The first obstacle is /q/. As far as I know, most languages using the Cyrillic script tend to add little hooks to letters to create new ones, so I /q/ would be something like қ ҝ ҟ ҡ. But that looks a bit too indistinct from plain к, I think. (Similarly for ң for /ŋ/ – though ҥ and ӈ look more distinctive, IMO.)

So I thought about using digraphs instead, inspired a little by old orthographies for Inuit languages as well as the writing conventions for Syllabics.

Here’s what I came up with:

Vowels – a = а, i = і, u = у; aa = аа, ii = іі, uu = уу; ai = аі, au = ау, ia = іа, iu = іу, ua = уа, ui = уі.

Consonants – voiceless: p = п, t = т, s = с,  ł[5] = дл, k = , q = [1]; voiced: b[2] = б, v = в, l = л, j = я/ї/ю[7], g = г, r[3] = р; nasal: m = м, n = н, ŋ = нг[6], ɴ[4] = р.

Notes:

1. кр- at the beginning of a syllable, -рк at the end of a syllable, -ркк- when geminate, -р- as the first part of a consonant cluster.
2. occurs only as the first part of a consonant cluster, e.g. mibvik = мібвік, kublu = кублу.
3. As I mentioned elsewhere, the realisation of this phoneme (in particular, its manner of articulation: stop, fricative, or continuant) varies.
4. Not a separate phoneme; only occurs as the first part of a consonant cluster.
5. Only occurs as the second part of a consonant cluster, or geminate (then spelled simply дл); never as the first part of a cluster or by itself.
6. ннг when geminate.
7. I’m not quite certain yet what to do with geminate -jj-. Perhaps I’ll take advantage of the fact that this sometimes has a fricative or affricate pronunciation (compare the spelling “dj” rather than “yy” or “jj” in Inuinnaqtun) and write this combination as -жя/-жї/-жю.

To reduce the number of -рк- sequences, I decided to write a uvular place of articulation as the first part of a consonant cluster with a fixed -р-, regardless of the manner of articulation; this would be signalled unambiguously by the second member of the consonant cluster except in the case of a geminate voiceless uvular stop. This bit is a bit like Nunavik orthographical conventions, as I understand them.

An alternative, with slightly fewer digraphs, is to use Serbian Cyrillic њ for /ŋ/ and љ for /ɬ/; it would then seem attractive to use ј for /j/ (and -јј- for geminate /jj/), and since this gets rid of ї, we can use и for /i/. The only digraph then would be for /q/. I’ll admit that this leads to rather unusual values for the letters њ and љ.

I suppose it might even be possible to get rid of that final digraph by using х for /q/, as in Greek-script Ilatamiutitut. (Or ԛ, but font support for that is likely to be bad.)

That would lead to the following two spellings of the Lord’s Prayer as reproduced in the article on Greek-script Ilatamiutitut:

“Russian” –

Атаатавут крілангміітутіт: ангінірпангунііт ікпігіяуттіарлі;

атаніувііт краілаурлі; піюмаяіт атуртаулі нунамі суурлу крілангмі атуртаунгмат.

Ублумі нікріксаптіннік тунісівітігут.

Таммарніптіннік ісумакрарквінгіюнгніірлута саіммаутітігут, суурлу інуукратіпта уваптінгнут таммарнгнінгіт саіммаутіяраангаптігут ісумакрірквігіюннірпак каттігут.

Ууктуртаутіттаілітігут аюртаптігут, сапутітігублі саатаанасіміт. (Атаніунірк, аюгакрангірнірлу, інукранірюарнірлу пігігавігіт, ісукрангітумут.)

“Serbian” –

Атаатавут хилањмиитутит: ањинирпањуниит икпигијауттиарли;

атаниувиит хаилаурли; пијумајаит атуртаули нунами суурлу хилањми атуртауњмат.

Ублуми нихиксаптинник тунисивитигут.

Таммарниптинник исумахахвињијуњниирлута саиммаутитигут, суурлу инуухатипта уваптињнут таммарњнињит саиммаутијараањаптигут исумахихвигијуннирпак каттигут.

Ууктуртаутиттаилитигут ајуртаптигут, сапутитигубли саатаанасимит. (Атаниуних, ајугахањирнирлу, инуханирјуарнирлу пигигавигит, исухањитумут.)

I thought I’d try to use that to create words for Ilatamiutitut, so I had a go at creating patterns that would sound like Inuktitut.

I started by taking all the headwords from Spalding’s Multi-dialectal Outline Dictionary (with an Aivilingmiutaq base) that were nouns and verbs, trying to omit clearly-derived ones. (This was also the reason for only taking lemmata at the beginning of paragraphs, since lemmata inside paragraphs were usually derived from the first one. But this is not always the case; going through and adding more by hand might make for a more balanced “corpus”, at the expense of a lot of additional work, given the size of the dictionary.)

Then I wrote a script to analyse the nouns for things such as the distribution of initials, finals, medial consonant clusters, number of syllables, and the like.

After that, I turned the results into a set of rules for Awkwords and gave the script a spin.

I was fairly satisfied with many of the results; they do look fairly Inuktitut-like, and I even spotted a few of the words I already knew, such as nanuq or sila, amongst the results.

Here’s a run of 16 items I did just now: asanuuq – puqqulumukłuq – tuamiptamaq – iritiq – paaq – ajijaq – ulit – nuqqi – kivaqiq – aak – mimat – isaaluq – utakpiq – maqsaapi – sirlapatiq – tasaq. What do you think?

Since the dictionary I used a base is based on a phonologically fairly conservative (western) dialect, there are some consonant clusters that you won’t see in more eastern dialects of Inuktitut such as North Baffin or South Baffin, let alone Greenlandic, where consonant clusters such as -pt- tend to get assimilated. (For some details on such assimilation in various dialects, see Section G of Inuktitut Linguistics for Technocrats.)

After all, if they had influence from Greece, then perhaps they would write the language in (a variation of) the Greek alphabet (rather than, say, in Syllabics or in Roman).

On the other hand, while Unicode has a whole lot of precomposed Roman letters with diacritics, and even a fair number of Cyrillic letters with diacritics, there are next to no Greek letters with diacritics, so font support for Greek letters with diacritics is likely to be poor. So for OOC reasons (I’d like the result to look nice given the fonts I have), I’d probably have to stick to the basic Greek alphabet.

(Historically, I suppose the reason for the lack of letters is the fact that the Greek alphabet wasn’t used much for other languages – and especially not as the official orthography of a language after the day of computing. For example, the use of Greek to write Albanian is quite a while ago, and Greek dialects aren’t written much except by specialists.)

Some of the phoneme-to-letter assignments seem fairly straightforward; others are more difficult.

For example, the three vowels /a i u/ most obviously map to α ι υ, respectively. (Or, theoretically, α ι ου, but having one vowel be a digraph, especially when neither ο nor υ will be used by itself, seems ugly. Though I suppose that didn’t stop the French from using “ou” to transcribe Arabic /u/.)

Long vowels could be αα ιι υυ, like in Roman-script Inuktitut. Or possibly some diacritic, perhaps even the tone: ά ί ύ. Or, I suppose, a macron ᾱ ῑ ῡ or circumflex ᾶ ῖ ῦ. (Here, it comes in handy that α ι υ are the vowel letters that could be either short or long in Ancient Greek.)

The nasals /m n/ are fairly obviously μ ν, respectively, and the voiceless stops /p t k/ are π τ κ. Also, the voiceless fricative /s/ will presumably be σ, and the voiced lateral /l/ will be λ.

That’s about where the easy bits end.

For /q/, it’s tempting to use χ, since that used to be /kʰ/ and is nowadays /x/, so a kind of dorsal consonant (albeit still a velar, rather than uvular, one).

An alternative would be not to differentiate between /k/ and /q/ (which is what many Europeans did at first when writing down Eskimo-Aleut languages), but that seems unsatisfactory to me.

Another alternative is inspired by the Syllabics orthography: introduce a digraph along the lines of “rk”, with the first letter showing the place of articulation (uvular) and the second showing the manner of articulation (voiceless stop).

But that runs into problems with gemination; you’d need to decide whether to write “rkrk” or “rrk” (syllabics uses the latter route, and it parallels the use of “nng” for geminate /ŋ/ in the Latin orthography).

And that’s even before you run into the problem of how to represent the voiced uvular in the first place. (I’m not quite sure what its value is; in Greenlandic, my understanding is that it’s a voiced uvular fricative [ʁ] or possibly an approximant [ʁ̞], while in Canada, it seems to be a voiced uvular stop, [ɢ].)

Apparently, one reason why the letter “r” is used for this sound is that Samuel Kleinschmidt, when he drew up his orthography for Greenlandic, was influenced by Danish and/or German (his other languages), in which /r/ is often [ʁ]. And then this convention simply stuck and became used even in Canada, where the phoneme is a stop and where the local qallunaaq language had a completely different typical sound value for “r”.

But I’m not sure whether ρ would be a natural choice for Greek philologists coming across the Ilatamiut; their /r/ would presumably be an alveolar trill rather than anything uvular.

So I’m tentatively assigning ρ to the voiced uvular, but I’m not sure whether that will stay. (I’m also not sure whether in Ilatamiutitut, the voiced uvular will be a stop, a fricative, or an approximant. I’m leaning to one of the latter two, though; that is, a continuant.)

That leaves the voiced sounds /v j g/ (which are, apparently, all approximants in Greenlandic, so [ʋ j ɰ] rather than [v j g] or even [v j ɣ]; Wikipedia implies that /g/ is typically [g] in Inuktitut but may be [ɣ] “between vowels or vowels and approximants”, though the fricative pronunciation is said to be universal for Siglitun); the voiceless lateral /ɬ/, and the nasals /ŋ/ and [ɴ].

For the voiced continuants (which is how I’ll treat them), I’m tempted to use β γ, based on their modern Greek fricative sound values, for /v g/, but that leaves open the question of how to represent /j/; modern Greek would point towards γ again, but /ja ji ju ga gi gu/ all exist (as far as I know), whereas in modern Greek, you could only have γα γι γυ /ga ji gu/—or if you use a digraph, για γι γιυ γα γυ /ja ji ju ga – gu/, which still leaves a gap for /gi/ [ɣi ~ ɰi] that I’m not sure how to spell. So I’d prefer to have separate letters for /j/ and /g/. Maybe something completely off-the-wall for /j/ such as ζ or ψ?

For /ŋ/, the obvious Greek solution would be—yet again!—γ, which obviously won’t do. (Especially since /ŋ/ can occur at the beginning of a syllable in Ilatamiutitut.) Maybe a digraph νγ as in Latin Inuktitut? (And Syllabics, for that matter, though there the spelling is arguably even worse: a “ng” ligature in front of a “g”, with the geminate being “nng” ligature + “g”.) One good thing about the phonology is that “ng” would be unambiguous, since /ng/ can’t occur due to manner-of-articulation (MOA) assimilation.

The next sound, the uvular nasal [ɴ] is a special case; it’s not a phoneme in its own right and can’t start a syllable, but only occurs before other nasals. Wikipedia treats it as an allophone of /ʁ/ in that position (and that is also how it is spelled in Inuktitut orthography).

However, due to the pervasive MOA assimilation, one could do away with two separate letters and simply use one letter for the first component of clusters with a uvular first component; as far as I know, this is indeed done in Nunavik (in Quebec), where “r” is used in all cases, being automatically read as a voiceless uvular stop, a voiced uvular (stop?), or a voiced uvular nasal, depending on the following letter.

(This could theoretically also work for the labial place of articulation [POA], with—say—”mn mt ml” being read /mn pt vl/, respectively, and the velar POA, with—say—”gm gp gv” being read /ŋm kp gv/, respectively. A bit more difficult for the POA in between labial and velar since alveolar nasal and stop /n t/ can correspond either to alveolar continuant /s/ or to palatal continuant /j/.)

So. Even though this phone can’t occur syllable-initially, it would be nice to have three separate letters for the three manners of articulation of uvular sounds, but I don’t know what a good Greek letter for [ɴ] could be.

The final sound is /ɬ/; again, nothing obvious jumps to mind, though I lean towards using θ for it since both sounds are coronal fricatives. (And Welsh “ll”—which represents this same sound—sounds like “thl” to at least some English speakers.)

Another thought that came to me was to use υ ι to represent the approximants I write /v j/; then I’d have to think of some other way to write the vowels.

Unless I can convince myself that the spelling would be unambiguous. But looking through my Spalding, I see quite a few occurrences of /vv/, which would be ambiguous with /uv/. These are mostly if not exclusively North Baffin forms, according to the notes there, typically alternating with /bv/ or simply /v/ in other places. (For example, in the suffix “place where something is done”: (pi)vik, (pi)bvik, (pi)vvik.) But there are even more occurrences of /jj/ (including _ijji_ “foreign body in the eye”!).

So that could only work if I used different letters for the vowel phonemes—say, η for /i/ and ω for /u/, or something. (Or even something silly like α αα ε η ο ω for /a aː i iː u uː/.)

So with a tentative orthography, here’s an attempt at a transcription of the Lord’s Prayer in “Eastern Arctic Eskimo”:

Ατάταβυτ χιλανγμίτυτιτ: ανγινιχπανγυνίτ ικπιγιζαυττιαρλι,

ατανιυβίτ χαιλαυρλι, πιζυμαζαιτ ατυχταυλι νυναμι σύχλυ χιλανγμι ατυχταυνγματ.

Υβλυμι νιχικσαπτιννικ τυνισιβιτιγυτ.

Ταμμαρνιπτιννικ ισυμαχαχβινγιζυνγνίρλυτα σαιμμαυτιτιγυτ, σύχλυ ινύχατιπτα υβαπτινγννυτ ταμμαρνγνινγιτ σαιμμαυτιζαράνγαπτιγυτ ισυμαχιχβιγιζυννιχπακ καττιγυτ.

Ύκτυχταυτιτταιλιτιγυτ αζυχταπτιγυτ, σαπυτιτιγυβλι σάτάνασιμιτ. (Ατανιυνιχ, αζυγαχανγινιρλυ, ινυχανιρζυαρνιρλυ πιγιγαβιγιτ, ισυχανγιτυμυτ.)

Original:

ᐊᑖᑕᕗᑦ ᕿᓚᖕᒦᑐᑎᑦ: ᐊᖏᓂᖅᐸᖑᓃᑦ ᐃᒃᐱᒋᔭᐅᑦᑎᐊᕐᓕ;

ᐊᑕᓂᐅᕖᑦ ᖀᓚᐅᕐᓕ; ᐱᔪᒪᔭᐃᑦ ᐊᑐᖅᑕᐅᓕ ᓄᓇᒥ ᓲᖅᓗ ᕿᓚᖕᒥ ᐊᑐᖅᑕᐅᖕᒪᑦ.

ᐅᑉᓗᒥ ᓂᕿᒃᓴᑉᑎᓐᓂᒃ ᑐᓂᓯᕕᑎᒍᑦ.

ᑕᒻᒪᕐᓂᑉᑎᓐᓂᒃ ᐃᓱᒪᖃᖅᕕᖏᔪᖕᓃᕐᓗᑕ ᓰᒻᒪᐅᑎᑎᒍᑦ, ᓲᖅᓗ ᐃᓅᖃᑎᑉᑕ ᐅᕙᑉᑎᖕᓐᓄᑦ ᑕᒻᒪᕐᖕᓂᖏᑦ ᓰᒻᒪᐅᑎᔭᕌᖓᑉᑎᒍᑦ ᐃᓱᒪᕿᖅᕕᒋᔪᓐᓂᖅᐸᒃ ᑲᑦᑎᒍᑦ.

ᐆᒃᑐᖅᑕᐅᑎᑦᑏᓕᑎᒍᑦ ᐊᔪᖅᑕᑉᑎᒍᑦ, ᓴᐳᑎᑎᒍᑉᓕ ᓵᑖᓇᓯᒥᑦ.  (ᐊᑕᓂᐅᓂᖅ, ᐊᔪᒐᖃᖏᓂᕐᓗ, ᐃᓄᖃᓂᕐᔪᐊᕐᓂᕐᓗ ᐱᒋᒐᕕᒋᑦ, ᐃᓱᖃᖏᑐᒧᑦ.)

Ataatavut qilangmiitutit: anginiqpanguniit ikpigijauttiarli;

ataniuviit qailaurli; pijumajait atuqtauli nunami suuqlu qilangmi atuqtaungmat.

Ublumi niqiksaptinnik tunisivitigut.

Tammarniptinnik isumaqaqvingijungniirluta saimmautitigut, suuqlu inuuqatipta uvaptingnnut tammarngningit saimmautijaraangaptigut isumaqiqvigijunniqpak kattigut.

Uuktuqtautittailitigut ajuqtaptigut, saputitigubli saataanasimit. (Ataniuniq, ajugaqangirnirlu, inuqanirjuarnirlu pigigavigit, isuqangitumut.)

(There are undoubtedly errors, either in my transcription or possibly already in the source document; there are a few places where the combination of letters used looks doubtful to me. There are also a few marks where I’m not sure whether I interpreted them correctly; for example, whether some of the dots above are open circles instead and if so, whether they stand for “ai”, or what the occasional lone open circle in the middle of a word means.)

And since I’m an a posteriori sort of person, I’ll probably base it on Inuktitut, with probably some influence from (Modern) Greek. For example, I expect Greek will take the place that English does in Inuktitut, and Danish in Kalaallisut, such as as the source of loanwords for things like higher numbers. I expect there’ll be some sort of concultural explanation for why the Greeks ended up having linguistic influence on an arctic people, but OOC, it’s mostly that it’s a language I know a fair bit about and that is already the basis for a number of my previous conlang projects.

Depending on how creative I get, it may be fairly close to Inuktitut, a relex with the same phonology, a relex with different phonology, or merely inspired by the idea, using it as an example of what sort of things polysynthetic languages do. (I suspect that it’ll fall somewhere near the beginning of the list. Which may also have the positive side effect of teaching me more about Inuktitut, a language I wouldn’t mind knowing more about anyway. On the other hand, I might also end up learning a bunch of wrong things which I would have to un-learn later, if I wanted to learn Inuktitut seriously.)

The tentative code-name for the language is Ilatamiutitut, from Ilata “(name of the place where the Ilatamiut live)” < Greek Ελλάδα “Greece” + -miu(t) “inhabitants (of a place)” + -titut “(similative case ending: like …)”; that is, “(speaking) like the inhabitants of Ilata”.

Though the area where they live should probably have a better name than Ilata eventually; it seems rather unlikely to me to simply call it that, without at least a “New” qualifier. (But that would break the single-word language name. So I’ll have to see.)

CIT v1.1 !l cN:R:S:H a- y1 n36:2d B+ A+ E+ L N4 Im k– ia@ p* s* m* o+ P- S Engadinese

Now, I hadn’t heard of -met, but Wiktionary has an article on it, saying:

meaning “self“, and it intensifies substantive and less frequently adjective personal pronouns, it is usually followed by “ipse“

At roughly that point I recognised it from Temet nosce, the phrase used in The Matrix to translate “know thyself” (rather than the usual translation of the original Greek Γνώθι σαυτόν, Nosce te ipsum).

It stands to reason that Romansh medem/madem also comes from the same source, so I thought I might derive the Engadinese word for “same” similarly. (That’s only western and central Romansh, though; Ladin has listess, presumably from the same route as Italian stesso, i.e. st + ipsu < istum ipsum.)

So I thought, what would be the equivalent of -met ipsissimum?

I don’t know of a clitic for “self”, but ipse is presumably εμαυτόν σ(ε)αυτόν etc., or perhaps αυτός; and -issimus would be -(ό/-ώ)τατος (with the vowel belonging to the stem of the adjective).

So ipsissimum would be something like αυτότατον—though I’m not certain of the stress. For “Greek stressed like Latin”, I’d have to know the quality of the alpha in the ending, which I can’t find out right now, so I’ll assume it’s short, in which case the stress would go on the antepenult.

So perhaps the word for “same” could end up something like τόδατ todat? Not sure, but it’s an idea.

(While doing research for the vowel, I came across -γε as in έγωγε έμοιγε, but that clitic seems to be restricted to only three or four forms. Still, it might give -γε αυτότατος > γαυτόδατ gautodat?)

So, for example, the word for “head” in some languages derives from a word for “pot”; the word for “liver” from “figgy” or “fig-stuffed”; and so on.

One of those changes is “to eat”, which in some languages (e.g. French manger, Italian mangiare, Romansh mangiar) derives from manducare, which (I gather) originally meant something along the lines of “chew” or perhaps “gnaw”.

So I thought that for Engadinese, I might derive the basic “eating” verb from an Ancient Greek meaning “chew”.

So I turned to my trusty Langenscheidt German–Greek dictionary, looked up “kauen”, and found—τρώγω.

Imagine my disappointed when I saw that Modern Greek had beat me to the punch! (Because that’s the basic verb for “eat” nowadays, either in that form or, perhaps more commonly, in the shortened form τρώω. Its aorist stem is φαγ-, though, which harkens back to the suppletive aorist 2 stem  of the basic Ancient Greek verb εσθίω.)

Ah well

Most importantly, a word such as σοφός could not really give σοφ or the like, so I settled on something like σοφό—but then you have masculine/neuter words ending on consonant or stressed omicron, which seemed odd to me. Similarly, contract verbs with their final stress also didn’t act much like non-contract verbs.

So, perhaps something that I could do would be to do a wholesale stress change, to make Greek stress more Latinate, before going on to do sound changes.

That’s not unheard-of; after all, that’s what happened at some point in Proto-Germanic, where the PIE accent turned into a stem-initial stress. And, though I know less about such things, presumably also more or less what happened on the way from Proto-Slavic (with, presumably, variable stress) to Polish with its fixed penultimate stress.

It would also be interesting to see what changes to inflectional morphology (and indeed, to the entire feel of the language) such a stress change would bring about. (For instance, it might be easier to merge final -η with -α if the final syllable is never stressed.)

I think I’d have to study Latin stress a bit more carefully first, in order to get a bellyfeel for it.

Engadinese would, of course, not have no final stress; it would almost certainly acquire it through dropping of formerly-final syllables. (See e.g. Romansh -ziún from -TIÓNEM.)

In modern Romance languages, of course, this is typically accomplished with a preposition derived from Latin DE. The question then becomes, how to do this in Engadinese?

We can’t turn to Modern Greek for help, because the genitive is still alive and well there. And Ray called my first attempt, από, unsatisfactory, instead proposing εκ for TAKE. So that’s one possibility.

But a thought came through my mind: why not look at how other languages do this—specifically, non-Romance languages? And even more specifically, Maltese?

If I recall Bonġornu! Kif int? correctly, the Maltese possessive particle ta’ comes from a construction with an Arabic noun mata‘ or similar, meaning something along the lines of “possession”—so something like il-kiteb ta’ ommi “my mother’s book” comes from something like al-kitāb mata‘ ’ummi “the book, the possession of my mother”.

So perhaps that might be an idea worth following up. Though Greek does not, of course, have the construct state that inspired this construction in Maltese.

But perhaps it’ll lead to something more “organic” than just using εκ.

Though I don’t have the same problem with other contractions of preposition + article such as “cul, pel (pil?), dil (dal?)”.

I think it’d be clearer if it were spelled “e’l” as is (I believe) the case in Ladin.