun: di ce.

Relative clauses follow the modified noun phrase: Gen ce beu crivis e tiul mi. 'The person who drinks bear is my aunt/uncle'.

The relative clause itself uses a modified word order in that the reference pronoun ce is fronted and is not mentioned later in the sentence in the position it would normally appear. I.e., Tirkunan uses a gap construction instead of a resumptive pronoun. In the following sentence, the gap is marked with X (which is not pronounced): Gen ce te vis X e tiul mi 'The person you see X is my aunt/uncle'.

Relative clauses may be extrapositioned across adjuncts or even full clauses: Gen ni cambra ce beu crivis e tiul mi 'The person in the room who drinks beer is my aunt/uncle', or in some cases, Gen e ni cambra ce e tiul mi 'The person who is my aunt/uncle is in the room'. Note that for extrapositioned relative clauses, it is possible that another noun precedes ce to which ce does not refer: ...cambra ce..., while the reference is gen. This leads to syntactic ambiguities which are resolved only by semantics. Such constructions are generally only used if the context and the stressing disambiguates the construction clearly.


Conjunctions are used to join to sentences, either on the same level, or in main clause + subordinate clause relationship. Conjunctions start a conjoined clause: Mi ro vol sinu nu pos. 'I want to fly but can't.'.

The (more) ... the (more) ...

The construction "the X the Y" is expressed in Tirkunan as pru X, pru Y. Since pru means more and usually a comperative follows, no additional word to indicate the comperative is used. The comperative is separated in Tirkunan so the normal word order can be used.

Pru ti prot, pru ti da. The more you carry the more you give.
Pru le e vecra, pru le e grau. The older he is, the more serious he is.

The One Who ...

The construction "the one who X Y" can be expressed in two ways in Tirkunan: cul ce X, Y or cul Y ce X. The second form can be analysed as a relative clause extraposition, and again, it may cause semantic difficulties if Y ends in a noun phrase.

Cul ce ariu prim, sa cumpis. The one who arrives first shall win.
Cul ce ariu prim, sa cumpis joc. The one who arrives first shall win the game.
Cul sa cumpis ce ariu prim. ?That one shall win who arrives first.
?Cul sa cumpis joc ce ariu prim. ?That one shall win the game who arrives first.


FIXME: This chapter is missing. It is all done intuitively like in typical Romance languages...


Generally, whitespace is inserted after . ! ? , ; :, but not before.

For whitespace around apostrophes, please refer to the elision rules.

One Letter Words

This section lists some words that reduce to only one letter in some phonological contexts.

a ad
d di
i i
c ci
l li
n ni
p pi
u u, un


Given Names

Aemilia Mili
Alexandra, Alexander Lisandra
Aloysius Alesi
Ambrosius Mbrosi
Antonia, Antonius Ntoni
Arturius Ratul
Benedicta, Benedictus Bindic
Caesar Cestra
Caietana, Caietanus Caitan
Cecilia Ciceli
Christina Cristin
Christus Crist
Claudia, Claudius Crodi
Clemens Criment
Dulcia Droci
Felix Filic
Flora Flur
Francisca, Franciscus Francisc
Gaia, Gaius Gai
Hadriana, Hadrianus Drian
Helena Ren
Henrik Ndric
Hieronyma, Hieronymus Irom
Isabella Savel
John Iuan
Jake Iacou
Jesus Isu
Julia, Julius Iuli
Judia Iudi
Clara Clar
Laura Lur
Marcus Mbrac
Marcella, Marcellus Mbracel
Maria, Marius Mari
Martina, Martinus Mbratin
Maximilianus Masimblan
Michaela, Michael Micil
Paula, Paulus Pul
Philippa, Philippus Filip
Richarda, Richardus Ricalda
Sibylla Sivil
Stephana, Stephanus Stefra
Suetonius Svitoni
Theophila, Theophilus Tiufil
Tulia, Tulius Toli
Victoria, Victorius Vitori
Vincentius Vincenti
Walter Balteri
Werhard Biraldi
Wernhard Brinaldi


Surnames in Tirkunan are often combinations of a verb and an object or adject. Like in other romance languages, such nouns have a typical verb-noun form: Mumvrucan 'volcano mounter, lit. mount-volcano' (compare French cassegrain 'graincracker').

Often the fused proper noun will show additional elision or fusion or morpheme/word variants, like Umimpreti from Umint+preti or Saltinlac from salt+ni+lac. The fused proper noun will bear only one stress at the end (last or penultimate syllabel), resulting in all vowesl before that to collapse to a i u as for any unstressed syllables, like Numampan from nu+manc+pan.

Other typical surname formations derive from nouns using -us to express 'the one who is like...'.

Further, -ari to express 'the one who believes in/practices/works with/deals with/likes...'. Apart from names, this ending is often used for the dealer of the suffixed thing. Accordingly, any profession or generic agent may be a surname, too. These often end in the same -ari, or in the simple agent ending -atur. These verb based forms carrying a nominaliser cannot incorporate any object or adject -- only the non-suffixed verbs can do that.

Adjectives may be used as is for surnames. Also for adjectives, a -us ending, which may seem redundant, is frequent.

Tirkunan surnames are often very colourful, even borderline ridiculous to speakers of other languages.

Ritrunatur returner
Divatur who had to
Sucricatur who searched alone
Nascrutatur who does not listen
Umimpreti who raised the price
Sulutot solve-everything
Mumvrucan who mounts the volcano
Numampan doesn't eat bread
Udinisun who heard no-one
Avamar who went to the ocean
Saltinlac who jumped into the lake
Nucuntan not telling
Iacatinlac thrown into the lake
Sucricat the only one who was searched
Suprapasat surpassed
Pridat lost
Nupanus like there is no bread
Focus like fire
Miracus like an illusion
Ambilus like an eel
Cavalus like a horse
Cavalinus like a small horse
Centanus like a hundred years old
Racunus like someone
Marelus yellowish
Mitisalicari self-salter
Istatari liking the summer
Frumacari liking/dealing with yoghurt
Vecrivisari liking/dealing with old beer
Cipulinari liking/dealing with small onions
Casvari cheese monger
Baltur dancer


Pater Noster

Patra No [ˈpatrɐ no] Our Father
Patra no, ce e ni cel, [ˈpatrɐ no ke e ni kɛl] Our father, who is in heaven.
Numbra ti sa bindicat. [ˈnʊmbrɐ ti sa bɪndɪˈkat] Your name be hallowed.
Reni ti sa veni. [ˈrenɪ ti sa ˈvenɪ] Your kingdom come.
Roi ti sa fit, [ro͡ɪ̯ ti sa fɪt] Your will shall be done.
Comu ni cel simu ni ter. [ˈkomʊ ni kɛl ˈsimʊ ni tɛr] How in heaven so on earth.
Va dun aoi a no pan pi catu' di. [va dʊ nɐˈo ͡ɪ̯a no pam pi kɐˈtu di] Give us today our bread for each day.
I va pidun a no divat no, [i va pɪˈdʊ na no dɪˈvat no] And forgive us our debts.
Tan comu no pidun le a divatur no. [taŋ ˈkomʊ no pɪˈdʊn le a dɪvɐˈtʊr no] Like we forgive them of our debtors.
I va nu nduc no ni tintati, [i va nun dʊk no ni tɪnˈtatɪ] And do not lead us into temptation.
Sinu va libra no di mal. [ˈsinʊ va ˈlibrɐ no di mal] But liberate us from evil.
Ca di ti e reni i pot i glori, [ka di ti e ˈrenɪ i pɔ ti ˈɡlorɪ] As yours is the kingdom, the power, the glory.
Ni trinitat, [ni trɪnɪˈtat] In eternity.
Amin. [ɐˈmɪn] Amen.


Mi am ti. [mi an ti] I love you.
No sa va! [no sa va] Let's go!
Bon matin! [bɔn mɐˈtɪn] Good morning!
Bon di! [bɔn di] Good afternoon!
Bon sir! [bɔn sɪr] Good evening!
Bon not! [bɔn nɔt] Good night!
Salut! [sɐˈlʊt] Hello!
Aur! [ɐˈʊr] Bye, bye!
Sa rest nigatiu! [sa rɛst nɪɡɐˈti͡ʊ̯] Stay negative!
A rivis! [a rɪˈvɪs] Good bye!
A pru trar! [a pru trar] See you later!
..., pi prac. [ pi prak] ... , please.
..., pi faur. [ pi fɐˈʊr] ... , please.
Mbrut ubligat! [m̩brʊ tʊblɪˈɡat] Thank you!
Nu pi cul! [nu pi kʊl] Don't mention it!
Sa nur mi. [sa nʊr mi] Please ignore me.

The Northwind and the Sun

Vent Bural i Sul [vɛnt bʊˈra li sʊl] Northwind and Sun
N'u di, vent bural i sul discus sopra ce di lur du e pru frot, candu ni cul mument, viacatur pas mvrutat ni mantil crar. [nu di vɛnt bʊˈra li sʊl dɪsˈkʊs ˈsoprɐ ke di lʊr du e pru frɔt ˈkandʊ ni kʊl mʊˈmɛnt vjɐkɐˈtʊr pas ɱ̩vrʊˈtat ni mɐnˈtɪl krar] One day, the northwind and the sun were discussing about who of the two is stronger, when in that moment, a traveller passed by wrapped in a warm coat.
Cuveni ci cul eri cusidrat pru frot ce ariu fi viacatur trai mantil. [kʊˈvenɪ ki kʊ ˈlerɪ kʊsɪˈdrat pru frɔt ke ɐˈri͡ʊ̯ fi vjɐkɐˈtʊr tra͡ɪ̯ mɐnˈtɪl] They agreed that that one will be considered the one stronger who manages to make the traveller take off the coat.
Vent bural cumint sofla cu tut pot, sinu pru le sofla, pru viacatur ten frir i stric se ni mantil, i ni fin, vent bural deu riseni. [vɛnt bʊˈral kʊˈmɪnt ˈsoflɐ ku tʊt pɔt ˈsinʊ pru le ˈsoflɐ pru vjɐkɐˈtʊr tɛɱ frɪ ris trɪk se ni mɐnˈtɪl i ni fɪn vɛnt bʊˈral de͡ʊ̯ rɪˈsenɪ] The northwind began to blow with all strength, but the more he blew, the more the traveller felt cold and nuzzled into the coat, and in the end, the northwind had to give up.
Laur, sul cumint bril ni cel, i pruntament, viacatur rep crar i trai mantil. [lɐˈʊr sʊl kʊˈmɪnt brɪl ni kɛl i prʊntɐˈmɛnt vjɐkɐˈtʊr rɛp kra ri tra͡ɪ̯ mɐnˈtɪl] Then, the sun began to shine in the sky, and immediately, the traveller got warm and took off the coat.
Acur, vent bural deu ricunus supriltat di sul. [ɐˈkʊr vɛnt bʊˈral de͡ʊ̯ rɪkʊˈnʊs sʊprɪlˈtat̚ di sʊl] Now, the northwind had to recognise the superiority of the sun.

Iuan 1:1

Iuan [jʊˈan] John
Ni princepi fu parul, i parul fu cu Deu, i parul fu Deu. [ni prɪŋˈkepɪ fu pɐˈrʊl i pɐˈrʊl fu ku de͡ʊ̯ i pɐˈrʊl fu de͡ʊ̯] In the beginning was the word, and the word was with God, and the word was God.
Le fu, ni princepi, cu Deu. [le fu ni prɪŋˈkepɪ ku de͡ʊ̯] It was, in the beginning, with God.
Tot e fit pi le, i sinti le, nint e fit ci e fit. [tɔ te fɪt pi le i ˈsɪntɪ le nɪn te fɪt ki e fɪt] All is made for him, and without him, nothing is made that is made.
Vit fu ni le, i vit fu luc d'ombra. [vɪt fu ni le i vɪt fu lʊk ˈdɔmbrɐ] Life was with him, and life was the light of man.
Luc bril ni tembra, i tembra nu cuprin le. [lʊk brɪl ni ˈtɛmbrɐ i ˈtɛmbrɐ nu kʊˈprɪn le] Light shone in the darkness, and the darkness did not understand it.
Ombra fu mandat pi Deu, i numbra se fu Iuan. [ˈɔmbrɐ fu mɐnˈdat pi de͡ʊ̯ i ˈnʊmbrɐ se fu jʊˈan] A man was sent by God, and his name was John.


Gines [ɡɪˈnɛs] Genesis
Ni princepi, Deu fi cel i ter. [ni prɪŋˈkepɪ de͡ʊ̯ fi kɛ li tɛr] In the beginning, God made heaven and earth.
Ter fu vot i sinti from, i tembra fu sopra faci d'avis i sprit di Deu mou sopr' ap. [tɛr fu vɔ ti ˈsɪntɪ frɔm i ˈtɛmbrɐ fu ˈsoprɐ ˈfakɪ dɐˈvɪ sis prɪt̚ di de͡ʊ̯ mo͡ʊ̯ sɔp rap] The earth was empty and without shape, and darkness was above the surface of the abyss, and the spirit of God moved above the water.
Deu di: luc sa fit! I luc fu fit. [de͡ʊ̯ di lʊk sa fɪt i lʊk fu fɪt] God said: Let light be made! And light was made.
Deu vis ci luc fu bon i sipar luc di tembra. [de͡ʊ̯ vɪs ki lʊk fu bɔ ni sɪˈpar lʊk di ˈtɛmbrɐ] God saw that the light was good and he separated light from darkness.
Deu cram luc «di» i tembra «not». [de͡ʊ̯ kram lʊk di i ˈtɛmbrɐ nɔt] God called the light `day' and the darkness `night'.
I fu sir i fu matin: di un. [i fu sɪ ri fu mɐˈtɪn di ʊn] And it was evening and it was morning: day one.

Golden Rule

Regla d'Or [ˈreɡlɐ dɔr] Golden Rule
Ce nu ro fit a ti, nu fi a racun! [ke nu ro fɪ ta ti nu fi a rɐˈkʊn] What you don't want done to yourself, don't do to anyone.


Babil [bɐˈbɪl] Babel
N'u tem, tut ter ten sul limba i pal parul. [nu tɛm tʊt̚ tɛr tɛn sʊl ˈlɪmbɐ i pal pɐˈrʊl] Once upon a time, all earth had a single language and the same word.

Hovercraft of Eels

Nau cusin ar mi e prin d'ambil. [na͡ʊ̯ kʊˈsɪ nar mi e prɪn dɐmˈbɪl] My hovercraft is full of eels.


Prandi sir i ser cric [ˈprandɪ sɪ ri sɛr krɪk] Dinner and Circular Saw
Bon di! [bɔn di] Good Afternoon!
Ni cul sir Oda i mi au mancat past farat di fung cu pestu i singra burata. [ni kʊl sɪ ˈrodɐ i mi a͡ʊ̯ mɐŋˈkat past fɐˈrat̚ di fʊŋ ku ˈpɛstʊ i ˈsɪŋrɐ bʊˈratɐ] This evening, Uta and I ate pasta filled with mushrooms with pesto and a burrata each.
Si, puc mbrut, sinu past fu frisc di mbricat, pestu casifit fu ni mfriratur, i burata ni cumbra e tutur bon. [si pʊk m̩brʊt ˈsinʊ past fu frɪsk dim brɪˈkat ˈpɛstʊ kɐsɪˈfɪt fu niɱ frɪrɐˈtʊr i bʊˈratɐ ni ˈkʊmbrɐ e tʊˈtʊr bɔn] Yes, a bit much, but the pasta was fresh from the market, home-made pesto was in the fridge, and burrata on top is always good.
Ni rapan di pasat mi trival a ser cric tabra mi. [ni rɐˈpan di pɐˈsat mi trɪˈva la sɛr krɪk ˈtabrɐ mi] During the past few days, I have been working on my table saw.
Tabra e pres finit i se pos opra ser. [ˈtabrɐ e prɛs fɪˈnɪ ti se pɔ ˈsoprɐ sɛr] The table is almost done, and you can operate the saw.
Aoi mi au fit per pi le. [ɐˈo͡ɪ̯ mi a͡ʊ̯ fɪt pɛr pi le] Today, I made feet for it.
Vo sa rest nigatiu! [vo sa rɛst nɪɡɐˈti͡ʊ̯] Stay negative!
A pru trar, Ndric [a pru trar n̩drɪk] See you later, Henrik

Christmas Card Exchange 2009

The foreside reads 'Hjalri Nátli eð þælkt nó önn', which is Þrjótrunn, meaning 'Merry Christmas and a happy new year'. Our family is on holiday in Þrjótur, sending a card back to their friends at home in Tarragona, Lusitania.

Modern Version

Salut carisma ... [sɐˈlʊt kɐˈrɪsmɐ] Hello dear ...
Natal Filic i an nou pruspla di nort frir, [nɐˈtal fɪˈlɪ ki an no͡ʊ̯ ˈprʊsplɐ di nɔrt frɪr] Merry Christmas and a happy new year from the cold north,
di Friulter, u no cilebra Natal. Ca nivic [di frjʊlˈtɛr u no kɪˈlebrɐ nɐˈtal ka nɪˈvɪk] from Iceland, where we celebrate Christmas. Here, it snows
custantament i tut ciutat e prin d'Aret Sulan [kʊstɐntɐˈmɛn ti tʊt kjʊˈta te prɪn dɐˈrɛt sʊˈlan] constantly and the whole city is full of Sun-Rams
i di mfant cu froc. Fin acur, pi frutun, vis [i diɱ fant ku frɔk fɪ nɐˈkʊr pi frʊˈtʊn vɪs] and of children with scissors. Up to now, fortunately, we saw
nisu' lesi. Pi cilibrati se manc pisc [nɪˈsu ˈlesɪ pi kɪlɪˈbratɪ se maŋk pɪsk] no injury. For the celebration, people eat
pisc apistur. [pɪs kɐpɪsˈtʊr] stinking fish.
A pru trar, [a pru trar] See you later,
Oda i Ndric [ˈodɐ in drɪk] Uta and Henrik

Original Version

The language has changed since 2009. The original version was as follows:
Salu kar ... [ˈsalʊ kar]
Natal Filik id an nova pruspla di nort frida, [nɐˈtal fɪˈlɪ kɪ dan ˈnovɐ ˈprʊsplɐ di nɔrt ˈfridɐ]
di Friglater, ova nos kilebra Natal. Ka nivik [di frɪɡlɐˈtɛr ˈovɐ nɔs kɪˈlebrɐ nɐˈtal ka nɪˈvɪk]
kustantament i tot kiutat es prin d'Aret Sular [kʊstɐntɐˈmɛn ti tɔt kjʊˈta tɛs prɪn dɐˈrɛt sʊˈlar]
i di nfant ku frok. Fin akur, pi frutun, vis [i diɱ fant ku frɔk fɪ nɐˈkʊr pi frʊˈtʊn vɪs]
nisu' lisiun. Pi kilibratiun se manga pisk [nɪˈsu lɪsˈjʊn pi kɪlɪbrɐtˈjʊn se ˈmaŋɡɐ pɪsk]
apistan. [ɐpɪsˈtan]
A pru trada, [a pru ˈtradɐ]
Oda id Indrik [ˈodɐ ɪ dɪnˈdrɪk]

Hjalri Nátli! There are some minor variations to the text sometimes -- I wrote it by hand and sometimes changed a few words slightly (e.g. nu vis racu' lisiun ('we did not see any injury') instead of vis nisu' lisiun ('we saw no injury')).

When addressing a group of people (e.g., a family), I used Salu lis ..., where lis is the definite article in plural, which is used for vocatives (as in French usage).

History of Sound Changes

This section lists the key aspects of the Tirkunan sound shifts that derived words from Vulgar Latin into Tirkunan.



Obviously from Latin Lv̄sitānia, which was in the area of today's Portugal in Roman times. Some things seem to have gone differently, since 'Tarragona' is not in Portugal here, but in Spain, in Catalonia. Also, obviously neither 'Portugal' nor 'Spain' is named after the Latin Lv̄sitānia, but 'Lustani' seems to be a country there. And also, the city seems to be larger than here, provided that 'Tirkunis' is really our 'Tarragona', i.e. Tarracō in Roman times.
This is like in Latin, e.g. rē + dare becomes reddere, stressed on the antepenult (this verb also completely changes conjugation), and similarly with ab + ferre, which becomes avfere also stressed on the antepenult. This is unlike Italian, which has ri + dare as ridare with the 3.sg. form ridà, stressed on the last syllable, as Italian does not move stress to the prefix.
This is like in Latin: once a prefixed stem changes, it stays that way, e.g., in reddere, the a of the original stem of dare does not reemerge in conjugated forms.
The strong simplifications of pronouns and case system of Tirkunan can be observed in other Romance languages, too. E.g. the collapse of cases in pronouns into the accusative in ego, tv can be found in Ligurian, Piemontese, Venetian, and Lombard, which have 'mi, ti', and to a lesser degree and into the nominative in Catalan, which has 'jo, tu' (and also 'mi' in some contexts); accidentally, Ligurian singular pronouns are virtually identical to Tirkunan: 'mi, ti, lê' vs. 'mi, ti, le'.
Like in Romanian.
Similar to Catalan that likes to add -s to adverbs: 'llavors', 'fins', 'doncs', 'abans'.
Like in Sardinian.
Cognate to Italian: mo' 'now'.
Like Rumansch: vegn a offrir 'he will offer'.
Cognate to French: est-ce que.
Compare Italian 'acquistare', 'conquistare', 'diventare', 'voltare', and Catalan 'gausar', French 'oser', Portuguese 'ousar', Spanish 'osar', Italian: 'osare' < avdēre, avsvm. If Tirkunan has creole-like traces, English verbs might also be relevant for comparison as many are loaned from the Latin supine stem: 'discuss', 'prevent'.
Compare Romance here: Qvatere has not survived, but discvtere has survived in Italian 'discùtere' and Spanish 'discutir'
Like Italian '-unque', and similarly to Italian '-siasi', Romanian '-va', Spanish '-qier(a)', Portuguese '-quer', Sardinian '-sisiat, -casiat'.
Like in Romanian, and also like Italian 'altrove'.


November 7th, 2021
Comments? Suggestions? Corrections? You can drop me a line.