former yu

  • Randy Cunningham: Thinks they're watching a generic superhero show.
  • Howard Weinerman: Has very strong feelings about secondary characters.
  • The Ninjanomicon: Consistently fails to differentiate between canon and fanon.
  • Hannibal McFist: Watches more reality TV then they'd care to admit.
  • Willem Viceroy III: Is any person who watches "Randy Cunningham: 9th Grade Ninja."
  • The Sorcerer: Will watch anything with Tim Curry.
  • First Ninja: Weeaboo scum.
  • Theresa Fowler: Likes watching boys and girls KISS on the MOUTH!!!
  • Debbie Kang: Wishes there were more shows for kids about smart people. Why aren't there are more smart people, huh?! Why's the good guy always have to be as dumb as a box of rocks, huh?! WHAT ABOUT ALL THE SMART KIDS WHO NEED ROLE MODELS?!?!?!
  • Heidi Weinerman: Still thinks RandyxHeidi has a shot at being canon.
  • Morgan: Has a collection of fictional waifus.
  • Bucky Hensletter: Regularly adopts stray animals only to throw them back out once they tear up the furniture.
  • Julian: Hot Topic scum.
  • Flute Girl: Will watch anything with Grey deLisle.
  • Bash Johnson: Has very strong feelings about stupid people.
  • Marci McFist: Likes Fargo and A Prairie Home Companion.
  • Rachel: Only likes characters who are good role models for real people.
  • Stevens: Has a collection of fictional husbandos.
  • The Sorceress: A strong independent woman who don't need no man, who also loves star-crossed lovers narratives.
  • Catfish Booray: Has bizarre sexual interests and is more than okay with that.
  • Mac Antfee: Lives in the past. Still uses DeviantArt.
  • Ghoulian: Dark and edgy Hot Topic scum.
  • Levander Hart: Vivienne Westwood scum.
  • NomiRandy: Likes people more for their physical attributes than their personalities.
  • The Tengu: Has a FurAffinity account.
  • NG: Has very strong feelings about the background characters that are too niche even for the fandom.
  • Jacques: Franch, eet eez zee language of zee yaois! Hon hon hon!
  • The Art Teacher: Has a slightly less skeevy collection of Internet waifus.
  • S. Ward Smith: Know every single TV Trope by heart.
  • Mort Weinerman: Generally a good judge of character.
  • Doug: Former Yu-Gi-Oh fan who specialised in Téa-bashing fanfics.
  • The Messenger: Went through the mammoth effort of slowing down Big Hero 6's microbot montage so they could take pictures of every single second.
  • Accordion Dave: Has only just realized that they dress like Caliborn from Homestuck.
  • Cass Simonson: Does not play well with others.
  • The Disciplinarian: Gemini.
  • The Killer Potatoes: A little sh*t that likes making things harder for me!
  • Jed and Scott: Likes watching the commentary reels, getting to see inside the creator's heads, finding all those little behind-the-scenes secrets that make the show tick.
  • Danny Phantom or Jake Long: Not supposed to be here.

gohtcheese  asked:

In your book, you said that it is your "strong recommendation that infixes not be used ... unless the language is being developed from a proto-language" what are some ways to create realistic infixes?

The proper answer to this question is this book, written by Alan C. Yu (my former GSI at Berkeley!), and I highly recommend it. The advice I gave above, though, comes in response to a particular pattern beginning conlangers fall into that is especially destructive when it comes to infixes. Suffixes and prefixes it doesn’t matter so much (if the aim is a naturalistic language), because you’ll find a language like Turkish where (almost) everything is a suffix, and a superficial analysis will render a single meaning for each one. Thus you can just take the meanings, erase the forms, and replace them with your own—maybe even add a few meanings or alter them—and the result will be something one can defend as “naturalistic” since it can’t be proved to be altogether different from a superficial analysis of Turkish. Infixes, though, are much more restrictive in their distribution and positioning. The fact that they’re listed as a type of affix akin to prefix and suffix makes it seem as if they should be equally as common and equally as simple, and that’s a shame.

To give a concrete example, let’s look at Tagalog. If you take a look at the Wikipedia page on Tagalog grammar, you’ll come across a table that lists this information:

  • Actor Trigger I: -um-
  • Actor Trigger II: nag-
  • Actor Trigger III: na-
  • Actor Trigger IV: nang-
  • Object Trigger I: -in-
  • Object Trigger II: i-…-in-
  • Object Trigger III: -in-…-an
  • Locative Trigger: -in-…-an
  • Benefactive Trigger: i-…-in-
  • Instrument Trigger: ipinaN-
  • Reason Trigger: ikina-

Many have viewed this information and come away with exactly two pieces of information:

  1. Triggers exist!
  2. A language can have infixes!

Nomenclature aside, this is absolutely true. The unfortunate result, though, has been conlangs that do the following:

  • Actor Trigger: -ak-
  • Object Trigger: -mu-
  • Subject Trigger: -el-
  • Locative Trigger: -fu-
  • Benefactive Trigger: -gi-
  • Instrument Trigger: -op-
  • Reason Trigger: -an-
  • Illative Trigger: -so-
  • Inessive Trigger: -at-
  • Elative Trigger: -eb-
  • Dative Trigger: -or-
  • Ablative Trigger: -vi-
  • Allative Trigger: -zo-
  • Genitive Trigger: -ik-

Etc. There’s not necessarily anything wrong with this: It’s just monstrously unnatural. If the goal is NOT a naturalistic language, it’s fine to do something like this (in fact, I’ve done something like this, in a language that none could ever claim is trying to be naturalistic). Doing something like this in a naturalistic conlang, though, misses some dire generalizations that can be drawn from the Tagalog information presented above—namely:

  1. All told, there are (arguably) only 7 or 8 affixes in that entire list, and furthermore…
  2. Three of the actor triggers (II, III, and IV) look like simple morphophonological variants of one another; two pairs are identical in form (Object Trigger II and the Benefactive Trigger, as well as Object Trigger III and the Locative Trigger—in fact functionally, the only major difference is the benefactive and locative forms have no imperative form—but trivially so, as they can’t be used in the imperative); and the final two are actually combinations of several affixes, including the -in- infix.
  3. Despite the fact that there are 12 “triggers”, there are only 6 “meanings” (the first 7 comprise just the actor and object roles).
  4. There are only two infixes above; everything else is a prefix or suffix.
  5. There are only three affixes with the form VC, and they are only infixes or suffixes.

That’s just from the information I gave you, but it betrays greater complexity in Tagalog than the large list of “triggers” in the conlang above does. There’s some more information that’s even more revealing, though, regarding the use of the two infixes in Tagalog. Here’s a partial paradigm to help illustrate:

  • bilí “buy (stem)” ~ bumilí “buy (ATI)” ~ binilí “buy (OTI)”
  • alís “leave (stem)” ~ umalís “leave (ATI)” ~ inalis “leave (OTI)”

[Note: Apparently inalis is not a typo; there’s a stress shift—at least according to my book.]

Notice that the infixes become prefixes for the second stem. The difference? The word begins with a vowel. Now one final piece of information. The forms I gave you above are past tense forms. For the OTI, compare them to the imperative (switching to basâ “be wet” for reasons of clarity):

  • binasâ “was wet (OTI)” ~ basaín “be wet!”
  • inalis “was removed (OTI)” ~ alisin “leave!”

This should help to elucidate what’s going on here with the infixes. Whatever they were, they’re moving around now—probably because at one point in time they were separate words, and they moved around for emphasis. Then what happened is Tagalog doesn’t like VCC sequences. Many languages don’t. There’s a number of ways to take care of VCC sequences. One very novel way, though, is to simply metathesize them—in this case, an entire prefix. Thus, we have, with our two stems from above:

  • *inalis > inalis
  • *inbasâ > binasâ

Of course, this process is unnecessary when the stem begins with a vowel, as there’s no impermissible sequence. This rather reminds me of German, in fact. In German, you have regular verbs, like arbeiten, “to work”, and gehen, “to go”, that are used in the regular verb kind of way, but then you also have verbs that take a small set of prefixes and turn them into new words, like auf-. That produces words like aufarbeiten “to finish off” and aufgehen “to open”. (Note: Got these from here; heretofore unfamiliar with the meanings.) When used in a sentence as the main verb, the prefixes actually separate, like so:

  • Ich arbeite. “I work.”
  • Ich arbeite es auf. “I finish it off.”
  • Ich gehe. “I go.”
  • Die Tür geht auf. “The door opens.”

Notice this prefix moves to a kind of focus position when the verb is the main verb of the sentence. That is, it moves to the end—kind of like the Tagalog infixes. When it’s not the main verb, though, the prefix stays on as part of the infinitive:

  • Ich werde arbeiten. “I will work.”
  • Ich werde es aufarbeiten. “I will finish it off.”
  • Ich werde gehen. “I will go.”
  • Die Tür wird aufgehen. “The door will open.”

Now imagine for a moment that German didn’t like VCC sequences (perish the thought!). What could it do with these prefixes? Mightn’t it do something like this?

  • *aufarbeiten > aufarbeiten
  • *aufgehen > gaufehen

And there you’d have naturalistic, Tagalog-style infixation in German. (Well, naturalistic apart from the fact that German LOVES its consonant clusters and would never do anything to a VCC sequence like [aʊfg].)

So, looking at Tagalog from a design standpoint, one might ask, “Why the infixes?”, and the answer is, “Because those prefixes happened to be of the form VC, and the language infixed them to resolve the VCC sequences”. These are two infixes from one language. All infixes have stories like this one that are much more interesting that the story any prefix or suffix tells—or circumfix, for that matter. And they’re very different. You see “infixes” that are paradigmatic (like the -t- that shows up in certain Arabic verb forms but has no specific meaning; it’s just a part of a verb form), infixing suffixes (as opposed to the infixing prefixes of Tagalog), positional infixes (attached to where the stress is in the word, or near it, or attached to a certain syllabic position in the word), imitative infixes (i.e. if Rosa > Rosita, then Oscar > Osquítar), etc. They all have reasons for existing that warrant explanation, and that result in certain types of behavior unique to their history, and they can’t be created as an afterthought the way (at least morphophonologically) prefixes and suffixes can. That’s the reason I suggest not doing anything with infixes unless some thought has gone into their etymologies. Otherwise you end up with an absolute mess, like I had in my language Sathir, with its “first” infixes and “second” infixes… Bleh! Total trash fire.

(Note: Don’t be surprised if this becomes a video on my YouTube channel. It’s due.)



A country of 8 000 000 people, almost destroyed by wars, demonized with propaganda and shit like that, and always SO SUCCESSFUL in SPORT… GOD DAMN, WE ROCK!

[social sandbox] plucking out a quote or fact, YO, Spotify Map

PLUCKING OUT A QUOTE OR FACT CAN BE THE BASIS FOR A BLOG POST: Former Kroc Fellow Alan Yu writes: “I read Gizmodo regularly and I came across this post, which is pretty much just one line from an interview on Fresh Air. I thought it was nice that the blogger just pull one fact out of a story and made it a separate post, kind of like what I remember you said you’re trying to do with Snapchat. I imagine it’s also driving traffic back to the NPR site as well.”


PEOPLE LISTENING TO THE SAME PIECES ON SPOTIFY AT THE SAME TIME ARE NOW ON A MAP: Map.  (Brainstorm: Could we do this for radio stories? There’s a social element there if you can chat or talk with people listening to the same pieces at the same time you are.)

ADDING SOUNDS TO GOOGLE STREET VIEW: ProjectCode  (Another interesting thought: what happens when you add geographic metadata to audio, particularly for mobile audiences?)

NEED AN EASY WAY TO SORT THROUGH ALL OF THE LINKS YOUR FRIENDS SEND ON TWITTER? Vellum is a reading layer for your Twitter feed, created by the NYTimes R&D dept. 

See something? Let us know!