This year Peterson opened the contest to entries in High Valyrian, with the stipulation that whereas Dothraki entries should be based on syllable count (just as we do with haiku in English), HV entries should be based on mora-count, just as is done in Japanese.
This turns out to be easier said than done. For one thing there is a lot of uncertainty as to how to treat word-final consonants (a problem Japanese doesn't really have, except in the case of –ん). I treated them as closing a syllable if the next word begins with a consonant (and we will see why below), others treated them as always closing the syllable, but the official ruling was that they should be "extrametrical"—i.e. we should ignore them entirely. This rule was eventually amended to be optional. This was probably a wise move, as people are having endless difficulties composing meaningful HV haiku in a 5:7:5-mora structure (which is not to say that there haven't been some terrific submissions that made the cut!)
Mr. Peterson is leaning towards declaring High Valyrian unsuited to haiku, and suggests there must be some verse form that would work better.
My answer is that it works perfectly for Greco-Roman quantitative verse! The meters used by the Greeks and Romans (and the somewhat similar meters used in the Indo-Iranian languages) were based on patterns of "heavy" and "light" syllables. Traditionally, though, the preferred terminology is "long" and "short" syllables, which is unfortunate, because it can easily be confused with the concept of long and short vowels, a related, but not precisely identical issue.
- A syllable is "short" (light) if it ends with a short vowel.
- A syllable is "long" (heavy) if it ends in a long vowel, a diphthong, or a consonant.
That thing about consonants is particularly important. As a rule, it boils down to this: if a vowel is followed by two consonants—and, critically, the two consonants don't even have to be in the same word—the syllable counts as long. So, for instance the Latin word dăt "he/she/it gives" is considered a short syllable if the next word begins with a vowel, and a long syllable if the next word starts with a consonant!
For more information on how syllable "weight" or "length" works in general, see here. For High Valyrian in particular, see here.
So, the terms "long" and "short," while potentially confusing, are quite literal: they refer to the amount of time required to pronounce a syllable. Whereas the usual forms of poetry in English rely on accent (to put it simply, patterns of "loud" and "soft"), ancient Greek and Roman poetry relies on timing (patterns of "long" and "short"). It's somewhat like the difference between "beat" and "rythm."
By far the most famous of the classical meters is the "dactylic hexameter," the verse used for epic poetry (among other things), so that is what I will be writing here. Now, this entry is going to get pretty long if I explain every detail, but in sum, the basic structure of dactylic hexameter is as follows:
- Each line consists of six "feet" (which is why it's called "hexameter")
- A foot may be either a dactyl (¯ ˘ ˘), or a spondee (¯ ¯).
- The first four feet can be either, interchangeably.
- The fifth is almost always a dactyl, but a spondee is admissable.
- The sixth foot MUST be a spondee—with the caveat that the last syllable of a line is always considered to be of indeterminate length, so in fact either ¯ ¯ or ¯ ˘ are considered acceptable. (This last syllable is usually marked with a symbol looking something like ⦣, or ¯̀ instead of a macron, to reflect this fact.)
Now that that's out of the way, let's get to business. Here is my first ever hexametric composition in High Valyrian. It's only two lines, but written to sound like the beginning of a lengthy epic:
Ābre se zaldrīzī bone ivāedan hen Essot jitte
ēlī Pento se Dothrakoti Embraro rȳ ondoso vējo...
Some of my readers will know Latin but not HV, so here's a literal Latin translation:
Feminam, et dracones, illam cano ex Esso missam
primum per Penton et Dothrachorum Mare, manu Fati...
I won't even try to do it in order in English:
I sing of that woman, and dragons, sent from Essos
by the hand of fate, first through Pentos and the Dothraki sea...
I'm sure that if I'd written a few more lines, she would have eventually made it va Vestero rāenȳti ("to the shores of Westeros"), but I decided to quit while I was ahead. Given that I kind of doubt even GRRM has gotten her there yet (no spoilers!), I can't feel too bad about it.
In terms of the language, the only thing I've really made up here is ivāedan "I sing of." This is the applicative form of vāedan "I sing," nothing really inventive about that. The only issue is that I'm using it to specify the topic of singing, whereas there's a very good chance DJP will declare that ivāedagon means not "sing of," but "sing to." Clearly this hypothetical epic would not be sung to the woman and dragons (at least not exclusively), so I hope I can get away with this.
Ā͕brĕ sĕ | zāldrī͕|zī bŏ͕ne͜ ĭ|vā͕edăn † hĕn | Ē͕ssōt | jī͕ttḕMost of these symbols are conventional. The only one that isn't is the "musical accent" mark (˲) I used to mark stress—I figured there were enough squiggles over the words as it was, and didn't want to worsen it with something that is not that important to the meter.
ē͕lī | Pē͕ntŏ † sĕ | Dōthrăkŏ͕|ti͜ Ē͕mbrărŏ | rȳ͜ ō͕ndŏsŏ | vē͕jṑ
Still, for those who haven't studied Greco-Roman poetry before, here are some explanations:
- ¯ represents a long/heavy syllable (not necessarily a long vowel!)
- ˘ represents a short/light syllable.
- ¯̀ marks a syllaba anceps—the final syllable of a line, which is always counted as long, irrespective of its actual quantity.
- | demarcates a foot.
- † marks the primary cæsura (if you don't already know what that means, don't worry about it.)
- ͜ marks an elision, a very important concept I will explain below, in the Adoptation issues section.
Hexameters are difficult to write (which is why this entry will almost certainly be disqualified). They are also difficult to recite. This is particularly true for speakers of languages like English, where vowel length is non-contrastive, and essentially exists only as a secondary feature of word stress (and yes, a few other things). Generally the easiest way for an English speaker to draw out a syllable is to emphasize it. But in languages like Greek, Latin, and High Valyrian it is quite possible for an unstressed syllable to be long, or a stressed syllable short, and English speakers, even ones who have studied the languages and meters for decades, are apt to mispronounce them. For instance, almost every Anglophone Classicist I know says Árma virúmque canó, even though the word is cánō. Heck, if anything we even exaggerate this false-stress when reciting poetry.
But there are a few people who make the extra effort to be able to do this, and do it well. This is a valuable skill, because it really helps one understand that this is genuine poetry, and not just a pointless, arbitrary, and at best ostentatious mathematical exercise. I have convinced two such people to make recordings of my composition for me.
The first performer would prefer to remain anonymous. You may find his recitation here, or just press the play button below. It is a near perfect recreation of how hexameters are supposed to have been recited in Roman times:
The second performance is by a friend known to the conlanging community as Cachorro Muzkiz. He has developed an unconventional, singsong approach to reciting hexameters that very likely resembles the way they were originally performed in the Archaic Era. You can listen to some more extensive recitations by Cachorro here.
When the Romans adapted quantitative verse forms, such as dactylic hexameter, from the Greeks, they changed some rules to account for differences between Greek and Latin, but kept others whether or not they made sense for their own language. We have a similar situation here: on the one hand, I've made some changes to the rules to account for particular characteristics of High Valyrian; on the other hand, for the most part I've stuck with the rules for Latin, whether or not they strictly apply to HV.
Particular to HV:
- ‹Z› counts as one consonant. In Classical Greek, ‹ζ› originally represented a complex sound (either [zd] or [dz], or most likely both, depending on the time and place) that gradually assimilated to [zz], then finally [z]. But, whatever the chronology of this change, Greek verse always treated it as a double consonant, no matter what the era. The same applied to ‹z› in Latin verse. So a Greek or Roman would be a bit disturbed by se being scanned as a short syllable before zaldrīzī, but it makes good sense for HV.
- Although it didn't come up in this particular composition, HV treats rising diphthongs very differently from Latin, in a way that should be reflected in verse scansion. In HV, a diphthong that consists of /i/ or /u/, followed by another vowel, is treated as a single vowel having the quantity of the second vowel. Thus, a word like obūljarion would scan as ŏbūljăriŏn (presuming the next word began with a vowel, that is). This is totally alien to Latin, where the ending -arion (e.g. in logarion "little account book") would normally scan -ărĭŏn (again, presuming the next word begins with a vowel); you could get away with treating the i as a consonant (i.e. [j]) to reduce io to one syllable, but then it also affects the previous syllable: -ārjŏn. This is not an issue for HV.
- Final -m always counts as a consonant. In Latin verse, final -m does count towards making a syllble long if the next word begins with a consonant, but when the next word begins with a vowel, the -m is ignored, and elision (see below) occurs. Apparently in this position, /m/ was simply realized as nasalization. To be fair, word-final /m/ is very rare in HV.
Particular to Latin:
- As I have repeatedly mentioned, if a word ends with a short vowel followed by a single consonant, then that final syllable counts as "short" if the next word begins in a vowel, and "long" if it begins with a consonant. Although it is uncertain how a final consonant should be treated in HV haiku, for hexameters I definitely think the rule used by Greek and Roman (and I believe Indo-Iranian as well) verse should apply.
- H is invisible to scansion. It doesn't count as a consonant, for the purpose of making a syllable count as long, nor for preventing elision (see next item.) Thus, ivāedan hen scans as ĭ|vā͕edăn hĕn. This may not make sense for HV, in which /h/ appears to always count as a consonant.
- Elision: if a word ending in a vowel is followed by one beginning with a vowel (or h), then that final vowel does not count at all. Thus, in my composition, bone ivāedan scans as bŏn(e)͜ ĭvāedăn. This is the still true in where the first word ends in a long vowel, and the second begins with a short syllable: even then the two syllables count as one short one! Elided syllables do not seem to have been totally silent (though many English-speakers performing ancient poetry will pronounce them that way), more like they slurred into the next syllable, perhaps on a "grace note." I think the way final vowels are treated in modern Spanish and Italian music is exactly the same. Think, for example, of "The Macarena" (sorry to make you do that!), in which the chorus begins: "Dale͜ a tu cuerpo͜ alegría Macarena." There has been no hint of such a phenomenon across word-boundaries in HV, but the way rising diphthongs are treated suggests a very similar phenomenon.
- A stop followed by a liquid (e.g. pl, dr, etc.) may be counted as either one consonant or two. This is one of those rules that come from the Greek language (so far as we can tell from the Romance languages, these combinations always counted as two consonants in Latin!) In HV this rule should maybe be expanded to include other combinations. For example, DJP has said that iksos should scan as only two morae for the purpose of a haiku. But in a hexameter the first syllable could count as either long or short, as the line required (and the second would, of course, depend on what the next word was.)
High Valyrian is explicitly designed to be reminiscent of Latin, and consequently it should come as no surprise that it is suited to the same sorts of verse. Indeed, in many ways it is more suited: it is so much easier to get a short syllable in HV than in Latin!
Mind you, the two lines I composed for this contest are not perfect. There are some stylistic rules I have broken, for instance:
- I have too many words ending at foot divisions in the first half of each line (it's considered bad form in the first four feet, and good form in the last two)
- The very first line of my great epic has a spondee in the fifth foot (talk about "getting off on the wrong foot!")
So I do hope I'll get to do this kind of composition again in the future. In particular, here's something cool I noticed: if we translate "White Walker" literally into HV, we get Timpys Dekurūbaros in the nom.sing. When followed by a vowel, this would scan Tīm|pȳs dĕkŭ|rūbărŏs, which would fit perfectly into a hexameter! If we prefer to call them "The Others," we get Tolossa → Tŏ|lōssă, which will also be easy to incorporate.
So the path lies straight, as it were. Let's see what this year brings!