At Home with the Hansard

Is it cold in here?

When you think “Third World country” there are certain universals:

. . . and so on. Et cetera. ᐊᓯᖏᓪᓗ.

Uhm. I never was very good at those “Can you spot the difference?” puzzles. I know that something in that list doesn’t belong. Can’t put my finger on it, though.

But that’s OK, because neither can anyone else. As I write this, the standard habitation throughout the Territory is a single-family detached dwelling, made primarily of wood, raised well above the ground on pilings to allow free air circulation under the floor. This is a design traditionally used in . . . drumroll . . . tropical rain forests. The Nunavut version does come with the climate-appropriate shallowly pitched roof. But I’m sure this is merely an oversight, and we will shortly see the introduction of A-frames.

The Air Bubble of Consensus Floats Away

The version of the Hansard that I’ve got runs out in 2007. Eight years earlier, the newly elected Speaker said in his inaugural address:

ᐊᔭᐅᖅᑕᐅᕙᓐᓂᐊᖅᑐᒍᑦ ᐃᓱᒻᒥᖅᑕᐅᓂᕐᒧᑦ ᐊᖏᖃᑎᒌᓇᓱᑉᐸᓂᕐᒥᒃ
ajauqtauvanniaqtugut isummiqtaunirmut angiqatigiinasuppanirmik
(the spirit of consensus guides us)

OK, so it isn’t Long years ago, we made a tryst with destiny or Today I consider myself the luckiest man on the face of the earth. And he was only talking about the Territory’s decision not to have political parties. But it’s still a pretty nice line. I wish the translators had taken it a little more to heart. Or that it had occurred to someone, somewhere, to sit down with the translators and codify a list: this is our word for “inter­pretation”; this is how we say “MLA”; this is how we’re going to render “colonial legacy”, and so on. Instead . . .

Friends, Members and Congressmen!


For starters, consider the Member’s Statement. Most of them can be stripped down to the same word, maliga­liuqti: someone whose job, -ti, is to make, -liuq, the law, maliga. That ti, officially known as ji, is an interesting affix. “Job” is a fairly modern concept; it requires an urban society with a division of labor. The dictionarist helpfully tells me that ji particularly denotes a “menial or stooge”. This makes it all luminously clear. In the early days of the affix, this was the guy who had to clean up after the dogs when they’d eaten something that disagreed with them, as dogs will. Today its meaning seems to have sunk still lower.

Before the translators can regularize the Member’s Statement in Inuktitut they’ll have to do so in English. Is it “Member’s” (the statement of one member) or is it “Members’” (statements made by any and all members)? Repre­senting the plural contingent we’ve got

maligaliuqtit, maligaliuqtiit, maligaliurtiit

and probably a few others I overlooked. Those are all the same word; the translators just couldn’t decide whether to write the plural as -it or -iit. They also disagreed on whether to represent that syllable-final sound with q or r. This has apparently been a heated issue for several decades. Good thing we’ve got a Dubi­tative mode for just such situations: taitar­mangaakkit, taitar­mangaattigit, taitar­mangaarmigit . . .

Over on the singular side there’s

maligaliuttiup and maligaliuqtiup

with a third spin on the q-versus-r dilemma: If you live anywhere within spitting distance of Baffin, aka Qikiqtaaluk (“the big island”), you can avoid the whole issue—and lots of others—by ruthlessly assimilating any and all consonant pairs. You won’t see any igloos in Iqaluit; they’ve all collapsed into illuit.

Finally, there are the translators who just couldn’t leave well enough alone, giving us




One of the Members must have put in a demand for more syllables. He was rewarded with another five, almost doubling his original allotment. Our final Member—or Members, because there’s a cluster of almost a hundred going by the ilagijaujut name—dropped the “lawmaker” motif altogether. Instead they went with the “member” aspect, building on ila “friend or companion” and slathering on the affixes until they reached something properly legislatorial.

Let me State for the Record . . .

* At least that’s one thing we’ll never have to worry about. Inuktitut is utterly free of gender. When you consider that some languages have gendered verbs, you will realize how lucky you are.

All of this is before anyone has even opened his or her* mouth. Filter out the aberrations like nalunaiq­siguti­ngit or, for variety’s sake, nalunaiq­sijjuti­ngit. I really wanted this to mean “my esteemed colleague is a blithering imbecile”, but it turns out to mean only that the members—collectively or plurally—will clear up our confusion (nalu). Isn’t that nice. Apply the same filter to one-offs like unikkaangit, meaning . . . Well, let’s be diplomatic and say “account” or “narrative”, even if most words that build on the unipkaa or unikkaa root lean heavily to the “myth, fable, fairy tale” end of the storytelling spectrum. To go with ilagi­jaujut, there’s a cluster of words suggesting that the translator was simply fond of the letter l, though it probably has more to do with the dialectal influence of the Big Island:

uqalimautiksangit, uqallautingit, uqallautiksangit

And the rest:

uqausinga, uqausingit
uqausissangit, uqausitsangit
uqausiuksangit, uqausiksangi(t), uqausiksangit

That’s -nga, one thing belonging to someone else, -ngit, lots of things belonging to someone else—or to lots of someones, but definitely not two—and -ngat, one thing belonging to lots of other people. It all depends on whether the Member is an -up or an -it. The “statement” words all start out harmlessly enough with uqaq-, “speak”, and they all end up as nouns. But how they get there is entirely up to the translator.

Says Who?

Finally, there’s interpretation. No time to consult the dictionary; this is live, real-time, on the spot. And it’s duly logged in the Hansard—all but the trifling detail of what language the speaker is using, and what language it’s being inter­preted into. English, French, Inuktitut, Inuinnaqtun, a visiting American’s Iñupiat: who knows?

mirror, mirror

Sometimes you’ll meet an unmodified interpreter: ᑐᓵᔨ (tusaaji), the person whose job it is, -ji, to listen closely, tusaa-. But most often it’s ᑐᓵᔨᑎᒍᑦ (tusaajitigut). That’s the Vialis case, here doing a job that early Indo-European languages gave to the Instru­mental: “by way of an inter­preter”. Or, for variety’s sake, the plural form ᑐᓵᔨᒃᑯᑦ (tusaa­jikkut), “by way of lots of inter­preters”.

If you can’t bear to let a noun go wandering around unsuper­vised, you can add the noun-to-verb converter -uq- “go through”. The affix predictably eats the final t of the noun ending—you are only allowed to use -uq- with the Vialis, so it will always be a t—to make tusaajik­kuuq-. Wrap it up with a verb ending to get ᑐᓵᔨᒃᑰᖅᑐᖅ (tusaajik­kuuqtuq) or -ᔨᑎᒎᖅᑐᖅ (-jitiguuqtuq).

Or, for something completely different, try tukiliuqtausimajuq. This word tries to become a noun two separate times, but is firmly stomped back into verbdom each time:

If you want to be all grammatical about it, that’s:
* the “passive participle”
the “present perfect”.

ᑐᑭ tuki, noun, sense or understanding

ᓕᐅᖅ liuq, noun-to-verb, make or construct

ᑕᖅ taq, verb-to-noun, the person or thing that has this done to them*

u, noun-to-verb, be

ᓯᒪ sima, verb affix, completed action

ᔪᖅ juq, verb ending

For “interpretation ends”, the tusaa-whatever may be followed with ᓄᖅᑲᖅᑐᖅ (nuqqaqtuq) or ᐃᓱᓕᒃᑐᖅ (isuliktuq), each with its predictable spelling variants. For those who prefer the one-piece look, there’s ᑐᓵᔨᑎᒎᕈᓐᓃᖅᑐᓂ (tusaa­jitiguu­runniiqtuni) and similar forms. That’s

tusaa + ji (tigut) + uq + gunniiq + tuni

Start with tusaajitiguuq—or singular tusaajikkuuq—as above. Add the “cessation” affix -gunniiq-. It was happening before, but it ain’t no more. In a process that should by now be tediously familiar, the final q of -uq- fuses with the initial g of -gunniiq- to become r: -guurunniiq-. Finally, add a verb ending to taste—either a simple indicative -tuq or the “participle-tuni—and you’re done.

Almost. In one exceptionally long speech we’re treated to a tusaaji asijjiqtuq, meaning that we will now make use of (-jjiq-) a different (asi-) inter­preter. I don’t know why it has to be a completely different one. You would think another inter­preter of the same kind—the first guy’s brother, say—would be a better choice to preserve continuity. But since the speaker has been uttering things like

attuaniqarniaqtut kiinaujat isirpalliajunit atuqtunillu 2005-mit aturumaanniaq­tunit amiakkuk­saqa­lauqsima­galuarmata $3.6 milianit maanna $12.3 milia­nnguqtut akiliksaqarluta $8.7 milianit. ᐊᑦᑐᐊᓂᖃᕐᓂᐊᖅᑐᑦ ᑮᓇᐅᔭᑦ ᐃᓯᕐᐸᓪᓕᐊᔪᓂᑦ ᐊᑐᖅᑐᓂᓪᓗ 2005-ᒥᑦ ᐊᑐᕈᒫᓐᓂᐊᖅᑐᓂᑦ ᐊᒥᐊᒃᑯᒃᓴᖃ­ᓚᐅᖅᓯᒪ­ᒐᓗᐊᕐᒪᑕ $3.6 ᒥᓕᐊᓂᑦ ᒫᓐᓇ $12.3 ᒥᓕᐊᙳᖅᑐᑦ ᐊᑭᓕᒃᓴᖃᕐᓗᑕ $8.7 ᒥᓕᐊᓂᑦ. The net impact of these revenue and expen­diture adjust­ments to the 2005 Budget means that the original budgeted surplus of $3.6 million dollars, has now changed by $12.3 million dollars to a projected deficit of $8.7 million dollars.

you can see the original tusaaji’s point.

I Heard “marruk” Too

* You don’t really need an attribution, do you?

Interpreter,* n. One who enables two persons of different languages to under­stand each other by repeating to each what it would have been to the inter­preter's advantage for the other to have said.

I may have found the answer to the interpretation puzzle. But it wasn’t a from-the-horse’s-mouth source, so let’s hold on to that “may”. Call it ᓇᓂᓯᒪᔭᕋᖃᐃ (nanisima­jaraqai) or possibly even ᓇᓂᓯᒪᔭᕋᑭᐊᖅ (nanisima­jarakiaq).

The explanation goes like this: in the early years, the Hansard was recorded only in English. If someone spoke in Inuktitut, the official record did not show their words but the inter­preter’s. The Inuktitut Hansard was produced after the fact by trans­lating the entire text—including the parts that were spoken in Inuktitut in the first place—from English into Inuktitut, with accom­panying “inter­pretation” label, if any. So what you’re reading is not what the speaker really said, but a trans­lation into Inuktitut of an inter­pretation into English of what the speaker said.

Is that too confusing? Pay close attention now:

no “interpretation” = the original was in English, so you are reading exactly what was said;

“interpretation” = the speaker used Inuktitut, so you are reading the inter­preter’s rendition;


no “tusaajitigut” or equivalent = the original was in English, so you are reading the translator’s Inuktitut version;

“tusaajitigut” or equivalent = the original was in Inuktitut, rendered into English by an inter­preter, and subsequently retro-translated into Inuktitut by someone else.

Mr. Bierce, you don’t know the half of it.

A quick eyeballing of the Hansard reveals that this two-step process lasted from Session 1.2 (beginning May 1999) through Session 2.2 (ending May 2005). Concurrent English and Inuktitut versions don’t come in until Session 2.3 (beginning November 2005). From there on, you will see “inter­pretation” or “tusaajitigut” on either the English side or the Inuktitut side, but not both. Unless, of course, the speaker used French. Or Inuinnaqtun, or Iñupiat, or . . .

What about that historic Session 1.1? Lacking original transcripts, it’s impossible to tell. When the word “inter­pretation” or “tusaajitigut” does appear, it’s only on one side or the other. But generally it’s absent from both sides—and I’m tolerably certain all those people from Ottawa weren’t talking out of both sides of their mouths concurrently.

Back in May 1999, in the intervals of trying to get the Speaker to pronounce his name right, the Hon. Mr. Enoki Irqittuq ran into a further tangle.

iksivautaq (tusaajitigut):
uqaravit 5.3-mik matuiqqaugakkit amma iqqaq­tuiviliri­nikkun­nuungajut.
ᐃᒃᓯᕙᐅᑕᖅ (ᑐᓵᔨᑎᒍᑦ):
ᐅᖃᕋᕕᑦ 5.3-ᒥᒃ ᒪᑐᐃᖅᑲᐅᒐᒃᑭᑦ ᐊᒻᒪ ᐃᖅᑲᖅ­ᑐᐃᕕᓕᕆ­ᓂᒃᑯᓐ­ᓅᖓᔪᑦ.
Chairperson (interpretation):
When you said 5.3 I open the page and it says Department of Justice.
mista iqqittuq (tusaajitigut):
uqaqqaunngintunga 5.7-mik. uqaqqaujunga 7.3-mik. qujannamiik.
ᒥᔅᑕ ᐃᖅᑭᑦᑐᖅ (ᑐᓵᔨᑎᒍᑦ):
ᐅᖃᖅᑲᐅᙱᓐᑐᖓ 5.7-ᒥᒃ. ᐅᖃᖅᑲᐅᔪᖓ 7.3-ᒥᒃ. ᖁᔭᓐᓇᒦᒃ.
Mr. Irqittuq (interpretation):
I didn’t say 5.7 I said 7.3. Thank you.

Let’s assume for the sake of discussion that he neither said nor denied saying “5.7”. Blame the clerk. A bit further along, it becomes clear that the question is “7.3” vs. “5.3”.

Incidentally, the translation seems a big slapdash:

* When the Legislative Assembly has only nineteen members, representatives who do not hold a minis­terial portfolio are few and far between.

uqaravit = uqaq + gavit, (be)causative present, “You say”

matuiqqaugakkit = matu + iq + qqau + gakkit, (be)causative past with plural object “I opened (the pages)”

iqqaqtuivilirinikkunnuungajut . . . OK, can we stipulate that this is “Department of Justice”?

Mr. Erqittuq’s colleage, the equally Honorable* Mr. Anawak, is no help:

niqtunaqtuq anaruaq (tusaajitigut):
tusarataaqtunga 5.32-mik.
ᓂᖅᑐᓇᖅᑐᖅ ᐊᓇᕈᐊᖅ (ᑐᓵᔨᑎᒍᑦ):
ᑐᓴᕋᑖᖅᑐᖓ 5.32-ᒥᒃ.
Hon. Jack Anawak (inter­pretation):
Well I heard 5.3 too . . .
iglaqtut ᐃᒡᓚᖅᑐᑦ Laughter

Do you suppose the laughter was directed at the interpreter? As recorded for posterity, Mr. Anawak said “5.3 too”. The inter­preter heard it as “5.32”, which is dutifully rendered in the Inuktitut text as “5.32-mik”.

That’s an assumption, of course. But it seems safe. I can’t think of anything that could be misheard as ᒪᕐᕈᒃ (marruk), “two”, leading to a further cascade of mishearings or mistrans­lations until we reach “too”. Unless someone wants to postulate that Mr. Anawak is further muddying the waters by claiming to have heard neither 5.3 nor 7.3 but 5.32.

The ᒪᕐᕈᒃ (marruk) form, with short u, is only for when you’re counting on your fingers or plodding through subclauses. If the two whatevers are actively doing something, the word would naturally take the dual form marruuk. Or marruungnik, if it’s you doing something to the two of them. The “5.32-mik” is exempt because it’s a single utterance: Mr. Anawak does not claim to have heard “5.3” twice.

Numbers in Inuktitut are officially nouns. But you could make a case for numbers being the only “true” adjectives. They go in tandem with nouns and match their inflection, just like in Indo-European languages.

So let’s work through what we know about those two lines of text.

Given: This is 1999, so only the English text was recorded. The word “inter­pretation” or equivalent will therefore be present on both sides or on neither side.

Given: The presence of the word “interpretation” means that the speaker originally used Inuktitut, but only the English inter­pretation was recorded.

Given: The only plausible explanation of the “5.3 too” vs. “5.32” error is that the speaker used English, which was misheard by an on-the-spot inter­preter. Each version was then separately recorded.

Mr. Anawak left the Assembly before Session 2.3, so there’s no way to be certain about his language. In Session 1.1 he spoke twice. One is flagged as “inter­preta­tion” in English; the other is tusaa­jikut in Inuktitut. That does not get us far. Subsequent Hansards would have us believe he spoke through an inter­preter about two-thirds of the time. Or, if you prefer, that the clerk forgot the word “inter­pretation” one-third of the time.

If I find the Inuktitut word for “oxymoron”, will “logical paradox” be right behind it?

Marruuk Redux

* Known, naturally, as the inuksuk.

When all else fails, read the instructions. Problem is, Inuktitut doesn’t come with a manual. Option two—asking someone who knows the language—is also out. It seems that the person responsible for feeding Nunavut’s communications satellite* has been sleeping on the job, so the satellite has to sustain itself by eating my e-mail. Well, except when it’s addressed to someone so wildly famous that the inuksuk could never get away with it. Then it bounces back with a code 5.1.1, e-mail-speak for “Addressee Unknown”. If I ever had occasion to communicate with the Premier of Nunavut, the e-mail server at would undoubtedly tell me they’d never heard of the lady.

Let’s try option three: Ask the world at large. This netted me a series of explana­tions, most of them requiring some spin on the original “mishearing” inter­pretation.

Hypothesis #1: The typist had fat fingers. Or was wearing mittens—a perfectly reasonable idea on the face of it. But don’t be fooled by those photographs of men dressed in layers of shaggy fur, pecking away at laptops. Computers will not work in below-freezing tempera­tures. You can’t even store them safely below about -4°F (-20°C)—a balmy spring day by northern standards.

Hypothesis #2: The written translation, like the spoken inter­pretation, involved two people. One read out the English text; the other listened and prepared the translation. In other words, exactly what logic says happened in the first place—only shifted to a different pair of people.

Hypothesis #3: Again two people, but this time the two are the trans­lator and the typist. This becomes more likely when you remember that this was the Prosyl Era. Syllabics couldn’t be typed by just anyone—at least not if you wanted it done overnight, as the Blues were. The typist had to be familiar with the keyboard. So the translator does her stuff, speaks it into a Dictaphone or 1999 equivalent, and hands the tape over to the typist.

Sorry, I’m Going Birds’-Nesting That Day

Old-school journalists had a canonical list of questions: who, what, when, where, why. Oh, and how, even if it does spoil the alliteration. We’ve already met some of the “who”. The “where” presents no problems. It’s that shiny-looking new building in Iqaluit with the plumbers and heating contractors standing around scratching their heads and asking questions beginning “Why did . . .?” or “How can . . .?” Or, depending on ethnic background, ᓱᒻᒪᑦ summat . . .? or ᖃᓄᖅ qanuq . . .?

When you get to the “when”, things get messy. Let’s not even talk about calendrical months. The English ones, that is. There’s a perfectly respect­able set of Inuktitut names for traditional months. At least the ones in the warmer half of the year; for the other half all you need is Ay-yup. Still below zero.

Look, instead, at the days of the week. For Thursday alone, we’ve got:

ᐅᒡᓚᕐᓇᖅ uglarnaq
ᕿᑎᖅᑐᑦ qitiqtut, ᕿᑎᖅᑰ(ᖅ) qitiqquu(q)
ᓯᑕᒻᒥᑐᖅ sitammituq, ᓯᑕ(ᒻ)ᒥᖅ sitam(m)iq, ᑎᓴᒻᒥᖅ tisammiq

* This makes perfect sense—so long as the week starts on Sunday. By the time the German calendar was revised to make the week start on Monday, it was too late to rename the day. And no use saying Mittwoch refers to the five-day workweek: that, too, is a recent innovation. Labor-movement posters showed a little girl with her father, saying On Saturdays, Daddy belongs to me.

The word uglarnaq is apparently the day you go visiting, since that’s what the verb uglaq- means. Anything in qitiq- means “the middle of the week”. Everything else is the fourth day, with a double-wide range of forms thanks to dialectal variation: “four” can be either sitamat or tisamat. Which is inter­esting, by the way. In the rest of the world we think of the middle of the week as Wednesday. Mittwoch, that kind of thing.* But the Inuit name is mathematically unim­peachable, because the middle of a series of seven is not the third, it’s the fourth. The seven-day week is a pretty recent introduction to the far North; it needed fresh eyes to look at this kind of thing logically.

Friday may be just as bad, even without the t-for-s swap:

ᑕᓪᓕᒥᖅ tallimiq, ᑕᓪᓕᕐᒥ tallirmi
ᐅᓪᓗᑐᐃᓐᓇ(ᖅ) ullutuinna(q)
ᓂᕿᑦᑎᔪᖅ niqittijuq

* Latin ullus, -a, -um, giving rise to the more familiar negative form nullus.

The first two are simply variations on the number five. Yawn. But at the first sight of ullutuinnaq, severe linguistic contami­nation set in and my mind attempted to make it mean “any old day”.* It would be fun if it were the day everyone goes birds’-nesting, that being what the noun ublu or ullu means, but I guess we’ll have to be grownup and stick with ubluq or ulluq, “day”. With added -tuinnaq, it appears to mean “just another day”. In other words, exactly what my language-confused mind originally wanted it to mean—only I’d got the elements backward. There is nothing in Latin remotely resembling -tuinnaq. In any language, it seems an odd way to describe the end of the work week.

The last of the batch, niqittijuq, means “We’re all good Catholics up here”. The equally well attested variant niqituqtaili is mysteriously absent from the Hansard.

Stop that non-alphabetic! It’s stealing a syllable!

I don’t know how much the Government of Nunavut paid the typists and computer folks to put those early years’ transcripts into Prosyl. But I am absolutely certain it was too much. Back when I first looked into duct tape, I was mystified to find that the preliminary version (the “Blues”) had


while the final, official—and presumably corrected—version had


Or, in Prosyl, h9lo9MEsD83m5 in place of h9lo9MEs>/D8~i3m5.

* Feel free to insert a witticism about MLAs who say nii.

Note the mysterious departure of two entire syllables, ᔮ jaa (>/) and ᓃ nii* (~i). Since the subject under discussion is deteriorating water pipes, there’s a certain poignancy to the word’s fate. But the linguistic destruction isn’t limited to sulluliit. As we all know, the Inuktitut language has an ineradi­cable habit of eating single letters, especially final consonants. Eating whole syllables—long ones, at that—is definitely beyond the pale. In the Hansard, it was a recurring problem throughout the Prosyl years. You can see it in the first recorded proceedings, in May 1999, and it’s still happening on the very last Prosyl day, in March 2002. The sullulilla­riu[jaa]­run[nii]rmat paragraph begins with maannaujumi, similarly truncated to a slightly improbable ᓐᓇᐅᔪᒥ nnaujumi. Members’ names may be abbreviated to ᑕ or ᓪ (ᐲᑕ piita and ᐹᓪ paal, respectively), if they don’t vanish altogether.

* When the Government of Nunavut hires a new IT person—which appears to be every six months or so—the first thing they do is rearrange everything on the site, including the Hansard archives. Result: the page you visited last week has now vanished without a trace. Familiarity with the 301 Redirect or mod_rewrite is apparently not part of the job description.

To make up for it, pages with a clear and unambi­guous expiration date will stick around until radio­active decay sets in.

I’m not kidding. Just when you think you’ve successfully navigated the calendars and popups that point to the French text, you’re uncere­moni­ously dumped back into the English version. Not so much as a n’existe pas of warning.

You don’t have to take my unsupported word for it; see for yourself. If your browser is crotchety about following the exact link—or if the page has been moved or renamed*—backtrack to the main Hansard page, go to the Archives popup and find Session 1.3 (ᓯᕗᓕᖅᐹᑦ ᒪᓕᒐᓕᐅᖅᑏᐃᑦ—ᐱᖓᔪᐊᓐᓂᒃ ᑲᑎᒪᓇᖅ), proceedings for 29 February 2000, page 773. (This is not as horrifying as it sounds. Pagination is continuous, so the text for that day starts with page 748.) Or any other day you like. They are all in PDF format, so you may need to save them into a different application if you want to search or select text. Here’s our crumbling infrastructure:

screen shot

Digression: May 1999? What about April? Sorry, ᒪᒥᐊᓇᖅ, je regrette. Only in English. The Inuktitut Hansard does not kick in until the second session, May-June 1999. French is anyone’s guess. An Inuktitut transcript of the inaugural session must exist—it’s in the Blues—but the official version was either never published, or was subse­quently yanked. ᒫᓐᓇ ᐅᑯᓇᓂ ᑲᑎᒪᕐᔪᐊᕐᓂᕐᓂ ᑲᑎᒪᔾᔪᑎᕕᓂᕐᓂᒃ ᐱᑕᖃᙱᑦᑐᖅ (maanna ukunani katimar­juarnirni katima­jjuti­vinirnik pitaqanngittuq), says the web page. They’re not here now. You can have the unofficial version, if you like, but it’s not the same thing.

I don’t know whether the problem arose with the original Prosyl transcription, or whether—this seems more likely—it happened during the conversion from text to PDF. But here’s a final oddity. You may have noticed that the missing pieces both involve non-alphabetics: >/ (ᔮ jaa) and ~i (ᓃ nii). In Prosyl as in Roman transliteration, those long vowels aren’t single characters. They’re combinations, made by letting one of five non-advancing dots ride on top of an ordinary dotless syllable. At the left end of the spectrum you’ve got ᐆ, ᐴ, ᑑ, ᒫ, and the like. Midway along there’s ᐄ, ᐲ, ᑏ, ᑮ, ᒦ and more. Continue heading right and you’ve got ᑖ, ᑰ and ᒎ all by themselves in shortstop position. And, at the rightmost edge of possibility, there’s another good cluster: ᐋ, ᐹ, ᒨ, ᔫ and so on, including all forms of ᖐ. Finally there are the lower-level or mezzanine dots for everything involving ᓃ, ᓖ, or ᖡ.

But when a long vowel disappears from the Hansard, the dot doesn’t just pack up and roll away. It takes the whole syllable along. So either someone’s keyboard had severe jamming problems, or there were bugs—qumait, perhaps—in the Prosyl-to-PDF program.

Whatever the reason, one or more long syllables will suddenly cut out just when you least expect it. There’s no pattern or consistency; a character might be present on one page and disappear on another. Spend much time with the final, official Hansards and you will soon get used to forms like ᖁᔭᓐᓇᒃ ᐅᖃᖅ qujannak uqaq for ᖁᔭᓐᓇᒦᒃ ᐅᖃᖅᑏ qujannamiik uqaqtii. On the very last day of the Prosyl era, 7 March 2002, they ran about one ᑐᔨᑎᒍᑦ tujitigut for every two ᑐᓵᔨᑎᒍᑦ tusaa­jitigut. I was disappointed not to find any ᑐᔨᑎᕈᓐᖅᑐᖅ tujiti­runqtuq (for ᑐᔨᑎ­ᕈᓐᖅᑐᖅ tusaa­jitiguu­runniiqtuq, that day’s preferred wording). It would have lent some piquancy to the record. As well as, of course, saving a good deal of ink.